summaryrefslogtreecommitdiff
path: root/macros/latex/contrib/l3kernel/l3tl.dtx
blob: fb2a14b6f05ec725437a09f4c3ac7262b0c84500 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
1893
1894
1895
1896
1897
1898
1899
1900
1901
1902
1903
1904
1905
1906
1907
1908
1909
1910
1911
1912
1913
1914
1915
1916
1917
1918
1919
1920
1921
1922
1923
1924
1925
1926
1927
1928
1929
1930
1931
1932
1933
1934
1935
1936
1937
1938
1939
1940
1941
1942
1943
1944
1945
1946
1947
1948
1949
1950
1951
1952
1953
1954
1955
1956
1957
1958
1959
1960
1961
1962
1963
1964
1965
1966
1967
1968
1969
1970
1971
1972
1973
1974
1975
1976
1977
1978
1979
1980
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
2020
2021
2022
2023
2024
2025
2026
2027
2028
2029
2030
2031
2032
2033
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
2052
2053
2054
2055
2056
2057
2058
2059
2060
2061
2062
2063
2064
2065
2066
2067
2068
2069
2070
2071
2072
2073
2074
2075
2076
2077
2078
2079
2080
2081
2082
2083
2084
2085
2086
2087
2088
2089
2090
2091
2092
2093
2094
2095
2096
2097
2098
2099
2100
2101
2102
2103
2104
2105
2106
2107
2108
2109
2110
2111
2112
2113
2114
2115
2116
2117
2118
2119
2120
2121
2122
2123
2124
2125
2126
2127
2128
2129
2130
2131
2132
2133
2134
2135
2136
2137
2138
2139
2140
2141
2142
2143
2144
2145
2146
2147
2148
2149
2150
2151
2152
2153
2154
2155
2156
2157
2158
2159
2160
2161
2162
2163
2164
2165
2166
2167
2168
2169
2170
2171
2172
2173
2174
2175
2176
2177
2178
2179
2180
2181
2182
2183
2184
2185
2186
2187
2188
2189
2190
2191
2192
2193
2194
2195
2196
2197
2198
2199
2200
2201
2202
2203
2204
2205
2206
2207
2208
2209
2210
2211
2212
2213
2214
2215
2216
2217
2218
2219
2220
2221
2222
2223
2224
2225
2226
2227
2228
2229
2230
2231
2232
2233
2234
2235
2236
2237
2238
2239
2240
2241
2242
2243
2244
2245
2246
2247
2248
2249
2250
2251
2252
2253
2254
2255
2256
2257
2258
2259
2260
2261
2262
2263
2264
2265
2266
2267
2268
2269
2270
2271
2272
2273
2274
2275
2276
2277
2278
2279
2280
2281
2282
2283
2284
2285
2286
2287
2288
2289
2290
2291
2292
2293
2294
2295
2296
2297
2298
2299
2300
2301
2302
2303
2304
2305
2306
2307
2308
2309
2310
2311
2312
2313
2314
2315
2316
2317
2318
2319
2320
2321
2322
2323
2324
2325
2326
2327
2328
2329
2330
2331
2332
2333
2334
2335
2336
2337
2338
2339
2340
2341
2342
2343
2344
2345
2346
2347
2348
2349
2350
2351
2352
2353
2354
2355
2356
2357
2358
2359
2360
2361
2362
2363
2364
2365
2366
2367
2368
2369
2370
2371
2372
2373
2374
2375
2376
2377
2378
2379
2380
2381
2382
2383
2384
2385
2386
2387
2388
2389
2390
2391
2392
2393
2394
2395
2396
2397
2398
2399
2400
2401
2402
2403
2404
2405
2406
2407
2408
2409
2410
2411
2412
2413
2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
2431
2432
2433
2434
2435
2436
2437
2438
2439
2440
2441
2442
2443
2444
2445
2446
2447
2448
2449
2450
2451
2452
2453
2454
2455
2456
2457
2458
2459
2460
2461
2462
2463
2464
2465
2466
2467
2468
2469
2470
2471
2472
2473
2474
2475
2476
2477
2478
2479
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
2511
2512
2513
2514
2515
2516
2517
2518
2519
2520
2521
2522
2523
2524
2525
2526
2527
2528
2529
2530
2531
2532
2533
2534
2535
2536
2537
2538
2539
2540
2541
2542
2543
2544
2545
2546
2547
2548
2549
2550
2551
2552
2553
2554
2555
2556
2557
2558
2559
2560
2561
2562
2563
2564
2565
2566
2567
2568
2569
2570
2571
2572
2573
2574
2575
2576
2577
2578
2579
2580
2581
2582
2583
2584
2585
2586
2587
2588
2589
2590
2591
2592
2593
2594
2595
2596
2597
2598
2599
2600
2601
2602
2603
2604
2605
2606
2607
2608
2609
2610
2611
2612
2613
2614
2615
2616
2617
2618
2619
2620
2621
2622
2623
2624
2625
2626
2627
2628
2629
2630
2631
2632
2633
2634
2635
2636
2637
2638
2639
2640
2641
2642
2643
2644
2645
2646
2647
2648
2649
2650
2651
2652
2653
2654
2655
2656
2657
2658
2659
2660
2661
2662
2663
2664
2665
2666
2667
2668
2669
2670
2671
2672
2673
2674
2675
2676
2677
2678
2679
2680
2681
2682
2683
2684
2685
2686
2687
2688
2689
2690
2691
2692
2693
2694
2695
2696
2697
2698
2699
2700
2701
2702
2703
2704
2705
2706
2707
2708
2709
2710
2711
2712
2713
2714
2715
2716
2717
2718
2719
2720
2721
2722
2723
2724
2725
2726
2727
2728
2729
2730
2731
2732
2733
2734
2735
2736
2737
2738
2739
2740
2741
2742
2743
2744
2745
2746
2747
2748
2749
2750
2751
2752
2753
2754
2755
2756
2757
2758
2759
2760
2761
2762
2763
2764
2765
2766
2767
2768
2769
2770
2771
2772
2773
2774
2775
2776
2777
2778
2779
2780
2781
2782
2783
2784
2785
2786
2787
2788
2789
2790
2791
2792
2793
2794
2795
2796
2797
2798
2799
2800
2801
2802
2803
2804
2805
2806
2807
2808
2809
2810
2811
2812
2813
2814
2815
2816
2817
2818
2819
2820
2821
2822
2823
2824
2825
2826
2827
2828
2829
2830
2831
2832
2833
2834
2835
2836
2837
2838
2839
2840
2841
2842
2843
2844
2845
2846
2847
2848
2849
2850
2851
2852
2853
2854
2855
2856
2857
2858
2859
2860
2861
2862
2863
2864
2865
2866
2867
2868
2869
2870
2871
2872
2873
2874
2875
2876
2877
2878
2879
2880
2881
2882
2883
2884
2885
2886
2887
2888
2889
2890
2891
2892
2893
2894
2895
2896
2897
2898
2899
2900
2901
2902
2903
2904
2905
2906
2907
2908
2909
2910
2911
2912
2913
2914
2915
2916
2917
2918
2919
2920
2921
2922
2923
2924
2925
2926
2927
2928
2929
2930
2931
2932
2933
2934
2935
2936
2937
2938
2939
2940
2941
2942
2943
2944
2945
2946
2947
2948
2949
2950
2951
2952
2953
2954
2955
2956
2957
2958
2959
2960
2961
2962
2963
2964
2965
2966
2967
2968
2969
2970
2971
2972
2973
2974
2975
2976
2977
2978
2979
2980
2981
2982
2983
2984
2985
2986
2987
2988
2989
2990
2991
2992
2993
2994
2995
2996
2997
2998
2999
3000
3001
3002
3003
3004
3005
3006
3007
3008
3009
3010
3011
3012
3013
3014
3015
3016
3017
3018
3019
3020
3021
3022
3023
3024
3025
3026
3027
3028
3029
3030
3031
3032
3033
3034
3035
3036
3037
3038
3039
3040
3041
3042
3043
3044
3045
3046
3047
3048
3049
3050
3051
3052
3053
3054
3055
3056
3057
3058
3059
3060
3061
3062
3063
3064
3065
3066
3067
3068
3069
3070
3071
3072
3073
3074
3075
3076
3077
3078
3079
3080
3081
3082
3083
3084
3085
3086
3087
3088
3089
3090
3091
3092
3093
3094
3095
3096
3097
3098
3099
3100
3101
3102
3103
3104
3105
3106
3107
3108
3109
3110
3111
3112
3113
3114
3115
3116
3117
3118
3119
3120
3121
3122
3123
3124
3125
3126
3127
3128
3129
3130
3131
3132
3133
3134
3135
3136
3137
3138
3139
3140
3141
3142
3143
3144
3145
3146
3147
3148
3149
3150
3151
3152
3153
3154
3155
3156
3157
3158
3159
3160
3161
3162
3163
3164
3165
3166
3167
3168
3169
3170
3171
3172
3173
3174
3175
3176
3177
3178
3179
3180
3181
3182
3183
3184
3185
3186
3187
3188
3189
3190
3191
3192
3193
3194
3195
3196
3197
3198
3199
3200
3201
3202
3203
3204
3205
3206
3207
3208
3209
3210
3211
3212
3213
3214
3215
3216
3217
3218
3219
3220
3221
3222
3223
3224
3225
3226
3227
3228
3229
3230
3231
3232
3233
3234
3235
3236
3237
3238
3239
3240
3241
3242
3243
3244
3245
3246
3247
3248
3249
3250
3251
3252
3253
3254
3255
3256
3257
3258
3259
3260
3261
3262
3263
3264
3265
3266
3267
3268
3269
3270
3271
3272
3273
3274
3275
3276
3277
3278
3279
3280
3281
3282
3283
3284
3285
3286
3287
3288
3289
3290
3291
3292
3293
3294
3295
3296
3297
3298
3299
3300
3301
3302
3303
3304
3305
3306
3307
3308
3309
3310
3311
3312
3313
3314
3315
3316
3317
3318
3319
3320
3321
3322
3323
3324
3325
3326
3327
3328
3329
3330
3331
3332
3333
3334
3335
3336
3337
3338
3339
3340
3341
3342
3343
3344
3345
3346
3347
3348
3349
3350
3351
3352
3353
3354
3355
3356
3357
3358
3359
3360
3361
3362
3363
3364
3365
3366
3367
3368
3369
3370
3371
3372
3373
3374
3375
3376
3377
3378
3379
3380
3381
3382
3383
3384
3385
3386
3387
3388
3389
3390
3391
3392
3393
3394
3395
3396
3397
3398
3399
3400
3401
3402
3403
3404
3405
3406
3407
3408
3409
3410
3411
3412
3413
3414
3415
3416
3417
3418
3419
3420
3421
3422
3423
3424
3425
3426
3427
3428
3429
3430
3431
3432
3433
3434
3435
3436
3437
3438
3439
3440
3441
3442
3443
3444
3445
3446
3447
3448
3449
3450
3451
3452
3453
3454
3455
3456
3457
3458
3459
3460
3461
3462
3463
3464
3465
3466
3467
3468
3469
3470
3471
3472
3473
3474
3475
3476
3477
3478
3479
3480
3481
3482
3483
3484
3485
3486
3487
3488
3489
3490
3491
3492
3493
3494
3495
3496
3497
3498
3499
3500
3501
3502
3503
3504
3505
3506
3507
3508
3509
3510
3511
3512
3513
3514
3515
3516
3517
3518
3519
3520
3521
3522
3523
3524
3525
3526
3527
3528
3529
3530
3531
3532
3533
3534
3535
3536
3537
3538
3539
3540
3541
3542
3543
3544
3545
3546
3547
3548
3549
3550
3551
3552
3553
3554
3555
3556
3557
3558
3559
3560
3561
3562
3563
3564
3565
3566
3567
3568
3569
3570
3571
3572
3573
3574
3575
3576
3577
3578
3579
3580
3581
3582
3583
3584
3585
3586
3587
3588
3589
3590
3591
3592
3593
3594
3595
3596
3597
3598
3599
3600
3601
3602
3603
3604
3605
3606
3607
3608
3609
3610
3611
3612
3613
3614
3615
3616
3617
3618
3619
3620
3621
3622
3623
3624
3625
3626
3627
3628
3629
3630
3631
3632
3633
3634
3635
3636
3637
3638
3639
3640
3641
3642
3643
3644
3645
3646
3647
3648
3649
3650
3651
3652
3653
3654
3655
3656
3657
3658
3659
3660
3661
3662
3663
3664
3665
3666
3667
3668
3669
3670
3671
3672
3673
3674
3675
3676
% \iffalse meta-comment
%
%% File: l3tl.dtx
%
% Copyright (C) 1990-2021 The LaTeX Project
%
% It may be distributed and/or modified under the conditions of the
% LaTeX Project Public License (LPPL), either version 1.3c of this
% license or (at your option) any later version.  The latest version
% of this license is in the file
%
%    https://www.latex-project.org/lppl.txt
%
% This file is part of the "l3kernel bundle" (The Work in LPPL)
% and all files in that bundle must be distributed together.
%
% -----------------------------------------------------------------------
%
% The development version of the bundle can be found at
%
%    https://github.com/latex3/latex3
%
% for those people who are interested.
%
%<*driver>
\documentclass[full,kernel]{l3doc}
\begin{document}
  \DocInput{\jobname.dtx}
\end{document}
%</driver>
% \fi
%
% \title{^^A
%   The \pkg{l3tl} package\\ Token lists^^A
% }
%
% \author{^^A
%  The \LaTeX{} Project\thanks
%    {^^A
%      E-mail:
%        \href{mailto:latex-team@latex-project.org}
%          {latex-team@latex-project.org}^^A
%    }^^A
% }
%
% \date{Released 2021-05-11}
%
% \maketitle
%
% \begin{documentation}
%
% \TeX{} works with tokens, and \LaTeX3 therefore provides a number of
% functions to deal with lists of tokens.  Token lists may be present
% directly in the argument to a function:
% \begin{verbatim}
%   \foo:n { a collection of \tokens }
% \end{verbatim}
% or may be stored in a so-called \enquote{token list variable}, which
% have the suffix \texttt{tl}: a token list variable can also be used as
% the argument to a function, for example
% \begin{verbatim}
%   \foo:N \l_some_tl
% \end{verbatim}
% In both cases, functions are available to test and manipulate the lists
% of tokens, and these have the module prefix \texttt{tl}.
% In many cases, functions which can be applied to token list variables
% are paired with similar functions for application to explicit lists
% of tokens: the two \enquote{views} of a token list are therefore collected
% together here.
%
% A token list (explicit, or stored in a variable) can be seen either
% as a list of \enquote{items},
% or a list of \enquote{tokens}. An item is whatever \cs{use:n} would
% grab as its argument: a single non-space token or a brace group,
% with optional leading explicit space characters (each item is thus
% itself a token list). A token is either a normal \texttt{N} argument,
% or \verb*| |, |{|, or |}| (assuming normal \TeX{} category codes).
% Thus for example
% \begin{verbatim}
%   { Hello } ~ world
% \end{verbatim}
% contains six items (\texttt{Hello}, \texttt{w}, \texttt{o}, \texttt{r},
% \texttt{l} and \texttt{d}), but thirteen tokens (|{|, \texttt{H}, \texttt{e},
% \texttt{l}, \texttt{l}, \texttt{o}, |}|, \verb*| |, \texttt{w}, \texttt{o},
% \texttt{r}, \texttt{l} and \texttt{d}).
% Functions which act on items are often faster than their analogue acting
% directly on tokens.
%
% \section{Creating and initialising token list variables}
%
% \begin{function}{\tl_new:N, \tl_new:c}
%   \begin{syntax}
%     \cs{tl_new:N} \meta{tl~var}
%   \end{syntax}
%   Creates a new \meta{tl~var} or raises an error if the
%   name is already taken. The declaration is global. The
%   \meta{tl~var} is initially empty.
% \end{function}
%
% \begin{function}{\tl_const:Nn, \tl_const:Nx, \tl_const:cn, \tl_const:cx}
%   \begin{syntax}
%     \cs{tl_const:Nn} \meta{tl~var} \Arg{token list}
%   \end{syntax}
%   Creates a new constant \meta{tl~var} or raises an error
%   if the name is already taken. The value of the
%   \meta{tl~var} is set globally to the \meta{token list}.
% \end{function}
%
% \begin{function}{\tl_clear:N, \tl_clear:c, \tl_gclear:N, \tl_gclear:c}
%   \begin{syntax}
%     \cs{tl_clear:N} \meta{tl~var}
%   \end{syntax}
%   Clears all entries from the \meta{tl~var}.
% \end{function}
%
% \begin{function}
%   {\tl_clear_new:N, \tl_clear_new:c, \tl_gclear_new:N, \tl_gclear_new:c}
%   \begin{syntax}
%     \cs{tl_clear_new:N} \meta{tl~var}
%   \end{syntax}
%   Ensures that the \meta{tl~var} exists globally by applying
%   \cs{tl_new:N} if necessary, then applies \cs[index=tl_clear:N]{tl_(g)clear:N} to leave
%   the \meta{tl~var} empty.
% \end{function}
%
% \begin{function}
%   {
%     \tl_set_eq:NN,  \tl_set_eq:cN,  \tl_set_eq:Nc,  \tl_set_eq:cc,
%     \tl_gset_eq:NN, \tl_gset_eq:cN, \tl_gset_eq:Nc, \tl_gset_eq:cc
%   }
%   \begin{syntax}
%     \cs{tl_set_eq:NN} \meta{tl~var_1} \meta{tl~var_2}
%   \end{syntax}
%   Sets the content of \meta{tl~var_1} equal to that of
%   \meta{tl~var_2}.
% \end{function}
%
% \begin{function}[added = 2012-05-18]
%   {
%     \tl_concat:NNN,  \tl_concat:ccc,
%     \tl_gconcat:NNN, \tl_gconcat:ccc
%   }
%   \begin{syntax}
%     \cs{tl_concat:NNN} \meta{tl~var_1} \meta{tl~var_2} \meta{tl~var_3}
%   \end{syntax}
%   Concatenates the content of \meta{tl~var_2} and \meta{tl~var_3}
%   together and saves the result in \meta{tl~var_1}. The \meta{tl~var_2}
%   is placed at the left side of the new token list.
% \end{function}
%
% \begin{function}[EXP, pTF, added=2012-03-03]{\tl_if_exist:N, \tl_if_exist:c}
%   \begin{syntax}
%     \cs{tl_if_exist_p:N} \meta{tl~var}
%     \cs{tl_if_exist:NTF} \meta{tl~var} \Arg{true code} \Arg{false code}
%   \end{syntax}
%   Tests whether the \meta{tl~var} is currently defined.  This does not
%   check that the \meta{tl~var} really is a token list variable.
% \end{function}
%
% \section{Adding data to token list variables}
%
% \begin{function}
%   {
%     \tl_set:Nn, \tl_set:NV, \tl_set:Nv, \tl_set:No, \tl_set:Nf, \tl_set:Nx,
%     \tl_set:cn, \tl_set:cV, \tl_set:cv, \tl_set:co, \tl_set:cf, \tl_set:cx,
%     \tl_gset:Nn, \tl_gset:NV, \tl_gset:Nv,
%     \tl_gset:No, \tl_gset:Nf, \tl_gset:Nx,
%     \tl_gset:cn, \tl_gset:cV, \tl_gset:cv,
%     \tl_gset:co, \tl_gset:cf, \tl_gset:cx
%   }
%   \begin{syntax}
%     \cs{tl_set:Nn} \meta{tl~var} \Arg{tokens}
%   \end{syntax}
%   Sets \meta{tl~var} to contain \meta{tokens},
%   removing any previous content from the variable.
% \end{function}
%
% \begin{function}
%   {
%     \tl_put_left:Nn,  \tl_put_left:NV,  \tl_put_left:No,  \tl_put_left:Nx,
%     \tl_put_left:cn,  \tl_put_left:cV,  \tl_put_left:co,  \tl_put_left:cx,
%     \tl_gput_left:Nn, \tl_gput_left:NV, \tl_gput_left:No, \tl_gput_left:Nx,
%     \tl_gput_left:cn, \tl_gput_left:cV, \tl_gput_left:co, \tl_gput_left:cx
%   }
%   \begin{syntax}
%     \cs{tl_put_left:Nn} \meta{tl~var} \Arg{tokens}
%   \end{syntax}
%   Appends \meta{tokens} to the left side of the current content of
%   \meta{tl~var}.
% \end{function}
%
% \begin{function}
%   {
%     \tl_put_right:Nn, \tl_put_right:NV, \tl_put_right:No, \tl_put_right:Nx,
%     \tl_put_right:cn, \tl_put_right:cV, \tl_put_right:co, \tl_put_right:cx,
%     \tl_gput_right:Nn, \tl_gput_right:NV, \tl_gput_right:No,
%     \tl_gput_right:Nx,
%     \tl_gput_right:cn, \tl_gput_right:cV, \tl_gput_right:co,
%     \tl_gput_right:cx
%   }
%   \begin{syntax}
%     \cs{tl_put_right:Nn} \meta{tl~var} \Arg{tokens}
%   \end{syntax}
%   Appends \meta{tokens} to the right side of the current content of
%   \meta{tl~var}.
% \end{function}
%
% \section{Modifying token list variables}
%
% \begin{function}[updated = 2011-08-11]
%   {
%     \tl_replace_once:Nnn,  \tl_replace_once:cnn,
%     \tl_greplace_once:Nnn, \tl_greplace_once:cnn
%   }
%   \begin{syntax}
%     \cs{tl_replace_once:Nnn} \meta{tl~var} \Arg{old tokens} \Arg{new tokens}
%   \end{syntax}
%   Replaces the first (leftmost) occurrence of \meta{old tokens} in the
%   \meta{tl~var} with \meta{new tokens}. \meta{Old tokens}
%   cannot contain |{|, |}| or |#|
%   (more precisely, explicit character tokens with category code $1$
%   (begin-group) or $2$ (end-group), and tokens with category code $6$).
% \end{function}
%
% \begin{function}[updated = 2011-08-11]
%   {
%     \tl_replace_all:Nnn, \tl_replace_all:cnn,
%     \tl_greplace_all:Nnn, \tl_greplace_all:cnn
%   }
%   \begin{syntax}
%     \cs{tl_replace_all:Nnn} \meta{tl~var} \Arg{old tokens} \Arg{new tokens}
%   \end{syntax}
%   Replaces all occurrences of \meta{old tokens} in the
%   \meta{tl~var} with \meta{new tokens}. \meta{Old tokens}
%   cannot contain |{|, |}| or |#|
%   (more precisely, explicit character tokens with category code $1$
%   (begin-group) or $2$ (end-group), and tokens with category code $6$).
%   As this function
%   operates from left to right, the pattern \meta{old tokens}
%   may remain after the replacement (see \cs{tl_remove_all:Nn}
%   for an example).
% \end{function}
%
% \begin{function}[updated = 2011-08-11]
%   {
%     \tl_remove_once:Nn,  \tl_remove_once:cn,
%     \tl_gremove_once:Nn, \tl_gremove_once:cn
%   }
%   \begin{syntax}
%     \cs{tl_remove_once:Nn} \meta{tl~var} \Arg{tokens}
%   \end{syntax}
%   Removes the first (leftmost) occurrence of \meta{tokens} from the
%   \meta{tl~var}. \meta{Tokens} cannot contain |{|, |}| or |#|
%   (more precisely, explicit character tokens with category code $1$
%   (begin-group) or $2$ (end-group), and tokens with category code $6$).
% \end{function}
%
% \begin{function}[updated = 2011-08-11]
%   {
%     \tl_remove_all:Nn,  \tl_remove_all:cn,
%     \tl_gremove_all:Nn, \tl_gremove_all:cn
%   }
%   \begin{syntax}
%     \cs{tl_remove_all:Nn} \meta{tl~var} \Arg{tokens}
%   \end{syntax}
%   Removes all occurrences of \meta{tokens} from the
%   \meta{tl~var}. \meta{Tokens} cannot contain |{|, |}| or |#|
%   (more precisely, explicit character tokens with category code $1$
%   (begin-group) or $2$ (end-group), and tokens with category code $6$).
%   As this function
%   operates from left to right, the pattern \meta{tokens}
%   may remain after the removal, for instance,
%   \begin{quote}
%     \cs{tl_set:Nn} \cs{l_tmpa_tl} |{abbccd}|
%     \cs{tl_remove_all:Nn} \cs{l_tmpa_tl} |{bc}|
%   \end{quote}
%   results in \cs{l_tmpa_tl} containing \texttt{abcd}.
% \end{function}
%
% \section{Reassigning token list category codes}
%
% These functions allow the rescanning of tokens: re-apply \TeX{}'s
% tokenization process to apply category codes different from those
% in force when the tokens were absorbed. Whilst this functionality is
% supported, it is often preferable to find alternative approaches
% to achieving outcomes rather than rescanning tokens (for example
% construction of token lists token-by-token with intervening category
% code changes or using \cs{char_generate:nn}).
%
% \begin{function}[updated = 2015-08-11]
%   {
%     \tl_set_rescan:Nnn,  \tl_set_rescan:Nno,  \tl_set_rescan:Nnx,
%     \tl_set_rescan:cnn,  \tl_set_rescan:cno,  \tl_set_rescan:cnx,
%     \tl_gset_rescan:Nnn, \tl_gset_rescan:Nno, \tl_gset_rescan:Nnx,
%     \tl_gset_rescan:cnn, \tl_gset_rescan:cno, \tl_gset_rescan:cnx
%   }
%   \begin{syntax}
%     \cs{tl_set_rescan:Nnn} \meta{tl~var} \Arg{setup} \Arg{tokens}
%   \end{syntax}
%   Sets \meta{tl~var} to contain \meta{tokens}, applying the category
%   code r\'{e}gime specified in the \meta{setup} before carrying out
%   the assignment. (Category codes applied to tokens not explicitly covered
%   by the \meta{setup} are those in force at the point of use of
%   \cs{tl_set_rescan:Nnn}.)
%   This allows the \meta{tl~var} to contain material
%   with category codes other than those that apply when \meta{tokens}
%   are absorbed.  The \meta{setup} is run within a group and may
%   contain any valid input, although only changes in category codes,
%   such as uses of \cs{cctab_select:N},
%   are relevant. See also \cs{tl_rescan:nn}.
%   \begin{texnote}
%     The \meta{tokens} are first turned into a string (using
%     \cs{tl_to_str:n}).  If the string contains one or more characters
%     with character code \tn{newlinechar} (set equal to
%     \tn{endlinechar} unless that is equal to $32$, before the user
%     \meta{setup}), then it is split into lines at these characters,
%     then read as if reading multiple lines from a file, ignoring
%     spaces (catcode $10$) at the beginning and spaces and tabs
%     (character code $32$ or $9$) at the end of every line.
%     Otherwise, spaces (and tabs) are retained at both ends of the
%     single-line string, as if it appeared in the middle of a line
%     read from a file.
%   \end{texnote}
% \end{function}
%
% \begin{function}[updated = 2015-08-11]{\tl_rescan:nn}
%   \begin{syntax}
%     \cs{tl_rescan:nn} \Arg{setup} \Arg{tokens}
%   \end{syntax}
%   Rescans \meta{tokens} applying the category code r\'{e}gime
%   specified in the \meta{setup}, and leaves the resulting tokens in
%   the input stream. (Category codes applied to tokens not explicitly covered
%   by the \meta{setup} are those in force at the point of use of
%   \cs{tl_rescan:nn}.)
%   The \meta{setup} is run within a group and may
%   contain any valid input, although only changes in category codes,
%   such as uses of \cs{cctab_select:N},
%   are relevant.  See also \cs{tl_set_rescan:Nnn}, which is more
%   robust than using \cs{tl_set:Nn} in the \meta{tokens} argument of
%   \cs{tl_rescan:nn}.
%   \begin{texnote}
%     The \meta{tokens} are first turned into a string (using
%     \cs{tl_to_str:n}).  If the string contains one or more characters
%     with character code \tn{newlinechar} (set equal to
%     \tn{endlinechar} unless that is equal to $32$, before the user
%     \meta{setup}), then it is split into lines at these characters,
%     then read as if reading multiple lines from a file, ignoring
%     spaces (catcode $10$) at the beginning and spaces and tabs
%     (character code $32$ or $9$) at the end of every line.
%     Otherwise, spaces (and tabs) are retained at both ends of the
%     single-line string, as if it appeared in the middle of a line
%     read from a file.
%   \end{texnote}
% \end{function}
%
% \section{Token list conditionals}
%
% \begin{function}[EXP,pTF, updated = 2019-09-04]
%   {\tl_if_blank:n, \tl_if_blank:e, \tl_if_blank:V, \tl_if_blank:o}
%   \begin{syntax}
%     \cs{tl_if_blank_p:n} \Arg{token list}
%     \cs{tl_if_blank:nTF} \Arg{token list} \Arg{true code} \Arg{false code}
%   \end{syntax}
%   Tests if the \meta{token list} consists only of blank spaces
%   (\emph{i.e.}~contains no item). The test is \texttt{true} if
%   \meta{token list} is zero or more explicit space characters
%   (explicit tokens with character code~$32$ and category code~$10$),
%   and is \texttt{false} otherwise.
% \end{function}
%
% \begin{function}[EXP,pTF]{\tl_if_empty:N, \tl_if_empty:c}
%   \begin{syntax}
%     \cs{tl_if_empty_p:N} \meta{tl~var}
%     \cs{tl_if_empty:NTF} \meta{tl~var} \Arg{true code} \Arg{false code}
%   \end{syntax}
%   Tests if the \meta{token list variable} is entirely empty
%   (\emph{i.e.}~contains no tokens at all).
% \end{function}
%
% \begin{function}[added = 2012-05-24, updated = 2012-06-05, EXP,pTF]
%   {\tl_if_empty:n, \tl_if_empty:V, \tl_if_empty:o}
%   \begin{syntax}
%     \cs{tl_if_empty_p:n} \Arg{token list}
%     \cs{tl_if_empty:nTF} \Arg{token list} \Arg{true code} \Arg{false code}
%   \end{syntax}
%   Tests if the \meta{token list} is entirely empty
%   (\emph{i.e.}~contains no tokens at all).
% \end{function}
%
% \begin{function}[EXP,pTF]
%   {\tl_if_eq:NN, \tl_if_eq:Nc, \tl_if_eq:cN, \tl_if_eq:cc}
%   \begin{syntax}
%     \cs{tl_if_eq_p:NN} \meta{tl~var_1} \meta{tl~var_2}
%     \cs{tl_if_eq:NNTF} \meta{tl~var_1} \meta{tl~var_2} \Arg{true code} \Arg{false code}
%   \end{syntax}
%   Compares the content of two \meta{token list variables} and
%   is logically \texttt{true} if the two contain the same list of
%   tokens (\emph{i.e.}~identical in both the list of characters they
%   contain and the category codes of those characters). Thus for example
%   \begin{verbatim}
%     \tl_set:Nn \l_tmpa_tl { abc }
%     \tl_set:Nx \l_tmpb_tl { \tl_to_str:n { abc } }
%     \tl_if_eq:NNTF \l_tmpa_tl \l_tmpb_tl { true } { false }
%   \end{verbatim}
%   yields \texttt{false}.
%   See also \cs{str_if_eq:nnTF} for a comparison that ignores category codes.
% \end{function}
%
% \begin{function}[TF, added = 2020-07-14]{\tl_if_eq:Nn, \tl_if_eq:cn}
%   \begin{syntax}
%     \cs{tl_if_eq:NnTF} \meta{tl~var_1} \Arg{token list_2} \Arg{true code} \Arg{false code}
%   \end{syntax}
%   Tests if the \meta{token list variable_1} and the \meta{token
%   list_2} contain the same list of tokens, both in respect of
%   character codes and category codes.  This conditional is not
%   expandable: see \cs{tl_if_eq:NNTF} for an expandable version when
%   both token lists are stored in variables, or \cs{str_if_eq:nnTF} if
%   category codes are not important.
% \end{function}
%
% \begin{function}[TF]{\tl_if_eq:nn}
%   \begin{syntax}
%     \cs{tl_if_eq:nnTF} \Arg{token list_1} \Arg{token list_2} \Arg{true code} \Arg{false code}
%   \end{syntax}
%   Tests if \meta{token list_1} and \meta{token list_2} contain the
%   same list of tokens, both in respect of character codes and category
%   codes.  This conditional is not expandable: see \cs{tl_if_eq:NNTF}
%   for an expandable version when token lists are stored in variables,
%   or \cs{str_if_eq:nnTF} if category codes are not important.
% \end{function}
%
% \begin{function}[TF]{\tl_if_in:Nn, \tl_if_in:cn}
%   \begin{syntax}
%     \cs{tl_if_in:NnTF} \meta{tl~var} \Arg{token list} \Arg{true code} \Arg{false code}
%   \end{syntax}
%   Tests if the \meta{token list} is found in the content of the
%   \meta{tl~var}. The \meta{token list} cannot contain
%   the tokens |{|, |}| or |#|
%   (more precisely, explicit character tokens with category code $1$
%   (begin-group) or $2$ (end-group), and tokens with category code $6$).
% \end{function}
%
% \begin{function}[TF]
%   {\tl_if_in:nn, \tl_if_in:Vn, \tl_if_in:on, \tl_if_in:no}
%   \begin{syntax}
%     \cs{tl_if_in:nnTF} \Arg{token list_1} \Arg{token list_2} \Arg{true code} \Arg{false code}
%   \end{syntax}
%   Tests if \meta{token list_2} is found inside \meta{token list_1}.
%   The \meta{token list_2} cannot contain the tokens |{|, |}| or |#|
%   (more precisely, explicit character tokens with category code $1$
%   (begin-group) or $2$ (end-group), and tokens with category code $6$).
% \end{function}
%
% \begin{function}[added = 2017-11-14, EXP,pTF]{\tl_if_novalue:n}
%   \begin{syntax}
%     \cs{tl_if_novalue_p:n} \Arg{token list}
%     \cs{tl_if_novalue:nTF} \Arg{token list} \Arg{true code} \Arg{false code}
%   \end{syntax}
%   Tests if the \meta{token list} is exactly equal to the special
%   \cs{c_novalue_tl} marker. This function is intended to allow construction
%   of flexible document interface structures in which missing optional
%   arguments are detected.
% \end{function}
%
% \begin{function}[updated = 2011-08-13, EXP,pTF]
%   {\tl_if_single:N, \tl_if_single:c}
%   \begin{syntax}
%     \cs{tl_if_single_p:N} \meta{tl~var}
%     \cs{tl_if_single:NTF} \meta{tl~var} \Arg{true code} \Arg{false code}
%   \end{syntax}
%   Tests if the content of the \meta{tl~var} consists of a single item,
%   \emph{i.e.}~is a single normal token (neither an explicit space
%   character nor a begin-group character) or a single brace group,
%   surrounded by optional spaces on both sides. In other words, such a
%   token list has token count $1$ according to \cs{tl_count:N}.
% \end{function}
%
% \begin{function}[updated = 2011-08-13, EXP,pTF]{\tl_if_single:n}
%   \begin{syntax}
%     \cs{tl_if_single_p:n} \Arg{token list}
%     \cs{tl_if_single:nTF} \Arg{token list} \Arg{true code} \Arg{false code}
%   \end{syntax}
%   Tests if the \meta{token list} has exactly one item, \emph{i.e.}~is
%   a single normal token (neither an explicit space character nor a
%   begin-group character) or a single brace group, surrounded by
%   optional spaces on both sides. In other words, such a token list has
%   token count $1$ according to \cs{tl_count:n}.
% \end{function}
%
% \begin{function}[EXP,pTF]{\tl_if_single_token:n}
%   \begin{syntax}
%   \cs{tl_if_single_token_p:n} \Arg{token list}
%   \cs{tl_if_single_token:nTF} \Arg{token list} \Arg{true code} \Arg{false code}
%   \end{syntax}
%   Tests if the token list consists of exactly one token, \emph{i.e.}~is
%   either a single space character or a single \enquote{normal} token.
%   Token groups (|{|\ldots|}|) are not single tokens.
% \end{function}
%
% \begin{function}[added = 2013-07-24, EXP, noTF]{\tl_case:Nn, \tl_case:cn}
%   \begin{syntax}
%     \cs{tl_case:NnTF} \meta{test token list variable} \\
%     ~~"{" \\
%     ~~~~\meta{token list variable case_1} \Arg{code case_1} \\
%     ~~~~\meta{token list variable case_2} \Arg{code case_2} \\
%     ~~~~\ldots \\
%     ~~~~\meta{token list variable case_n} \Arg{code case_n} \\
%     ~~"}" \\
%     ~~\Arg{true code}
%     ~~\Arg{false code}
%   \end{syntax}
%   This function compares the \meta{test token list variable} in turn
%   with each of the \meta{token list variable cases}. If the two
%   are equal (as described for \cs{tl_if_eq:NNTF})
%   then the associated \meta{code} is left in the input
%   stream and other cases are discarded. If any of the
%   cases are matched, the \meta{true code} is also inserted into the
%   input stream (after the code for the appropriate case), while if none
%   match then the \meta{false code} is inserted. The function
%   \cs{tl_case:Nn}, which does nothing if there is no match, is also
%   available.
% \end{function}
%
% \section{Mapping over token lists}
%
% All mappings are done at the current group level, \emph{i.e.}~any
% local assignments made by the \meta{function} or \meta{code} discussed
% below remain in effect after the loop.
%
% \begin{function}[updated = 2012-06-29, rEXP]
%   {\tl_map_function:NN, \tl_map_function:cN}
%   \begin{syntax}
%     \cs{tl_map_function:NN} \meta{tl~var} \meta{function}
%   \end{syntax}
%   Applies \meta{function} to every \meta{item} in the \meta{tl~var}.
%   The \meta{function} receives one argument for each iteration.
%   This may be a number of tokens if the \meta{item} was stored within
%   braces. Hence the \meta{function} should anticipate receiving
%   \texttt{n}-type arguments. See also \cs{tl_map_function:nN}.
% \end{function}
%
% \begin{function}[updated = 2012-06-29, rEXP]{\tl_map_function:nN}
%   \begin{syntax}
%     \cs{tl_map_function:nN} \Arg{token list} \meta{function}
%   \end{syntax}
%   Applies \meta{function} to every \meta{item} in the \meta{token list},
%   The \meta{function} receives one argument for each iteration.
%   This may be a number of tokens if the \meta{item} was stored within
%   braces. Hence the \meta{function} should anticipate receiving
%   \texttt{n}-type arguments. See also \cs{tl_map_function:NN}.
% \end{function}
%
% \begin{function}[updated = 2012-06-29]
%   {\tl_map_inline:Nn, \tl_map_inline:cn}
%   \begin{syntax}
%     \cs{tl_map_inline:Nn} \meta{tl~var} \Arg{inline function}
%   \end{syntax}
%   Applies the \meta{inline function} to every \meta{item} stored within the
%   \meta{tl~var}. The \meta{inline function} should consist of code which
%   receives the \meta{item} as |#1|. See also \cs{tl_map_function:NN}.
% \end{function}
%
% \begin{function}[updated = 2012-06-29]{\tl_map_inline:nn}
%   \begin{syntax}
%     \cs{tl_map_inline:nn} \Arg{token list} \Arg{inline function}
%   \end{syntax}
%   Applies the \meta{inline function} to every \meta{item} stored within the
%   \meta{token list}. The \meta{inline function}  should consist of code which
%   receives the \meta{item} as |#1|. See also \cs{tl_map_function:nN}.
% \end{function}
%
% \begin{function}[rEXP, added = 2019-09-02]
%   {\tl_map_tokens:Nn, \tl_map_tokens:cn, \tl_map_tokens:nn}
%   \begin{syntax}
%     \cs{tl_map_tokens:Nn} \meta{tl~var} \Arg{code}
%     \cs{tl_map_tokens:nn} \Arg{tokens} \Arg{code}
%   \end{syntax}
%   Analogue of \cs{tl_map_function:NN} which maps several tokens
%   instead of a single function.  The \meta{code} receives each item in
%   the \meta{tl~var} or in \meta{tokens} as a trailing brace group. For
%   instance,
%   \begin{verbatim}
%     \tl_map_tokens:Nn \l_my_tl { \prg_replicate:nn { 2 } }
%   \end{verbatim}
%   expands to twice each item in the \meta{tl~var}: for each item in
%   \cs[no-index]{l_my_tl} the function \cs{prg_replicate:nn} receives |2| and
%   \meta{item} as its two arguments.  The function
%   \cs{tl_map_inline:Nn} is typically faster but is not expandable.
% \end{function}
%
% \begin{function}[updated = 2012-06-29]
%   {\tl_map_variable:NNn, \tl_map_variable:cNn}
%   \begin{syntax}
%     \cs{tl_map_variable:NNn} \meta{tl~var} \meta{variable} \Arg{code}
%   \end{syntax}
%   Stores each \meta{item} of the \meta{tl~var} in turn in the (token
%   list) \meta{variable} and applies the \meta{code}.  The \meta{code}
%   will usually make use of the \meta{variable}, but this is not
%   enforced.  The assignments to the \meta{variable} are local.  Its
%   value after the loop is the last \meta{item} in the \meta{tl~var},
%   or its original value if the \meta{tl~var} is blank.  See also
%   \cs{tl_map_inline:Nn}.
% \end{function}
%
% \begin{function}[updated = 2012-06-29]{\tl_map_variable:nNn}
%   \begin{syntax}
%     \cs{tl_map_variable:nNn} \Arg{token list} \meta{variable} \Arg{code}
%   \end{syntax}
%   Stores each \meta{item} of the \meta{token list} in turn in the
%   (token list) \meta{variable} and applies the \meta{code}.  The
%   \meta{code} will usually make use of the \meta{variable}, but this
%   is not enforced.  The assignments to the \meta{variable} are local.
%   Its value after the loop is the last \meta{item} in the
%   \meta{tl~var}, or its original value if the \meta{tl~var} is blank.
%   See also \cs{tl_map_inline:nn}.
% \end{function}
%
% \begin{function}[updated = 2012-06-29, rEXP]{\tl_map_break:}
%   \begin{syntax}
%     \cs{tl_map_break:}
%   \end{syntax}
%   Used to terminate a \cs[no-index]{tl_map_\ldots} function before all
%   entries in the \meta{token list variable} have been processed. This
%   normally takes place within a conditional statement, for example
%   \begin{verbatim}
%     \tl_map_inline:Nn \l_my_tl
%       {
%         \str_if_eq:nnT { #1 } { bingo } { \tl_map_break: }
%         % Do something useful
%       }
%   \end{verbatim}
%   See also \cs{tl_map_break:n}.
%   Use outside of a \cs[no-index]{tl_map_\ldots} scenario leads to low
%   level \TeX{} errors.
%   \begin{texnote}
%     When the mapping is broken, additional tokens may be inserted
%     before the \meta{tokens} are
%     inserted into the input stream.
%     This depends on the design of the mapping function.
%   \end{texnote}
% \end{function}
%
% \begin{function}[updated = 2012-06-29, rEXP]{\tl_map_break:n}
%   \begin{syntax}
%     \cs{tl_map_break:n} \Arg{code}
%   \end{syntax}
%   Used to terminate a \cs[no-index]{tl_map_\ldots} function before all
%   entries in the \meta{token list variable} have been processed, inserting
%   the \meta{code} after the mapping has ended. This
%   normally takes place within a conditional statement, for example
%   \begin{verbatim}
%     \tl_map_inline:Nn \l_my_tl
%       {
%         \str_if_eq:nnT { #1 } { bingo }
%           { \tl_map_break:n { <code> } }
%         % Do something useful
%       }
%   \end{verbatim}
%   Use outside of a \cs[no-index]{tl_map_\ldots} scenario leads to low
%   level \TeX{} errors.
%   \begin{texnote}
%     When the mapping is broken, additional tokens may be inserted
%     before the \meta{code} is
%     inserted into the input stream.
%     This depends on the design of the mapping function.
%   \end{texnote}
% \end{function}
%
% \section{Using token lists}
%
% \begin{function}[EXP]{\tl_to_str:n, \tl_to_str:V}
%   \begin{syntax}
%     \cs{tl_to_str:n} \Arg{token list}
%   \end{syntax}
%   Converts the \meta{token list} to a \meta{string}, leaving the
%   resulting character tokens in the input stream. A \meta{string}
%   is a series of tokens with category code $12$ (other) with the exception
%   of spaces, which retain category code $10$ (space).
%   This function requires only a single expansion.
%   Its argument \emph{must} be braced.
%   \begin{texnote}
%     This is the \eTeX{} primitive \tn{detokenize}.
%     Converting a \meta{token list} to a \meta{string} yields a
%     concatenation of the string representations of every token in the
%     \meta{token list}.
%     The string representation of a control sequence is
%     \begin{itemize}
%       \item an escape character, whose character code is given by the
%         internal parameter \tn{escapechar}, absent if the
%         \tn{escapechar} is negative or greater than the largest
%         character code;
%       \item the control sequence name, as defined by \cs{cs_to_str:N};
%       \item a space, unless the control sequence name is a single
%         character whose category at the time of expansion of
%         \cs{tl_to_str:n} is not \enquote{letter}.
%     \end{itemize}
%     The string representation of an explicit character token is that
%     character, doubled in the case of (explicit) macro parameter
%     characters (normally |#|).
%     In particular, the string representation of a token list may
%     depend on the category codes in effect when it is evaluated, and
%     the value of the \tn{escapechar}: for instance |\tl_to_str:n {\a}|
%     normally produces the three character \enquote{backslash},
%     \enquote{lower-case a}, \enquote{space}, but it may also produce a
%     single \enquote{lower-case a} if the escape character is negative
%     and \texttt{a} is currently not a letter.
%   \end{texnote}
% \end{function}
%
% \begin{function}[EXP]{\tl_to_str:N, \tl_to_str:c}
%   \begin{syntax}
%     \cs{tl_to_str:N} \meta{tl~var}
%   \end{syntax}
%   Converts the content of the  \meta{tl~var} into a series of characters
%   with category code $12$ (other) with the exception of spaces, which
%   retain category code $10$ (space). This \meta{string} is then left
%   in the input stream. For low-level details, see the notes given for
%   \cs{tl_to_str:n}.
% \end{function}
%
% \begin{function}[EXP]{\tl_use:N, \tl_use:c}
%   \begin{syntax}
%     \cs{tl_use:N} \meta{tl~var}
%   \end{syntax}
%   Recovers the content of a \meta{tl~var} and places it
%   directly in the input stream. An error is raised if the variable
%   does not exist or if it is invalid. Note that it is possible to use
%   a \meta{tl~var} directly without an accessor function.
% \end{function}
%
% \section{Working with the content of token lists}
%
% \begin{function}[added = 2012-05-13, EXP]
%   {\tl_count:n, \tl_count:V, \tl_count:o}
%   \begin{syntax}
%     \cs{tl_count:n} \Arg{tokens}
%   \end{syntax}
%   Counts the number of \meta{items} in \meta{tokens} and leaves this
%   information in the input stream. Unbraced tokens count as one
%   element as do each token group (|{|\ldots|}|). This process
%   ignores any unprotected spaces within \meta{tokens}. See also
%   \cs{tl_count:N}. This function requires three expansions,
%   giving an \meta{integer denotation}.
% \end{function}
%
% \begin{function}[added = 2012-05-13, EXP]{\tl_count:N, \tl_count:c}
%   \begin{syntax}
%     \cs{tl_count:N} \meta{tl~var}
%   \end{syntax}
%   Counts the number of token groups in the \meta{tl~var}
%   and leaves this information in the input stream. Unbraced tokens
%   count as one element as do each token group (|{|\ldots|}|). This
%   process ignores any unprotected spaces within the \meta{tl~var}.
%   See also \cs{tl_count:n}. This function requires three expansions,
%   giving an \meta{integer denotation}.
% \end{function}
%
% \begin{function}[EXP, added = 2019-02-25]{\tl_count_tokens:n}
%   \begin{syntax}
%     \cs{tl_count_tokens:n} \Arg{tokens}
%   \end{syntax}
%   Counts the number of \TeX{} tokens in the \meta{tokens} and leaves
%   this information in the input stream. Every token, including spaces and
%   braces, contributes one to the total; thus for instance, the token count of
%   |a~{bc}| is $6$.
% \end{function}
%
% \begin{function}[updated = 2012-01-08, EXP]
%   {\tl_reverse:n, \tl_reverse:V, \tl_reverse:o}
%   \begin{syntax}
%     \cs{tl_reverse:n} \Arg{token list}
%   \end{syntax}
%   Reverses the order of the \meta{items} in the \meta{token list},
%   so that \meta{item_1}\meta{item_2}\meta{item_3} \ldots \meta{item_n}
%   becomes \meta{item_n}\ldots \meta{item_3}\meta{item_2}\meta{item_1}.
%   This process preserves unprotected space within the
%   \meta{token list}. Tokens are not reversed within braced token
%   groups, which keep their outer set of braces.
%   In situations where performance is important,
%   consider \cs{tl_reverse_items:n}.
%   See also \cs{tl_reverse:N}.
%   \begin{texnote}
%     The result is returned within \tn{unexpanded}, which means that the token
%     list does not expand further when appearing in an \texttt{x}-type
%     argument expansion.
%   \end{texnote}
% \end{function}
%
% \begin{function}[updated = 2012-01-08]
%   {\tl_reverse:N, \tl_reverse:c, \tl_greverse:N, \tl_greverse:c}
%   \begin{syntax}
%     \cs{tl_reverse:N} \meta{tl~var}
%   \end{syntax}
%   Reverses the order of the \meta{items} stored in \meta{tl~var}, so
%   that \meta{item_1}\meta{item_2}\meta{item_3} \ldots \meta{item_n}
%   becomes \meta{item_n}\ldots \meta{item_3}\meta{item_2}\meta{item_1}.
%   This process preserves unprotected spaces within the
%   \meta{token list variable}. Braced token groups are copied without
%   reversing the order of tokens, but keep the outer set of braces.
%   See also \cs{tl_reverse:n}, and, for improved performance,
%   \cs{tl_reverse_items:n}.
% \end{function}
%
% \begin{function}[added = 2012-01-08, EXP]{\tl_reverse_items:n}
%   \begin{syntax}
%     \cs{tl_reverse_items:n} \Arg{token list}
%   \end{syntax}
%   Reverses the order of the \meta{items} stored in \meta{tl~var},
%   so that \Arg{item_1}\Arg{item_2}\Arg{item_3} \ldots \Arg{item_n}
%   becomes \Arg{item_n} \ldots{} \Arg{item_3}\Arg{item_2}\Arg{item_1}.
%   This process removes any unprotected space within the
%   \meta{token list}. Braced token groups are copied without
%   reversing the order of tokens, and keep the outer set of braces.
%   Items which are initially not braced are copied with braces in
%   the result. In cases where preserving spaces is important,
%   consider the slower function \cs{tl_reverse:n}.
%   \begin{texnote}
%     The result is returned within \tn{unexpanded}, which means that the token
%     list does not expand further when appearing in an \texttt{x}-type
%     argument expansion.
%   \end{texnote}
% \end{function}
%
% \begin{function}[added = 2011-07-09, updated = 2012-06-25, EXP]
%   {\tl_trim_spaces:n, \tl_trim_spaces:o}
%   \begin{syntax}
%     \cs{tl_trim_spaces:n} \Arg{token list}
%   \end{syntax}
%   Removes any leading and trailing explicit space characters
%   (explicit tokens with character code~$32$ and category code~$10$)
%   from the \meta{token list} and leaves the result in the input
%   stream.
%   \begin{texnote}
%     The result is returned within \tn{unexpanded}, which means that the token
%     list does not expand further when appearing in an \texttt{x}-type
%     argument expansion.
%   \end{texnote}
% \end{function}
%
% \begin{function}[added = 2018-04-12, EXP]
%   {\tl_trim_spaces_apply:nN, \tl_trim_spaces_apply:oN}
%   \begin{syntax}
%     \cs{tl_trim_spaces_apply:nN} \Arg{token list} \meta{function}
%   \end{syntax}
%   Removes any leading and trailing explicit space characters (explicit
%   tokens with character code~$32$ and category code~$10$) from the
%   \meta{token list} and passes the result to the \meta{function} as an
%   \texttt{n}-type argument.
% \end{function}
%
% \begin{function}[added = 2011-07-09]
%   {
%     \tl_trim_spaces:N,  \tl_trim_spaces:c,
%     \tl_gtrim_spaces:N, \tl_gtrim_spaces:c
%   }
%   \begin{syntax}
%     \cs{tl_trim_spaces:N} \meta{tl~var}
%   \end{syntax}
%   Removes any leading and trailing explicit space characters
%   (explicit tokens with character code~$32$ and category code~$10$)
%   from the content of the \meta{tl~var}. Note that this therefore
%   \emph{resets} the content of the variable.
% \end{function}
%
% \begin{function}[added = 2017-02-06]
%   {\tl_sort:Nn, \tl_sort:cn, \tl_gsort:Nn, \tl_gsort:cn}
%   \begin{syntax}
%     \cs{tl_sort:Nn} \meta{tl var} \Arg{comparison code}
%   \end{syntax}
%   Sorts the items in the \meta{tl var} according to the
%   \meta{comparison code}, and assigns the result to
%   \meta{tl var}. The details of sorting comparison are
%   described in Section~\ref{sec:l3sort:mech}.
% \end{function}
%
% \begin{function}[added = 2017-02-06, EXP]{\tl_sort:nN}
%   \begin{syntax}
%     \cs{tl_sort:nN} \Arg{token list} \meta{conditional}
%   \end{syntax}
%   Sorts the items in the \meta{token list}, using the
%   \meta{conditional} to compare items, and leaves the result in the
%   input stream.  The \meta{conditional} should have signature |:nnTF|,
%   and return \texttt{true} if the two items being compared should be
%   left in the same order, and \texttt{false} if the items should be
%   swapped. The details of sorting comparison are
%   described in Section~\ref{sec:l3sort:mech}.
%   \begin{texnote}
%     The result is returned within \cs{exp_not:n}, which means that the
%     token list does not expand further when appearing in an
%     \texttt{x}-type or \texttt{e}-type argument expansion.
%   \end{texnote}
% \end{function}
%
% \section{The first token from a token list}
%
% Functions which deal with either only the very first item (balanced
% text or single normal token) in a token list, or the remaining tokens.
%
% \begin{function}[updated = 2012-09-09, EXP]
%   {\tl_head:N, \tl_head:n, \tl_head:V, \tl_head:v, \tl_head:f}
%   \begin{syntax}
%     \cs{tl_head:n} \Arg{token list}
%   \end{syntax}
%   Leaves in the input stream the first \meta{item} in the
%   \meta{token list}, discarding the rest of the \meta{token list}.
%   All leading explicit space characters
%   (explicit tokens with character code~$32$ and category code~$10$)
%   are discarded; for example
%   \begin{verbatim}
%     \tl_head:n { abc }
%   \end{verbatim}
%   and
%   \begin{verbatim}
%     \tl_head:n { ~ abc }
%   \end{verbatim}
%   both leave |a| in the input stream. If the \enquote{head} is a
%   brace group, rather than a single token, the braces are removed, and
%   so
%   \begin{verbatim}
%     \tl_head:n { ~ { ~ ab } c }
%   \end{verbatim}
%   yields \verb*| ab|.
%   A blank \meta{token list} (see \cs{tl_if_blank:nTF}) results in
%   \cs{tl_head:n} leaving nothing in the input stream.
%   \begin{texnote}
%     The result is returned within \cs{exp_not:n}, which means that the token
%     list does not expand further when appearing in an \texttt{x}-type
%     argument expansion.
%   \end{texnote}
% \end{function}
%
% \begin{function}[EXP]{\tl_head:w}
%   \begin{syntax}
%     \cs{tl_head:w} \meta{token list} | { } | \cs{q_stop}
%   \end{syntax}
%   Leaves in the input stream the first \meta{item} in the
%   \meta{token list}, discarding the rest of the \meta{token list}.
%   All leading explicit space characters
%   (explicit tokens with character code~$32$ and category code~$10$)
%   are discarded.
%   A blank \meta{token list} (which consists only of space characters)
%   results in a low-level \TeX{} error, which may be avoided by the
%   inclusion of an empty group in the input (as shown), without the need
%   for an explicit test. Alternatively, \cs{tl_if_blank:nF} may be used to
%   avoid using the function with a \enquote{blank} argument.
%   This function requires only a single expansion, and thus is suitable for
%   use within an \texttt{o}-type expansion. In general, \cs{tl_head:n} should
%   be preferred if the number of expansions is not critical.
% \end{function}
%
% \begin{function}[updated = 2012-09-01, EXP]
%   {\tl_tail:N, \tl_tail:n, \tl_tail:V, \tl_tail:v, \tl_tail:f}
%   \begin{syntax}
%     \cs{tl_tail:n} \Arg{token list}
%   \end{syntax}
%   Discards all leading explicit space characters
%   (explicit tokens with character code~$32$ and category code~$10$)
%   and the first \meta{item} in the \meta{token list}, and leaves the
%   remaining tokens in the input stream.  Thus for example
%   \begin{verbatim}
%     \tl_tail:n { a ~ {bc} d }
%   \end{verbatim}
%   and
%   \begin{verbatim}
%     \tl_tail:n { ~ a ~ {bc} d }
%   \end{verbatim}
%   both leave \verb*| {bc}d| in the input stream.  A blank
%   \meta{token list} (see \cs{tl_if_blank:nTF}) results
%   in \cs{tl_tail:n} leaving nothing in the input stream.
%   \begin{texnote}
%     The result is returned within \cs{exp_not:n}, which means that the
%     token list does not expand further when appearing in an \texttt{x}-type
%     argument expansion.
%   \end{texnote}
% \end{function}
%
% \begin{function}[updated = 2012-07-09, EXP, pTF]
%   {\tl_if_head_eq_catcode:nN, \tl_if_head_eq_catcode:oN}
%   \begin{syntax}
%     \cs{tl_if_head_eq_catcode_p:nN} \Arg{token list} \meta{test token}
%     \cs{tl_if_head_eq_catcode:nNTF} \Arg{token list} \meta{test token}
%     ~~\Arg{true code} \Arg{false code}
%   \end{syntax}
%   Tests if the first \meta{token} in the \meta{token list} has the
%   same category code as the \meta{test token}.  In the case where the
%   \meta{token list} is empty, the test is always \texttt{false}.
% \end{function}
%
% \begin{function}[updated = 2012-07-09, EXP, pTF]
%   {\tl_if_head_eq_charcode:nN, \tl_if_head_eq_charcode:fN}
%   \begin{syntax}
%     \cs{tl_if_head_eq_charcode_p:nN} \Arg{token list} \meta{test token}
%     \cs{tl_if_head_eq_charcode:nNTF} \Arg{token list} \meta{test token}
%     ~~\Arg{true code} \Arg{false code}
%   \end{syntax}
%   Tests if the first \meta{token} in the \meta{token list} has the
%   same character code as the \meta{test token}.  In the case where the
%   \meta{token list} is empty, the test is always \texttt{false}.
% \end{function}
%
% \begin{function}[updated = 2012-07-09, EXP, pTF]{\tl_if_head_eq_meaning:nN}
%   \begin{syntax}
%     \cs{tl_if_head_eq_meaning_p:nN} \Arg{token list} \meta{test token}
%     \cs{tl_if_head_eq_meaning:nNTF} \Arg{token list} \meta{test token}
%     ~~\Arg{true code} \Arg{false code}
%   \end{syntax}
%   Tests if the first \meta{token} in the \meta{token list} has the
%   same meaning as the \meta{test token}.  In the case where
%   \meta{token list} is empty, the test is always \texttt{false}.
% \end{function}
%
% \begin{function}[added = 2012-07-08, EXP, pTF]{\tl_if_head_is_group:n}
%   \begin{syntax}
%     \cs{tl_if_head_is_group_p:n} \Arg{token list}
%     \cs{tl_if_head_is_group:nTF} \Arg{token list} \Arg{true code} \Arg{false code}
%   \end{syntax}
%   Tests if the first \meta{token} in the \meta{token list}
%   is an explicit begin-group character (with category code~$1$
%   and any character code), in other words, if the \meta{token list}
%   starts with a brace group. In particular, the test is \texttt{false}
%   if the \meta{token list} starts with an implicit token such as
%   \cs{c_group_begin_token}, or if it is empty.
%   This function is useful to implement actions on token lists on
%   a token by token basis.
% \end{function}
%
% \begin{function}[added = 2012-07-08, EXP, pTF]{\tl_if_head_is_N_type:n}
%   \begin{syntax}
%     \cs{tl_if_head_is_N_type_p:n} \Arg{token list}
%     \cs{tl_if_head_is_N_type:nTF} \Arg{token list} \Arg{true code} \Arg{false code}
%   \end{syntax}
%   Tests if the first \meta{token} in the \meta{token list}
%   is a normal \texttt{N}-type argument. In other words,
%   it is neither an explicit space character
%   (explicit token with character code~$32$ and category code~$10$)
%   nor an explicit begin-group character
%   (with category code~1 and any character code). An empty
%   argument yields \texttt{false}, as it does not have a \enquote{normal}
%   first token.
%   This function is useful to implement actions on token lists on
%   a token by token basis.
% \end{function}
%
% \begin{function}[updated = 2012-07-08, EXP, pTF]{\tl_if_head_is_space:n}
%   \begin{syntax}
%     \cs{tl_if_head_is_space_p:n} \Arg{token list}
%     \cs{tl_if_head_is_space:nTF} \Arg{token list} \Arg{true code} \Arg{false code}
%   \end{syntax}
%   Tests if the first \meta{token} in the \meta{token list}
%   is an explicit space character
%   (explicit token with character code~$12$ and category code~$10$).
%   In particular, the test is \texttt{false} if the \meta{token list}
%   starts with an implicit token such as \cs{c_space_token}, or if it
%   is empty.
%   This function is useful to implement actions on token lists on
%   a token by token basis.
% \end{function}
%
% \section{Using a single item}
%
% \begin{function}[added = 2014-07-17, EXP]
%   {\tl_item:nn, \tl_item:Nn, \tl_item:cn}
%   \begin{syntax}
%     \cs{tl_item:nn} \Arg{token list} \Arg{integer expression}
%   \end{syntax}
%   Indexing items in the \meta{token list} from~$1$ on the left, this
%   function evaluates the \meta{integer expression} and leaves the
%   appropriate item from the \meta{token list} in the input stream.
%   If the \meta{integer expression} is negative, indexing occurs from
%   the right of the token list, starting at $-1$ for the right-most item.
%   If the index is out of bounds, then the function expands to nothing.
%   \begin{texnote}
%     The result is returned within the \tn{unexpanded}
%     primitive (\cs{exp_not:n}), which means that the \meta{item}
%     does not expand further when appearing in an \texttt{x}-type
%     argument expansion.
%   \end{texnote}
% \end{function}
%
% \begin{function}[EXP, added = 2016-12-06]
%   {\tl_rand_item:N, \tl_rand_item:c, \tl_rand_item:n}
%   \begin{syntax}
%     \cs{tl_rand_item:N} \meta{tl~var}
%     \cs{tl_rand_item:n} \Arg{token list}
%   \end{syntax}
%   Selects a pseudo-random item of the \meta{token list}.  If the
%   \meta{token list} is blank, the result is empty.
%   This is not available in older versions of \XeTeX{}.
%   \begin{texnote}
%     The result is returned within the \tn{unexpanded}
%     primitive (\cs{exp_not:n}), which means that the \meta{item}
%     does not expand further when appearing in an \texttt{x}-type
%     argument expansion.
%   \end{texnote}
% \end{function}
%
% \begin{function}[EXP, added = 2017-02-17, updated = 2017-07-15]
%   {\tl_range:Nnn, \tl_range:nnn}
%   \begin{syntax}
%     \cs{tl_range:Nnn} \meta{tl~var} \Arg{start index} \Arg{end index}
%     \cs{tl_range:nnn} \Arg{token list} \Arg{start index} \Arg{end index}
%   \end{syntax}
%   Leaves in the input stream the items from the \meta{start index} to the
%   \meta{end index} inclusive.  Spaces and braces are preserved between
%   the items returned (but never at either end of the list).
%   Here \meta{start index} and \meta{end index} should be \meta{integer expressions}.
%   For describing in detail the functions' behavior, let $m$ and $n$ be the start
%   and end index respectively. If either is $0$, the result is empty. A positive
%   index means `start counting from the left end', and a negative index means
%   `from the right end'. Let $l$ be the count of the token list.
%
%   The \emph{actual start point} is determined as $M=m$ if~$m>0$ and as $M=l+m+1$
%   if~$m<0$. Similarly the \emph{actual end point} is $N=n$ if~$n>0$ and $N=l+n+1$
%   if~$n<0$. If $M>N$, the result is empty. Otherwise it consists of all items from
%   position $M$ to position $N$ inclusive; for the purpose of this rule, we can
%   imagine that the token list extends at infinity on either side, with void items
%   at positions $s$ for $s\le0$ or $s>l$.
%
%   Spaces in between items in the actual range are preserved. Spaces at either end
%   of the token list will be removed anyway (think to the token list being passed to
%   |\tl_trim_spaces:n| to begin with.
%
%   Thus, with $l=7$ as in the examples below, all of the following are equivalent
%   and result in the whole token list
%   \begin{verbatim}
%     \tl_range:nnn { abcd~{e{}}fg } { 1 } { 7 }
%     \tl_range:nnn { abcd~{e{}}fg } { 1 } { 12 }
%     \tl_range:nnn { abcd~{e{}}fg } { -7 } { 7 }
%     \tl_range:nnn { abcd~{e{}}fg } { -12 } { 7 }
%   \end{verbatim}
%   Here are some more interesting examples. The calls
%   \begin{verbatim}
%     \iow_term:x { \tl_range:nnn { abcd{e{}}fg } { 2 } { 5 } }
%     \iow_term:x { \tl_range:nnn { abcd{e{}}fg } { 2 } { -3 } }
%     \iow_term:x { \tl_range:nnn { abcd{e{}}fg } { -6 } { 5 } }
%     \iow_term:x { \tl_range:nnn { abcd{e{}}fg } { -6 } { -3 } }
%   \end{verbatim}
%   are all equivalent and will print |bcd{e{}}| on the terminal; similarly
%   \begin{verbatim}
%     \iow_term:x { \tl_range:nnn { abcd~{e{}}fg } { 2 } { 5 } }
%     \iow_term:x { \tl_range:nnn { abcd~{e{}}fg } { 2 } { -3 } }
%     \iow_term:x { \tl_range:nnn { abcd~{e{}}fg } { -6 } { 5 } }
%     \iow_term:x { \tl_range:nnn { abcd~{e{}}fg } { -6 } { -3 } }
%   \end{verbatim}
%   are all equivalent and will print |bcd {e{}}| on the
%   terminal (note the space in the middle). To the contrary,
%   \begin{verbatim}
%     \tl_range:nnn { abcd~{e{}}f } { 2 } { 4 }
%   \end{verbatim}
%   will discard the space after `d'.
%
%   If we want to get the items from, say, the third to the last in a token
%   list |<tl>|, the call
%   is |\tl_range:nnn { <tl> } { 3 } { -1 }|. Similarly, for discarding
%   the last item, we can do |\tl_range:nnn { <tl> } { 1 } { -2 }|.
%
%^^A   The behavior of \cs{tl_range:Nnn} is exactly the same, acting on the
%^^A   contents of the tl variable.
%
%   For better performance, see  \cs{tl_range_braced:nnn} and
%   \cs{tl_range_unbraced:nnn}.
%   \begin{texnote}
%     The result is returned within the \tn{unexpanded}
%     primitive (\cs{exp_not:n}), which means that the \meta{item}
%     does not expand further when appearing in an \texttt{x}-type
%     argument expansion.
%   \end{texnote}
% \end{function}
%
% \section{Viewing token lists}
%
% \begin{function}[updated = 2021-04-29]{\tl_show:N, \tl_show:c}
%   \begin{syntax}
%     \cs{tl_show:N} \meta{tl~var}
%   \end{syntax}
%   Displays the content of the \meta{tl~var} on the terminal.
%   \begin{texnote}
%     This is similar to the \TeX{} primitive \tn{show}, wrapped to a
%     fixed number of characters per line.
%   \end{texnote}
% \end{function}
%
% \begin{function}[updated = 2015-08-07]{\tl_show:n}
%   \begin{syntax}
%     \cs{tl_show:n} \Arg{token list}
%   \end{syntax}
%   Displays the \meta{token list} on the terminal.
%   \begin{texnote}
%     This is similar to the \eTeX{} primitive \tn{showtokens}, wrapped
%     to a fixed number of characters per line.
%   \end{texnote}
% \end{function}
%
% \begin{function}[added = 2014-08-22, updated = 2021-04-29]{\tl_log:N, \tl_log:c}
%   \begin{syntax}
%     \cs{tl_log:N} \meta{tl~var}
%   \end{syntax}
%   Writes the content of the \meta{tl~var} in the log file.  See also
%   \cs{tl_show:N} which displays the result in the terminal.
% \end{function}
%
% \begin{function}[added = 2014-08-22, updated = 2015-08-07]{\tl_log:n}
%   \begin{syntax}
%     \cs{tl_log:n} \Arg{token list}
%   \end{syntax}
%   Writes the \meta{token list} in the log file.  See also
%   \cs{tl_show:n} which displays the result in the terminal.
% \end{function}
%
% \section{Constant token lists}
%
% \begin{variable}{\c_empty_tl}
%   Constant that is always empty.
% \end{variable}
%
% \begin{variable}[added = 2017-11-14]{\c_novalue_tl}
%   A marker for the absence of an argument. This constant |tl| can safely
%   be typeset (\emph{cf.}~\cs{q_nil}), with the result being |-NoValue-|.
%   It is important to note that \cs{c_novalue_tl} is constructed such that it
%   will \emph{not} match the simple text input |-NoValue-|, \emph{i.e.}
%   that
%   \begin{verbatim}
%     \tl_if_eq:NnTF \c_novalue_tl { -NoValue- }
%   \end{verbatim}
%   is logically \texttt{false}. The \cs{c_novalue_tl} marker is intended for
%   use in creating document-level interfaces, where it serves as an indicator
%   that an (optional) argument was omitted. In particular, it is distinct
%   from a simple empty |tl|.
% \end{variable}
%
% \begin{variable}{\c_space_tl}
%   An explicit space character contained in a token list (compare this with
%   \cs{c_space_token}). For use where an explicit space is required.
% \end{variable}
%
% \section{Scratch token lists}
%
% \begin{variable}{\l_tmpa_tl, \l_tmpb_tl}
%   Scratch token lists for local assignment. These are never used by
%   the kernel code, and so are safe for use with any \LaTeX3-defined
%   function. However, they may be overwritten by other non-kernel
%   code and so should only be used for short-term storage.
% \end{variable}
%
% \begin{variable}{\g_tmpa_tl, \g_tmpb_tl}
%   Scratch token lists for global assignment. These are never used by
%   the kernel code, and so are safe for use with any \LaTeX3-defined
%   function. However, they may be overwritten by other non-kernel
%   code and so should only be used for short-term storage.
% \end{variable}
%
% \end{documentation}
%
% \begin{implementation}
%
% \section{\pkg{l3tl} implementation}
%
%    \begin{macrocode}
%<*package>
%    \end{macrocode}
%
%    \begin{macrocode}
%<@@=tl>
%    \end{macrocode}
%
% A token list variable is a \TeX{} macro that holds tokens. By using the
% \eTeX{} primitive \tn{unexpanded} inside a \TeX{} \tn{edef} it is
% possible to store any tokens, including |#|, in this way.
%
% \subsection{Functions}
%
% \begin{macro}{\__kernel_tl_set:Nx,\__kernel_tl_gset:Nx}
%   These two are supplied to get better performance for macros which would
%   otherwise use \cs{tl_set:Nx} or \cs{tl_gset:Nx} internally.
%    \begin{macrocode}
\cs_new_eq:NN \__kernel_tl_set:Nx  \cs_set_nopar:Npx
\cs_new_eq:NN \__kernel_tl_gset:Nx \cs_gset_nopar:Npx
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\tl_new:N, \tl_new:c}
%   Creating new token list variables is a case of checking for an
%   existing definition and doing the definition.
%    \begin{macrocode}
\cs_new_protected:Npn \tl_new:N #1
  {
    \__kernel_chk_if_free_cs:N #1
    \cs_gset_eq:NN #1 \c_empty_tl
  }
\cs_generate_variant:Nn \tl_new:N { c }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\tl_const:Nn, \tl_const:Nx, \tl_const:cn, \tl_const:cx}
%   Constants are also easy to generate. They use \cs{cs_gset_nopar:Npx} instead
%   of \cs{__kernel_tl_gset:Nx} so that the correct scope checking is applied if
%   \pkg{l3debug} is used.
%    \begin{macrocode}
\cs_new_protected:Npn \tl_const:Nn #1#2
  {
    \__kernel_chk_if_free_cs:N #1
    \cs_gset_nopar:Npx #1 { \__kernel_exp_not:w {#2} }
  }
\cs_new_protected:Npn \tl_const:Nx #1#2
  {
    \__kernel_chk_if_free_cs:N #1
    \cs_gset_nopar:Npx #1 {#2}
  }
\cs_generate_variant:Nn \tl_const:Nn { c }
\cs_generate_variant:Nn \tl_const:Nx { c }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\tl_clear:N, \tl_clear:c}
% \begin{macro}{\tl_gclear:N, \tl_gclear:c}
%   Clearing a token list variable means setting it to an empty value.
%   Error checking is sorted out by the parent function.
%    \begin{macrocode}
\cs_new_protected:Npn \tl_clear:N  #1
  { \tl_set_eq:NN #1 \c_empty_tl }
\cs_new_protected:Npn \tl_gclear:N #1
  { \tl_gset_eq:NN #1 \c_empty_tl }
\cs_generate_variant:Nn \tl_clear:N  { c }
\cs_generate_variant:Nn \tl_gclear:N { c }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}{\tl_clear_new:N, \tl_clear_new:c}
% \begin{macro}{\tl_gclear_new:N, \tl_gclear_new:c}
%   Clearing a token list variable means setting it to an empty value.
%   Error checking is sorted out by the parent function.
%    \begin{macrocode}
\cs_new_protected:Npn \tl_clear_new:N  #1
  { \tl_if_exist:NTF #1 { \tl_clear:N #1 } { \tl_new:N #1 } }
\cs_new_protected:Npn \tl_gclear_new:N #1
  { \tl_if_exist:NTF #1 { \tl_gclear:N #1 } { \tl_new:N #1 } }
\cs_generate_variant:Nn \tl_clear_new:N  { c }
\cs_generate_variant:Nn \tl_gclear_new:N { c }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}{\tl_set_eq:NN, \tl_set_eq:Nc, \tl_set_eq:cN, \tl_set_eq:cc}
% \begin{macro}{\tl_gset_eq:NN, \tl_gset_eq:Nc, \tl_gset_eq:cN, \tl_gset_eq:cc}
%   For setting token list variables equal to each other.  To allow for
%   patching, the arguments have to be explicit.
%    \begin{macrocode}
\cs_new_protected:Npn \tl_set_eq:NN  #1#2 { \cs_set_eq:NN #1 #2 }
\cs_new_protected:Npn \tl_gset_eq:NN #1#2 { \cs_gset_eq:NN #1 #2 }
\cs_generate_variant:Nn \tl_set_eq:NN { cN, Nc, cc }
\cs_generate_variant:Nn \tl_gset_eq:NN { cN, Nc, cc }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}{\tl_concat:NNN, \tl_concat:ccc}
% \begin{macro}{\tl_gconcat:NNN, \tl_gconcat:ccc}
%   Concatenating token lists is easy.  When checking is turned on, all
%   three arguments must be checked: a token list |#2| or |#3| equal to
%   \cs{scan_stop:} would lead to problems later on.
%    \begin{macrocode}
\cs_new_protected:Npn \tl_concat:NNN #1#2#3
  {
    \__kernel_tl_set:Nx #1
      {
        \__kernel_exp_not:w \exp_after:wN {#2}
        \__kernel_exp_not:w \exp_after:wN {#3}
      }
  }
\cs_new_protected:Npn \tl_gconcat:NNN #1#2#3
  {
    \__kernel_tl_gset:Nx #1
      {
        \__kernel_exp_not:w \exp_after:wN {#2}
        \__kernel_exp_not:w \exp_after:wN {#3}
      }
  }
\cs_generate_variant:Nn \tl_concat:NNN  { ccc }
\cs_generate_variant:Nn \tl_gconcat:NNN { ccc }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}[pTF]{\tl_if_exist:N, \tl_if_exist:c}
%   Copies of the \texttt{cs} functions defined in \pkg{l3basics}.
%    \begin{macrocode}
\prg_new_eq_conditional:NNn \tl_if_exist:N \cs_if_exist:N { TF , T , F , p }
\prg_new_eq_conditional:NNn \tl_if_exist:c \cs_if_exist:c { TF , T , F , p }
%    \end{macrocode}
% \end{macro}
%
% \subsection{Constant token lists}
%
% \begin{variable}{\c_empty_tl}
%   Never full. We need to define that constant before using \cs{tl_new:N}.
%    \begin{macrocode}
\tl_const:Nn \c_empty_tl { }
%    \end{macrocode}
% \end{variable}
%
% \begin{variable}{\c_novalue_tl}
%   A special marker: as we don't have |\char_generate:nn| yet, has to be
%   created the old-fashioned way.
%    \begin{macrocode}
\group_begin:
\tex_lccode:D `A = `-
\tex_lccode:D `N = `N
\tex_lccode:D `V = `V
\tex_lowercase:D
  {
    \group_end:
    \tl_const:Nn \c_novalue_tl { ANoValue- }
  }
%    \end{macrocode}
% \end{variable}
%
% \begin{variable}{\c_space_tl}
%   A space as a token list (as opposed to as a character).
%    \begin{macrocode}
\tl_const:Nn \c_space_tl { ~ }
%    \end{macrocode}
% \end{variable}
%
% \subsection{Adding to token list variables}
%
% \begin{macro}
%   {
%     \tl_set:Nn, \tl_set:NV, \tl_set:Nv, \tl_set:No, \tl_set:Nf, \tl_set:Nx,
%     \tl_set:cn, \tl_set:cV, \tl_set:cv, \tl_set:co, \tl_set:cf, \tl_set:cx
%   }
% \begin{macro}
%   {
%     \tl_gset:Nn, \tl_gset:NV, \tl_gset:Nv,
%     \tl_gset:No, \tl_gset:Nf, \tl_gset:Nx,
%     \tl_gset:cn, \tl_gset:cV, \tl_gset:cv,
%     \tl_gset:co, \tl_gset:cf, \tl_gset:cx
%   }
%   By using \cs{exp_not:n} token list variables can contain |#| tokens,
%   which makes the token list registers provided by \TeX{}
%   more or less redundant. The \cs{tl_set:No} version is done
%   \enquote{by hand} as it is used quite a lot.
%    \begin{macrocode}
\cs_new_protected:Npn \tl_set:Nn #1#2
  { \__kernel_tl_set:Nx #1 { \__kernel_exp_not:w {#2} } }
\cs_new_protected:Npn \tl_set:No #1#2
  { \__kernel_tl_set:Nx #1 { \__kernel_exp_not:w \exp_after:wN {#2} } }
\cs_new_protected:Npn \tl_set:Nx #1#2
  { \__kernel_tl_set:Nx #1 {#2} }
\cs_new_protected:Npn \tl_gset:Nn #1#2
  { \__kernel_tl_gset:Nx #1 { \__kernel_exp_not:w {#2} } }
\cs_new_protected:Npn \tl_gset:No #1#2
  { \__kernel_tl_gset:Nx #1 { \__kernel_exp_not:w \exp_after:wN {#2} } }
\cs_new_protected:Npn \tl_gset:Nx #1#2
  { \__kernel_tl_gset:Nx #1 {#2} }
\cs_generate_variant:Nn \tl_set:Nn  {    NV , Nv , Nf }
\cs_generate_variant:Nn \tl_set:Nn  { c, cV , cv , cf }
\cs_generate_variant:Nn \tl_set:Nx  { c }
\cs_generate_variant:Nn \tl_set:No  { c }
\cs_generate_variant:Nn \tl_gset:Nn {    NV , Nv , Nf }
\cs_generate_variant:Nn \tl_gset:Nn { c, cV , cv , cf }
\cs_generate_variant:Nn \tl_gset:Nx { c }
\cs_generate_variant:Nn \tl_gset:No { c }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}
%   {
%     \tl_put_left:Nn, \tl_put_left:NV, \tl_put_left:No, \tl_put_left:Nx,
%     \tl_put_left:cn, \tl_put_left:cV, \tl_put_left:co, \tl_put_left:cx
%   }
% \begin{macro}
%   {
%     \tl_gput_left:Nn, \tl_gput_left:NV, \tl_gput_left:No, \tl_gput_left:Nx,
%     \tl_gput_left:cn, \tl_gput_left:cV, \tl_gput_left:co, \tl_gput_left:cx
%   }
% Adding to the left is done directly to gain a little performance.
%    \begin{macrocode}
\cs_new_protected:Npn \tl_put_left:Nn #1#2
  {
    \__kernel_tl_set:Nx #1
      { \__kernel_exp_not:w {#2} \__kernel_exp_not:w \exp_after:wN {#1} }
  }
\cs_new_protected:Npn \tl_put_left:NV #1#2
  {
    \__kernel_tl_set:Nx #1
      { \exp_not:V #2 \__kernel_exp_not:w \exp_after:wN {#1} }
  }
\cs_new_protected:Npn \tl_put_left:No #1#2
  {
    \__kernel_tl_set:Nx #1
      {
        \__kernel_exp_not:w \exp_after:wN {#2}
        \__kernel_exp_not:w \exp_after:wN {#1}
      }
  }
\cs_new_protected:Npn \tl_put_left:Nx #1#2
  { \__kernel_tl_set:Nx #1 { #2 \__kernel_exp_not:w \exp_after:wN {#1} } }
\cs_new_protected:Npn \tl_gput_left:Nn #1#2
  {
    \__kernel_tl_gset:Nx #1
      { \__kernel_exp_not:w {#2} \__kernel_exp_not:w \exp_after:wN {#1} }
  }
\cs_new_protected:Npn \tl_gput_left:NV #1#2
  {
    \__kernel_tl_gset:Nx #1
      { \exp_not:V #2 \__kernel_exp_not:w \exp_after:wN {#1} }
  }
\cs_new_protected:Npn \tl_gput_left:No #1#2
  {
    \__kernel_tl_gset:Nx #1
      {
        \__kernel_exp_not:w \exp_after:wN {#2}
        \__kernel_exp_not:w \exp_after:wN {#1}
      }
  }
\cs_new_protected:Npn \tl_gput_left:Nx #1#2
  { \__kernel_tl_gset:Nx #1 { #2 \__kernel_exp_not:w \exp_after:wN {#1} } }
\cs_generate_variant:Nn \tl_put_left:Nn  { c }
\cs_generate_variant:Nn \tl_put_left:NV  { c }
\cs_generate_variant:Nn \tl_put_left:No  { c }
\cs_generate_variant:Nn \tl_put_left:Nx  { c }
\cs_generate_variant:Nn \tl_gput_left:Nn { c }
\cs_generate_variant:Nn \tl_gput_left:NV { c }
\cs_generate_variant:Nn \tl_gput_left:No { c }
\cs_generate_variant:Nn \tl_gput_left:Nx { c }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}
%   {
%     \tl_put_right:Nn, \tl_put_right:NV, \tl_put_right:No, \tl_put_right:Nx,
%     \tl_put_right:cn, \tl_put_right:cV, \tl_put_right:co, \tl_put_right:cx
%   }
% \begin{macro}
%   {
%     \tl_gput_right:Nn, \tl_gput_right:NV, \tl_gput_right:No,
%     \tl_gput_right:Nx,
%     \tl_gput_right:cn, \tl_gput_right:cV, \tl_gput_right:co,
%     \tl_gput_right:cx
%   }
% The same on the right.
%    \begin{macrocode}
\cs_new_protected:Npn \tl_put_right:Nn #1#2
  { \__kernel_tl_set:Nx #1 { \__kernel_exp_not:w \exp_after:wN { #1 #2 } } }
\cs_new_protected:Npn \tl_put_right:NV #1#2
  {
    \__kernel_tl_set:Nx #1
      { \__kernel_exp_not:w \exp_after:wN {#1} \exp_not:V #2 }
  }
\cs_new_protected:Npn \tl_put_right:No #1#2
  {
    \__kernel_tl_set:Nx #1
      {
        \__kernel_exp_not:w \exp_after:wN {#1}
        \__kernel_exp_not:w \exp_after:wN {#2}
      }
  }
\cs_new_protected:Npn \tl_put_right:Nx #1#2
  { \__kernel_tl_set:Nx #1 { \__kernel_exp_not:w \exp_after:wN {#1} #2 } }
\cs_new_protected:Npn \tl_gput_right:Nn #1#2
  { \__kernel_tl_gset:Nx #1 { \__kernel_exp_not:w \exp_after:wN { #1 #2 } } }
\cs_new_protected:Npn \tl_gput_right:NV #1#2
  {
    \__kernel_tl_gset:Nx #1
      { \__kernel_exp_not:w \exp_after:wN {#1} \exp_not:V #2 }
  }
\cs_new_protected:Npn \tl_gput_right:No #1#2
  {
    \__kernel_tl_gset:Nx #1
      {
        \__kernel_exp_not:w \exp_after:wN {#1}
        \__kernel_exp_not:w \exp_after:wN {#2}
      }
  }
\cs_new_protected:Npn \tl_gput_right:Nx #1#2
  { \__kernel_tl_gset:Nx #1 { \__kernel_exp_not:w \exp_after:wN {#1} #2 } }
\cs_generate_variant:Nn \tl_put_right:Nn  { c }
\cs_generate_variant:Nn \tl_put_right:NV  { c }
\cs_generate_variant:Nn \tl_put_right:No  { c }
\cs_generate_variant:Nn \tl_put_right:Nx  { c }
\cs_generate_variant:Nn \tl_gput_right:Nn { c }
\cs_generate_variant:Nn \tl_gput_right:NV { c }
\cs_generate_variant:Nn \tl_gput_right:No { c }
\cs_generate_variant:Nn \tl_gput_right:Nx { c }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \subsection{Internal quarks and quark-query functions}
%
% \begin{variable}{\q_@@_nil,\q_@@_mark,\q_@@_stop}
%   Internal quarks.
%    \begin{macrocode}
\quark_new:N \q_@@_nil
\quark_new:N \q_@@_mark
\quark_new:N \q_@@_stop
%    \end{macrocode}
% \end{variable}
%
% \begin{variable}{\q_@@_recursion_tail,\q_@@_recursion_stop}
%   Internal recursion quarks.
%    \begin{macrocode}
\quark_new:N \q_@@_recursion_tail
\quark_new:N \q_@@_recursion_stop
%    \end{macrocode}
% \end{variable}
%
% \begin{macro}[EXP]{\@@_if_recursion_tail_break:nN}
% \begin{macro}[pTF]{\@@_if_recursion_tail_stop:n}
%   Functions to query recursion quarks.
%    \begin{macrocode}
\__kernel_quark_new_test:N \@@_if_recursion_tail_break:nN
\__kernel_quark_new_conditional:Nn \@@_quark_if_nil:n { TF }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \subsection{Reassigning token list category codes}
%
% \begin{variable}{\c_@@_rescan_marker_tl}
%   The rescanning code needs a special token list containing the same
%   character (chosen here to be a colon) with two different category
%   codes: it cannot appear in the tokens being rescanned since all
%   colons have the same category code.
%    \begin{macrocode}
\tl_const:Nx \c_@@_rescan_marker_tl { : \token_to_str:N : }
%    \end{macrocode}
% \end{variable}
%
% \begin{macro}
%   {
%     \tl_set_rescan:Nnn, \tl_set_rescan:Nno, \tl_set_rescan:Nnx,
%     \tl_set_rescan:cnn, \tl_set_rescan:cno, \tl_set_rescan:cnx
%   }
% \begin{macro}
%   {
%     \tl_gset_rescan:Nnn, \tl_gset_rescan:Nno, \tl_gset_rescan:Nnx,
%     \tl_gset_rescan:cnn, \tl_gset_rescan:cno, \tl_gset_rescan:cnx
%   }
% \begin{macro}{\tl_rescan:nn}
% \begin{macro}{\@@_rescan_aux:}
% \begin{macro}{\@@_set_rescan:NNnn, \@@_set_rescan_multi:nNN}
% \begin{macro}[EXP]{\@@_rescan:NNw}
%   In a group, after some initial setup explained below and the user
%   setup~|#3| (followed by \cs{scan_stop:} to be safe), there is a call
%   to \cs{@@_set_rescan:nNN}.  This shared auxiliary defined later
%   distinguishes single-line and multi-line ``files''.  In the simplest
%   case of multi-line files, it calls (with the same arguments)
%   \cs{@@_set_rescan_multi:nNN}, whose code is included here to help
%   understand the approach.  This function rescans its argument |#1|,
%   closes the group, and performs the assignment.
%
%   One difficulty when rescanning is that \tn{scantokens} treats the
%   argument as a file, and without the correct settings a \TeX{} error
%   occurs:
%   \begin{verbatim}
%    ! File ended while scanning definition of ...
%   \end{verbatim}
%   A related minor issue is a warning due to opening a group before the
%   \tn{scantokens} and closing it inside that temporary file; we avoid
%   that by setting \tn{tracingnesting}.  The standard solution to the
%   ``File ended'' error is to grab the rescanned tokens as a delimited
%   argument of an auxiliary, here \cs{@@_rescan:NNw}, that performs the
%   assignment, then let \TeX{} ``execute'' the end of file marker.  As
%   usual in delimited arguments we use \cs{prg_do_nothing:} to avoid
%   stripping an outer set braces: this is removed by using
%   \texttt{o}-expanding assignments.  The delimiter cannot appear
%   within the rescanned token list because it contains twice the same
%   character, with different catcodes.
%
%   For \cs{tl_rescan:nn} we cannot simply call \cs{@@_set_rescan:NNnn}
%   \cs{prg_do_nothing:} \cs{use:n} because that would leave the
%   end-of-file marker \emph{after} the result of rescanning.  If that
%   rescanned result is code that looks further in the input stream for
%   arguments, it would break.
%
%   For multi-line files the only subtlety is that \tn{newlinechar}
%   should be equal to \tn{endlinechar} because \tn{newlinechar}
%   characters become new lines and then become \tn{endlinechar}
%   characters when writing to an abstract file and reading back.  This
%   equality is ensured by setting \tn{newlinechar} equal to
%   \tn{endlinechar}.  Prior to this, \tn{endlinechar} is set to $-1$ if
%   it was $32$ (in particular true after \cs{ExplSyntaxOn}) to avoid
%   unreasonable line-breaks at every space for instance in error
%   messages triggered by the user setup.  Another side effect of
%   reading back from the file is that spaces (catcode $10$) are ignored
%   at the beginning of lines, and spaces and tabs (character code $32$
%   and $9$) are ignored at the end of lines.
%
%   The two \cs{if_false:} \ldots{} \cs{fi:} are there to prevent
%   alignment tabs to cause a change of tabular cell while rescanning.
%   We put the \enquote{opening} one after \cs{group_begin:} so that if
%   one accidentally \texttt{f}-expands \cs{tl_set_rescan:Nnn} braces
%   remain balanced.  This is essential in \texttt{e}-type arguments
%   when \tn{expanded} is not available.
%    \begin{macrocode}
\cs_new_protected:Npn \tl_rescan:nn #1#2
  {
    \tl_set_rescan:Nnn \l_@@_internal_a_tl {#1} {#2}
    \exp_after:wN \@@_rescan_aux:
    \l_@@_internal_a_tl
  }
\exp_args:NNo \cs_new_protected:Npn \@@_rescan_aux:
  { \tl_clear:N \l_@@_internal_a_tl }
\cs_new_protected:Npn \tl_set_rescan:Nnn
  { \@@_set_rescan:NNnn \tl_set:No }
\cs_new_protected:Npn \tl_gset_rescan:Nnn
  { \@@_set_rescan:NNnn \tl_gset:No }
\cs_new_protected:Npn \@@_set_rescan:NNnn #1#2#3#4
  {
    \group_begin:
      \if_false: { \fi:
      \int_set_eq:NN \tex_tracingnesting:D \c_zero_int
      \int_compare:nNnT \tex_endlinechar:D = { 32 }
        { \int_set:Nn \tex_endlinechar:D { -1 } }
      \int_set_eq:NN \tex_newlinechar:D \tex_endlinechar:D
      #3 \scan_stop:
      \exp_args:No \@@_set_rescan:nNN { \tl_to_str:n {#4} } #1 #2
    \if_false: } \fi:
  }
\cs_new_protected:Npn \@@_set_rescan_multi:nNN #1#2#3
  {
    \tex_everyeof:D \exp_after:wN { \c_@@_rescan_marker_tl }
    \exp_after:wN \@@_rescan:NNw
    \exp_after:wN #2
    \exp_after:wN #3
    \exp_after:wN \prg_do_nothing:
    \tex_scantokens:D {#1}
  }
\exp_args:Nno \use:nn
  { \cs_new:Npn \@@_rescan:NNw #1#2#3 } \c_@@_rescan_marker_tl
  {
    \group_end:
    #1 #2 {#3}
  }
\cs_generate_variant:Nn \tl_set_rescan:Nnn  {     Nno , Nnx }
\cs_generate_variant:Nn \tl_set_rescan:Nnn  { c , cno , cnx }
\cs_generate_variant:Nn \tl_gset_rescan:Nnn {     Nno , Nnx }
\cs_generate_variant:Nn \tl_gset_rescan:Nnn { c , cno }
%    \end{macrocode}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
%
% \begin{macro}{\@@_set_rescan:nNN}
% \begin{macro}{\@@_set_rescan_single:nnNN, \@@_set_rescan_single_aux:nnnNN}
% \begin{macro}[rEXP]{\@@_set_rescan_single_aux:w}
%   The function \cs{@@_set_rescan:nNN} calls \cs{@@_set_rescan_multi:nNN} or
%   \cs{@@_set_rescan_single:nnNN} |{ ' }| depending on whether its argument
%   is a single-line fragment of code/data or is made of multiple lines
%   by testing for the presence of a \tn{newlinechar} character.  If
%   \tn{newlinechar} is out of range, the argument is assumed to be a
%   single line.
%
%   For a single line, no \tn{endlinechar} should be added, so it is
%   set to $-1$, and spaces should not be removed.
%   Trailing spaces and tabs are a difficult matter, as \TeX{} removes
%   these at a very low level.  The only way to preserve them is to
%   rescan not the argument but the argument followed by a character
%   with a reasonable category code.  Here, $11$ (letter) and $12$ (other)
%   are accepted, as these are convenient, suitable for
%   delimiting an argument, and it is very unlikely that none of the
%   ASCII characters are in one of these categories.  To avoid
%   selecting one particular character to put at the end, whose
%   category code may have been modified, there is a loop through
%   characters from |'| (ASCII $39$) to |~| (ASCII $127$).  The choice
%   of starting point was made because this is the start of a very long
%   range of characters whose standard category is letter or other,
%   thus minimizing the number of steps needed by the loop (most often
%   just a single one).  If no valid character is found (very rare),
%   fall-back on \cs{@@_set_rescan_multi:nNN}.
%
%   Otherwise, once a valid character is found (let us use |'| in this
%   explanation) run some code very similar to \cs{@@_set_rescan_multi:nNN}
%   but with |'| added at both ends of the input.  Of course, we need to
%   define the auxiliary \cs{@@_set_rescan_single:NNww} on the fly to remove
%   the additional~|'| that is just before |::| (by which we mean
%   \cs{c_@@_rescan_marker_tl}).  Note that the argument must be
%   delimited by |'| with the current catcode; this is done thanks to
%   \cs{char_generate:nn}.  Yet another issue is that the rescanned
%   token list may contain a comment character, in which case the |'| we
%   expected is not there.  We fix this as follows: rather than just
%   |::| we set \tn{everyeof} to |::|\Arg{code1} |'::|\Arg{code2}
%   \cs{s_@@_stop}.  The auxiliary \cs{@@_set_rescan_single:NNww} runs the
%   \texttt{o}-expanding assignment, expanding either \meta{code1} or
%   \meta{code2} before its the main argument~|#3|.  In the typical case
%   without comment character, \meta{code1} is expanded, removing the
%   leading~|'|.  In the rarer case with comment character, \meta{code2}
%   is expanded, calling \cs{@@_set_rescan_single_aux:w}, which removes the
%   trailing |::|\Arg{code1} and the leading~|'|.
%    \begin{macrocode}
\cs_new_protected:Npn \@@_set_rescan:nNN #1
  {
    \int_compare:nNnTF \tex_newlinechar:D < 0
      { \use_ii:nn }
      {
        \exp_args:Nnf \tl_if_in:nnTF {#1}
          { \char_generate:nn { \tex_newlinechar:D } { 12 } }
      }
        { \@@_set_rescan_multi:nNN }
        {
          \int_set:Nn \tex_endlinechar:D { -1 }
          \@@_set_rescan_single:nnNN { `' }
        }
    {#1}
  }
\cs_new_protected:Npn \@@_set_rescan_single:nnNN #1
  {
    \int_compare:nNnTF
      { \char_value_catcode:n {#1} / 2 } = 6
      {
        \exp_args:Nof \@@_set_rescan_single_aux:nnnNN
          \c_@@_rescan_marker_tl
          { \char_generate:nn {#1} { \char_value_catcode:n {#1} } }
      }
      {
        \int_compare:nNnTF {#1} < { `\~ }
          {
            \exp_args:Nf \@@_set_rescan_single:nnNN
              { \int_eval:n { #1 + 1 } }
          }
          { \@@_set_rescan_multi:nNN }
      }
  }
\cs_new_protected:Npn \@@_set_rescan_single_aux:nnnNN #1#2#3#4#5
  {
    \tex_everyeof:D
      {
        #1 \use_none:n
        #2 #1 { \exp:w \@@_set_rescan_single_aux:w }
        \s_@@_stop
      }
    \cs_set:Npn \@@_rescan:NNw ##1##2##3 #2 #1 ##4 ##5 \s_@@_stop
      {
        \group_end:
        ##1 ##2 { ##4 ##3 }
      }
    \exp_after:wN \@@_rescan:NNw
    \exp_after:wN #4
    \exp_after:wN #5
    \tex_scantokens:D { #2 #3 #2 }
  }
\exp_args:Nno \use:nn
  { \cs_new:Npn \@@_set_rescan_single_aux:w #1 }
  \c_@@_rescan_marker_tl #2
  { \use_i:nn \exp_end: #1 }
%    \end{macrocode}
% \end{macro}
% \end{macro}
% \end{macro}
%
% \subsection{Modifying token list variables}
%
% \begin{macro}{\tl_replace_all:Nnn, \tl_replace_all:cnn}
% \begin{macro}{\tl_greplace_all:Nnn, \tl_greplace_all:cnn}
% \begin{macro}{\tl_replace_once:Nnn, \tl_replace_once:cnn}
% \begin{macro}{\tl_greplace_once:Nnn, \tl_greplace_once:cnn}
%   All of the \texttt{replace} functions call \cs{@@_replace:NnNNNnn}
%   with appropriate arguments.  The first two arguments are explained
%   later.  The next controls whether the replacement function calls
%   itself (\cs{@@_replace_next:w}) or stops (\cs{@@_replace_wrap:w})
%   after the first replacement.  Next comes an \texttt{x}-type
%   assignment function \cs{tl_set:Nx} or \cs{tl_gset:Nx} for local or
%   global replacements.  Finally, the three arguments \meta{tl~var}
%   \Arg{pattern} \Arg{replacement} provided by the user.  When
%   describing the auxiliary functions below, we denote the contents of
%   the \meta{tl~var} by \meta{token list}.
%    \begin{macrocode}
\cs_new_protected:Npn \tl_replace_once:Nnn
  { \@@_replace:NnNNNnn \q_@@_mark ? \@@_replace_wrap:w \__kernel_tl_set:Nx  }
\cs_new_protected:Npn \tl_greplace_once:Nnn
  { \@@_replace:NnNNNnn \q_@@_mark ? \@@_replace_wrap:w \__kernel_tl_gset:Nx }
\cs_new_protected:Npn \tl_replace_all:Nnn
  { \@@_replace:NnNNNnn \q_@@_mark ? \@@_replace_next:w \__kernel_tl_set:Nx  }
\cs_new_protected:Npn \tl_greplace_all:Nnn
  { \@@_replace:NnNNNnn \q_@@_mark ? \@@_replace_next:w \__kernel_tl_gset:Nx }
\cs_generate_variant:Nn \tl_replace_once:Nnn  { c }
\cs_generate_variant:Nn \tl_greplace_once:Nnn { c }
\cs_generate_variant:Nn \tl_replace_all:Nnn   { c }
\cs_generate_variant:Nn \tl_greplace_all:Nnn  { c }
%    \end{macrocode}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
%
% \begin{macro}
%   {
%     \@@_replace:NnNNNnn,
%     \@@_replace_auxi:NnnNNNnn,
%     \@@_replace_auxii:nNNNnn,
%     \@@_replace_next:w,
%     \@@_replace_next_aux:w,
%     \@@_replace_wrap:w,
%   }
%   To implement the actual replacement auxiliary
%   \cs{@@_replace_auxii:nNNNnn} we need a \meta{delimiter} with
%   the following properties:
%   \begin{itemize}
%     \item all occurrences of the \meta{pattern}~|#6| in
%       \enquote{\meta{token list} \meta{delimiter}} belong to the
%       \meta{token list} and have no overlap with the \meta{delimiter},
%     \item the first occurrence of the \meta{delimiter} in
%       \enquote{\meta{token list} \meta{delimiter}} is the trailing
%       \meta{delimiter}.
%   \end{itemize}
%   We first find the building blocks for the \meta{delimiter}, namely
%   two tokens \meta{A} and~\meta{B} such that \meta{A} does not appear
%   in~|#6| and |#6| is not~\meta{B} (this condition is trivial if |#6|
%   has more than one token).  Then we consider the delimiters
%   \enquote{\meta{A}} and \enquote{\meta{A} \meta{A}$^n$ \meta{B}
%     \meta{A}$^n$ \meta{B}}, for $n\geq 1$, where $\meta{A}^n$ denotes
%   $n$~copies of \meta{A}, and we choose as our \meta{delimiter} the
%   first one which is not in the \meta{token list}.
%
%   Every delimiter in the set obeys the first condition: |#6|~does not
%   contain~\meta{A} hence cannot be overlapping with the \meta{token
%     list} and the \meta{delimiter}, and it cannot be within the
%   \meta{delimiter} since it would have to be in one of the two
%   \meta{B} hence be equal to this single token (or empty, but this is
%   an error case filtered separately).  Given the particular form of
%   these delimiters, for which no prefix is also a suffix, the second
%   condition is actually a consequence of the weaker condition that the
%   \meta{delimiter} we choose does not appear in the \meta{token list}.
%   Additionally, the set of delimiters is such that a \meta{token list}
%   of $n$~tokens can contain at most $O(n^{1/2})$ of them, hence we
%   find a \meta{delimiter} with at most $O(n^{1/2})$ tokens in a time
%   at most $O(n^{3/2})$.  Bear in mind that these upper bounds are
%   reached only in very contrived scenarios: we include the case
%   \enquote{\meta{A}} in the list of delimiters to try, so that the
%   \meta{delimiter} is simply \cs{q_@@_mark} in the most common
%   situation where neither the \meta{token list} nor the \meta{pattern}
%   contains \cs{q_@@_mark}.
%
%   Let us now ahead, optimizing for this most common case.  First, two
%   special cases: an empty \meta{pattern}~|#6| is an error, and if
%   |#1|~is absent from both the \meta{token list}~|#5| and the
%   \meta{pattern}~|#6| then we can use it as the \meta{delimiter}
%   through \cs{@@_replace_auxii:nNNNnn} |{#1}|.  Otherwise, we end up
%   calling \cs{@@_replace:NnNNNnn} repeatedly with the first two
%   arguments \cs{q_@@_mark} |{?}|, |\?| |{??}|, |\??| |{???}|, and so on,
%   until |#6|~does not contain the control sequence~|#1|, which we take
%   as our~\meta{A}.  The argument~|#2| only serves to collect~|?|
%   characters for~|#1|.  Note that the order of the tests means that
%   the first two are done every time, which is wasteful (for instance,
%   we repeatedly test for the emptyness of~|#6|).  However, this is
%   rare enough not to matter.  Finally, choose~\meta{B} to be
%   \cs{q_@@_nil} or~\cs{q_@@_stop} such that it is not equal to~|#6|.
%
%   The \cs{@@_replace_auxi:NnnNNNnn} auxiliary receives \Arg{A} and
%   |{|\meta{A}$^n$\meta{B}|}| as its arguments, initially with $n=1$.
%   If \enquote{\meta{A} \meta{A}$^n$\meta{B} \meta{A}$^n$\meta{B}} is
%   in the \meta{token list} then increase~$n$ and try again.  Once it
%   is not anymore in the \meta{token list} we take it as our
%   \meta{delimiter} and pass this to the \texttt{auxii} auxiliary.
%    \begin{macrocode}
\cs_new_protected:Npn \@@_replace:NnNNNnn #1#2#3#4#5#6#7
  {
    \tl_if_empty:nTF {#6}
      {
        \__kernel_msg_error:nnx { kernel } { empty-search-pattern }
          { \tl_to_str:n {#7} }
      }
      {
        \tl_if_in:onTF { #5 #6 } {#1}
          {
            \tl_if_in:nnTF {#6} {#1}
              { \exp_args:Nc \@@_replace:NnNNNnn {#2} {#2?} }
              {
                \@@_quark_if_nil:nTF {#6}
                  { \@@_replace_auxi:NnnNNNnn #5 {#1} { #1 \q_@@_stop } }
                  { \@@_replace_auxi:NnnNNNnn #5 {#1} { #1 \q_@@_nil  } }
              }
          }
          { \@@_replace_auxii:nNNNnn {#1} }
          #3#4#5 {#6} {#7}
      }
  }
\cs_new_protected:Npn \@@_replace_auxi:NnnNNNnn #1#2#3
  {
    \tl_if_in:NnTF #1 { #2 #3 #3 }
      { \@@_replace_auxi:NnnNNNnn #1 { #2 #3 } {#2} }
      { \@@_replace_auxii:nNNNnn { #2 #3 #3 } }
  }
%    \end{macrocode}
%   The auxiliary \cs{@@_replace_auxii:nNNNnn} receives the following
%   arguments:
%   \begin{quote}
%     \Arg{delimiter} \meta{function} \meta{assignment} \\
%       \meta{tl~var} \Arg{pattern} \Arg{replacement}
%   \end{quote}
%   All of its work is done between
%   \cs{group_align_safe_begin:} and \cs{group_align_safe_end:} to avoid
%   issues in alignments.  It does the actual replacement within
%   |#3|~|#4|~|{...}|, an \texttt{x}-expanding \meta{assignment}~|#3| to
%   the \meta{tl~var}~|#4|.  The auxiliary \cs{@@_replace_next:w} is
%   called, followed by the \meta{token list}, some tokens including the
%   \meta{delimiter}~|#1|, followed by the \meta{pattern}~|#5|.
%   This auxiliary finds an argument delimited by~|#5| (the presence of
%   a trailing~|#5| avoids runaway arguments) and calls
%   \cs{@@_replace_wrap:w} to test whether this |#5| is found within the
%   \meta{token list} or is the trailing one.
%
%   If on the one hand it is found within the \meta{token list}, then
%   |##1| cannot contain the \meta{delimiter}~|#1| that we worked so
%   hard to obtain, thus \cs{@@_replace_wrap:w} gets~|##1| as its own
%   argument~|##1|, and protects it against
%   the \texttt{x}-expanding assignment.  It also finds \cs{exp_not:n}
%   as~|##2| and does nothing to it, thus letting through \cs{exp_not:n}
%   \Arg{replacement} into the assignment.  Note that
%   \cs{@@_replace_next:w} and \cs{@@_replace_wrap:w} are always called
%   followed by two empty brace groups.  These are safe because no
%   delimiter can match them.  They prevent losing braces when grabbing
%   delimited arguments, but require the use of \cs{exp_not:o} and
%   \cs{use_none:nn}, rather than simply \cs{exp_not:n}.
%   Afterwards, \cs{@@_replace_next:w} is called
%   to repeat the replacement, or \cs{@@_replace_wrap:w} if we only want
%   a single replacement.  In this second case, |##1| is the
%   \meta{remaining tokens} in the \meta{token list} and |##2| is some
%   \meta{ending code} which ends the assignment and removes the
%   trailing tokens |#5| using some \cs{if_false:} |{| \cs{fi:} |}|
%   trickery because~|#5| may contain any delimiter.
%
%   If on the other hand the argument~|##1| of \cs{@@_replace_next:w} is
%   delimited by the trailing \meta{pattern}~|#5|, then |##1| is
%   \enquote{\{ \} \{ \} \meta{token list} \meta{delimiter}
%     \Arg{ending code}}, hence \cs{@@_replace_wrap:w} finds
%   \enquote{\{ \} \{ \} \meta{token list}} as |##1| and the
%   \meta{ending code} as~|##2|.  It leaves the \meta{token list} into
%   the assignment and unbraces the \meta{ending code} which removes
%   what remains (essentially the \meta{delimiter} and
%   \meta{replacement}).
%    \begin{macrocode}
\cs_new_protected:Npn \@@_replace_auxii:nNNNnn #1#2#3#4#5#6
  {
    \group_align_safe_begin:
    \cs_set:Npn \@@_replace_wrap:w ##1 #1 ##2
      { \__kernel_exp_not:w \exp_after:wN { \use_none:nn ##1 } ##2 }
    \cs_set:Npx \@@_replace_next:w ##1 #5
      {
        \exp_not:N \@@_replace_wrap:w ##1
        \exp_not:n { #1 }
        \exp_not:n { \exp_not:n {#6} }
        \exp_not:n { #2 { } { } }
      }
    #3 #4
      {
        \exp_after:wN \@@_replace_next_aux:w
        #4
        #1
        {
          \if_false: { \fi: }
          \exp_after:wN \use_none:n \exp_after:wN { \if_false: } \fi:
        }
        #5
      }
    \group_align_safe_end:
  }
\cs_new:Npn \@@_replace_next_aux:w { \@@_replace_next:w { } { } }
\cs_new_eq:NN \@@_replace_wrap:w ?
\cs_new_eq:NN \@@_replace_next:w ?
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\tl_remove_once:Nn, \tl_remove_once:cn}
% \begin{macro}{\tl_gremove_once:Nn, \tl_gremove_once:cn}
%   Removal is just a special case of replacement.
%    \begin{macrocode}
\cs_new_protected:Npn \tl_remove_once:Nn #1#2
  { \tl_replace_once:Nnn #1 {#2} { } }
\cs_new_protected:Npn \tl_gremove_once:Nn #1#2
  { \tl_greplace_once:Nnn #1 {#2} { } }
\cs_generate_variant:Nn \tl_remove_once:Nn  { c }
\cs_generate_variant:Nn \tl_gremove_once:Nn { c }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}{\tl_remove_all:Nn, \tl_remove_all:cn}
% \begin{macro}{\tl_gremove_all:Nn, \tl_gremove_all:cn}
%   Removal is just a special case of replacement.
%    \begin{macrocode}
\cs_new_protected:Npn \tl_remove_all:Nn #1#2
  { \tl_replace_all:Nnn #1 {#2} { } }
\cs_new_protected:Npn \tl_gremove_all:Nn #1#2
  { \tl_greplace_all:Nnn #1 {#2} { } }
\cs_generate_variant:Nn \tl_remove_all:Nn  { c }
\cs_generate_variant:Nn \tl_gremove_all:Nn { c }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \subsection{Token list conditionals}
%
% \begin{macro}[pTF]{\tl_if_empty:N, \tl_if_empty:c}
%    These functions check whether the token list in the argument is
%    empty and execute the proper code from their argument(s).
%    \begin{macrocode}
\prg_new_conditional:Npnn \tl_if_empty:N #1 { p , T , F , TF }
  {
    \if_meaning:w #1 \c_empty_tl
      \prg_return_true:
    \else:
      \prg_return_false:
    \fi:
  }
\prg_generate_conditional_variant:Nnn \tl_if_empty:N
  { c } { p , T , F , TF }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}[pTF]{\tl_if_empty:n, \tl_if_empty:V}
%   The \cs{if:w} triggers the expansion of \cs{tl_to_str:n} which converts the
%   argument to a string: this is empty if and only if the argument is.  Then
%   |\if:w \scan_stop: ... \scan_stop:| is \texttt{true} if and only if the
%   string |...| is empty.
%   It could be tempting to use |\if:w \scan_stop: #1 \scan_stop:| directly.
%   But this fails on a token list expanding to anything starting with
%   \cs{scan_stop:} leaving everything that follows in the input stream.
%    \begin{macrocode}
\prg_new_conditional:Npnn \tl_if_empty:n #1 { p , TF , T , F }
  {
    \if:w \scan_stop: \tl_to_str:n {#1} \scan_stop:
      \prg_return_true:
    \else:
      \prg_return_false:
    \fi:
  }
\prg_generate_conditional_variant:Nnn \tl_if_empty:n
  { V } { p , TF , T , F }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}[pTF,documented-as=\tl_if_empty:nTF]{\tl_if_empty:o}
% \begin{macro}[EXP]{\@@_if_empty_if:o}
%   The auxiliary function \cs{@@_if_empty_if:o} is for use
%   in various token list conditionals which reduce to testing
%   if a given token list is empty after applying a simple function
%   to it.
%   The test for emptiness is based on \cs{tl_if_empty:nTF}, but
%   the expansion is hard-coded for efficiency, as this auxiliary
%   function is used in several places.
%   We don't put \cs{prg_return_true:} and so on in the definition of
%   the auxiliary, because that would prevent an optimization applied to
%   conditionals that end with this code.
%   Also the |\@@_if_empty_if:o| is expanded once in |\tl_if_empty:oTF| for
%   efficiency as well (and to reduce code doubling).
%    \begin{macrocode}
\cs_new:Npn \@@_if_empty_if:o #1
  {
    \if:w \scan_stop: \__kernel_tl_to_str:w \exp_after:wN {#1} \scan_stop:
  }
\exp_args:Nno \use:n
  { \prg_new_conditional:Npnn \tl_if_empty:o #1 { p , TF , T , F } }
  {
    \@@_if_empty_if:o {#1}
      \prg_return_true:
    \else:
      \prg_return_false:
    \fi:
 }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}[pTF]{\tl_if_blank:n, \tl_if_blank:V, \tl_if_blank:o}
% \begin{macro}{\@@_if_blank_p:NNw}
%   \TeX{} skips spaces when reading a non-delimited arguments. Thus,
%   a \meta{token list} is blank if and only if \cs{use_none:n}
%   \meta{token list} |?| is empty after one expansion.  The auxiliary
%   \cs{@@_if_empty_if:o} is a fast emptyness test, converting its
%   argument to a string (after one expansion) and using the test
%   \cs{if:w} \cs{scan_stop:} |...| \cs{scan_stop:}.
%    \begin{macrocode}
\exp_args:Nno \use:n
  { \prg_new_conditional:Npnn \tl_if_blank:n #1 { p , T , F , TF } }
  {
    \@@_if_empty_if:o { \use_none:n #1 ? }
      \prg_return_true:
    \else:
      \prg_return_false:
    \fi:
  }
\prg_generate_conditional_variant:Nnn \tl_if_blank:n
  { e , V , o } { p , T , F , TF }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}[pTF]{\tl_if_eq:NN, \tl_if_eq:Nc, \tl_if_eq:cN, \tl_if_eq:cc}
%   Returns \cs{c_true_bool} if and only if the two token list variables are
%   equal.
%    \begin{macrocode}
\prg_new_conditional:Npnn \tl_if_eq:NN #1#2 { p , T , F , TF }
  {
    \if_meaning:w #1 #2
      \prg_return_true:
    \else:
      \prg_return_false:
    \fi:
  }
\prg_generate_conditional_variant:Nnn \tl_if_eq:NN
  { Nc , c , cc } { p , TF , T , F }
%    \end{macrocode}
% \end{macro}
%
% \begin{variable}{\l_@@_internal_a_tl, \l_@@_internal_b_tl}
%   Temporary storage.
%    \begin{macrocode}
\tl_new:N \l_@@_internal_a_tl
\tl_new:N \l_@@_internal_b_tl
%    \end{macrocode}
% \end{variable}
%
% \begin{macro}[TF]{\tl_if_eq:Nn}
%   A simple store and compare routine.
%    \begin{macrocode}
\prg_new_protected_conditional:Npnn \tl_if_eq:Nn #1#2 { T , F , TF }
  {
    \group_begin:
      \tl_set:Nn \l_@@_internal_b_tl {#2}
      \exp_after:wN
    \group_end:
    \if_meaning:w #1 \l_@@_internal_b_tl
      \prg_return_true:
    \else:
      \prg_return_false:
    \fi:
  }
\prg_generate_conditional_variant:Nnn \tl_if_eq:Nn { c } { TF , T , F }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}[TF]{\tl_if_eq:nn}
%   A simple store and compare routine.
%    \begin{macrocode}
\prg_new_protected_conditional:Npnn \tl_if_eq:nn #1#2 { T , F ,  TF }
  {
    \group_begin:
      \tl_set:Nn \l_@@_internal_a_tl {#1}
      \tl_set:Nn \l_@@_internal_b_tl {#2}
      \exp_after:wN
    \group_end:
    \if_meaning:w \l_@@_internal_a_tl \l_@@_internal_b_tl
      \prg_return_true:
    \else:
      \prg_return_false:
    \fi:
  }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}[TF]{\tl_if_in:Nn, \tl_if_in:cn}
%   See \cs{tl_if_in:nnTF} for further comments. Here we simply
%   expand the token list variable and pass it to \cs{tl_if_in:nnTF}.
%    \begin{macrocode}
\cs_new_protected:Npn \tl_if_in:NnT  { \exp_args:No \tl_if_in:nnT  }
\cs_new_protected:Npn \tl_if_in:NnF  { \exp_args:No \tl_if_in:nnF  }
\cs_new_protected:Npn \tl_if_in:NnTF { \exp_args:No \tl_if_in:nnTF }
\prg_generate_conditional_variant:Nnn \tl_if_in:Nn
  { c } { T , F , TF }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}[TF]{\tl_if_in:nn, \tl_if_in:Vn, \tl_if_in:on, \tl_if_in:no}
%   Once more, the test relies on the emptiness test for robustness.
%   The function \cs{@@_tmp:w} removes tokens until the first occurrence
%   of |#2|. If this does not appear in |#1|, then the final |#2| is removed,
%   leaving an empty token list. Otherwise some tokens remain, and the
%   test is \texttt{false}. See \cs{tl_if_empty:nTF} for details on
%   the emptiness test.
%
%   Treating correctly cases like
%   |\tl_if_in:nnTF {a state}{states}|, where |#1#2| contains |#2| before
%   the end, requires special care.
%   To cater for this case, we insert |{}{}| between the two token
%   lists. This marker may not appear in |#2| because of \TeX{} limitations
%   on what can delimit a parameter, hence we are safe. Using two brace
%   groups makes the test work also for empty arguments.
%   The \cs{if_false:} constructions are a faster way to do
%   \cs{group_align_safe_begin:} and \cs{group_align_safe_end:}.
%   The \cs{scan_stop:} ensures that \texttt{f}-expanding
%   \cs{tl_if_in:nn} does not lead to unbalanced braces.
%    \begin{macrocode}
\prg_new_protected_conditional:Npnn \tl_if_in:nn #1#2 { T  , F , TF }
  {
    \scan_stop:
    \if_false: { \fi:
    \cs_set:Npn \@@_tmp:w ##1 #2 { }
    \tl_if_empty:oTF { \@@_tmp:w #1 {} {} #2 }
      { \prg_return_false: } { \prg_return_true: }
    \if_false: } \fi:
  }
\prg_generate_conditional_variant:Nnn \tl_if_in:nn
  { V , o , no } { T , F , TF }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}[pTF, EXP]{\tl_if_novalue:n}
% \begin{macro}[EXP]{\@@_if_novalue:w}
%   Tests for |-NoValue-|: this is similar to \cs{tl_if_in:nn} but set
%   up to be expandable and to check the value exactly.  The question
%   mark prevents the auxiliary from losing braces.
%    \begin{macrocode}
\cs_set_protected:Npn \@@_tmp:w #1
  {
    \prg_new_conditional:Npnn \tl_if_novalue:n ##1
      { p , T ,  F , TF }
      {
        \str_if_eq:onTF
          { \@@_if_novalue:w ? ##1 { } #1 }
          { ? { } #1 }
          { \prg_return_true: }
          { \prg_return_false: }
      }
    \cs_new:Npn \@@_if_novalue:w ##1 #1 {##1}
  }
\exp_args:No \@@_tmp:w { \c_novalue_tl }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}[EXP,pTF]{\tl_if_single:N}
%   Expand the token list and feed it to \cs{tl_if_single:n}.
%    \begin{macrocode}
\cs_new:Npn \tl_if_single_p:N { \exp_args:No \tl_if_single_p:n }
\cs_new:Npn \tl_if_single:NT  { \exp_args:No \tl_if_single:nT  }
\cs_new:Npn \tl_if_single:NF  { \exp_args:No \tl_if_single:nF  }
\cs_new:Npn \tl_if_single:NTF { \exp_args:No \tl_if_single:nTF }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}[EXP,pTF]{\tl_if_single:n}
% \begin{macro}[EXP]{\@@_if_single:nnw}
%   This test is similar to \cs{tl_if_empty:nTF}.  Expanding
%   \cs{use_none:nn} |#1| |??| once yields an empty result if |#1| is
%   blank, a single~|?| if |#1| has a single item, and otherwise yields
%   some tokens ending with |??|.  Then, \cs{__kernel_tl_to_str:w} makes sure
%   there are no odd category codes.  An earlier version would compare
%   the result to a single~|?| using string comparison, but the Lua call
%   is slow in \LuaTeX{}.  Instead, \cs{@@_if_single:nnw} picks the
%   second token in front of it.  If |#1| is empty, this token is
%   the trailing~|?| and the |\if:w| test yields \texttt{false}.  If
%   |#1| has a single item, the token is~|\scan_stop:| and the |\if:w| test
%   yields \texttt{true}.  Otherwise, it is one of the characters
%   resulting from \cs{tl_to_str:n}, and the |\if:w| test yields
%   \texttt{false}.  Note that \cs{if:w} and
%   \cs{__kernel_tl_to_str:w} are primitives that take care of
%   expansion.
%    \begin{macrocode}
\prg_new_conditional:Npnn \tl_if_single:n #1 { p , T , F , TF }
  {
    \if:w \scan_stop: \exp_after:wN \@@_if_single:nnw
        \__kernel_tl_to_str:w
          \exp_after:wN { \use_none:nn #1 ?? } \scan_stop: ? \s_@@_stop
      \prg_return_true:
    \else:
      \prg_return_false:
    \fi:
  }
\cs_new:Npn \@@_if_single:nnw #1#2#3 \s_@@_stop {#2}
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}[EXP,pTF]{\tl_if_single_token:n}
%   There are four cases: empty token list, token list starting with a
%   normal token, with a brace group, or with a space token.  If the
%   token list starts with a normal token, remove it and check for
%   emptiness.  For the next case, an empty token list is not a single
%   token.  Finally, we have a non-empty token list starting with a
%   space or a brace group.  Applying \texttt{f}-expansion yields an
%   empty result if and only if the token list is a single space.
%    \begin{macrocode}
\prg_new_conditional:Npnn \tl_if_single_token:n #1 { p , T , F , TF }
  {
    \tl_if_head_is_N_type:nTF {#1}
      { \@@_if_empty_if:o { \use_none:n #1 } }
      {
        \tl_if_empty:nTF {#1}
          { \if_false: }
          { \@@_if_empty_if:o { \exp:w \exp_end_continue_f:w #1 } }
      }
      \prg_return_true:
    \else:
      \prg_return_false:
    \fi:
  }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}[EXP, noTF]{\tl_case:Nn, \tl_case:cn}
% \begin{macro}[EXP]{\@@_case:nnTF}
% \begin{macro}[EXP]{\@@_case:Nw}
% \begin{macro}[EXP]{\@@_case_end:nw}
%   The aim here is to allow the case statement to be evaluated
%   using a known number of expansion steps (two), and without
%   needing to use an explicit \enquote{end of recursion} marker.
%   That is achieved by using the test input as the final case,
%   as this is always true. The trick is then to tidy up
%   the output such that the appropriate case code plus either
%   the \texttt{true} or \texttt{false} branch code is inserted.
%    \begin{macrocode}
\cs_new:Npn \tl_case:Nn #1#2
  {
    \exp:w
    \@@_case:NnTF #1 {#2} { } { }
  }
\cs_new:Npn \tl_case:NnT #1#2#3
  {
    \exp:w
    \@@_case:NnTF #1 {#2} {#3} { }
  }
\cs_new:Npn \tl_case:NnF #1#2#3
  {
    \exp:w
    \@@_case:NnTF #1 {#2} { } {#3}
  }
\cs_new:Npn \tl_case:NnTF #1#2
  {
    \exp:w
    \@@_case:NnTF #1 {#2}
  }
\cs_new:Npn \@@_case:NnTF #1#2#3#4
  { \@@_case:Nw #1 #2 #1 { } \s_@@_mark {#3} \s_@@_mark {#4} \s_@@_stop }
\cs_new:Npn \@@_case:Nw #1#2#3
  {
    \tl_if_eq:NNTF #1 #2
      { \@@_case_end:nw {#3} }
      { \@@_case:Nw #1 }
  }
\cs_generate_variant:Nn \tl_case:Nn   { c }
\prg_generate_conditional_variant:Nnn \tl_case:Nn
  { c } { T , F , TF }
%    \end{macrocode}
%   To tidy up the recursion, there are two outcomes. If there was a hit to
%   one of the cases searched for, then |#1| is the code to insert,
%   |#2| is the \emph{next} case to check on and |#3| is all of
%   the rest of the cases code. That means that |#4| is the \texttt{true}
%   branch code, and |#5| tidies up the spare \cs{s_@@_mark} and the
%   \texttt{false} branch. On the other hand, if none of the cases matched
%   then we arrive here using the \enquote{termination} case of comparing
%   the search with itself. That means that |#1| is empty, |#2| is
%   the first \cs{s_@@_mark} and so |#4| is the \texttt{false} code (the
%   \texttt{true} code is mopped up by |#3|).
%    \begin{macrocode}
\cs_new:Npn \@@_case_end:nw #1#2#3 \s_@@_mark #4#5 \s_@@_stop
  { \exp_end: #1 #4 }
%    \end{macrocode}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
%
% \subsection{Mapping over token lists}
%
% \begin{macro}{\tl_map_function:nN}
% \begin{macro}{\tl_map_function:NN, \tl_map_function:cN}
% \begin{macro}{\@@_map_function:Nn}
%   Expandable loop macro for token lists. These have the advantage of not
%   needing to test if the argument is empty, because if it is, the stop
%   marker is read immediately and the loop terminated.
%    \begin{macrocode}
\cs_new:Npn \tl_map_function:nN #1#2
  {
    \@@_map_function:Nn #2 #1
      \q_@@_recursion_tail
    \prg_break_point:Nn \tl_map_break: { }
  }
\cs_new:Npn \tl_map_function:NN
  { \exp_args:No \tl_map_function:nN }
\cs_new:Npn \@@_map_function:Nn #1#2
  {
    \@@_if_recursion_tail_break:nN {#2} \tl_map_break:
    #1 {#2} \@@_map_function:Nn #1
  }
\cs_generate_variant:Nn \tl_map_function:NN { c }
%    \end{macrocode}
% \end{macro}
% \end{macro}
% \end{macro}
%
% \begin{macro}{\tl_map_inline:nn}
% \begin{macro}{\tl_map_inline:Nn, \tl_map_inline:cn}
%   The inline functions are straight forward by now. We use a little
%   trick with the counter \cs{g__kernel_prg_map_int} to make
%   them nestable. We can also make use of \cs{@@_map_function:Nn}
%   from before.
%    \begin{macrocode}
\cs_new_protected:Npn \tl_map_inline:nn #1#2
  {
    \int_gincr:N \g__kernel_prg_map_int
    \cs_gset_protected:cpn
      { @@_map_ \int_use:N \g__kernel_prg_map_int :w } ##1 {#2}
    \exp_args:Nc \@@_map_function:Nn
      { @@_map_ \int_use:N \g__kernel_prg_map_int :w }
      #1 \q_@@_recursion_tail
    \prg_break_point:Nn \tl_map_break:
      { \int_gdecr:N \g__kernel_prg_map_int }
  }
\cs_new_protected:Npn \tl_map_inline:Nn
  { \exp_args:No \tl_map_inline:nn }
\cs_generate_variant:Nn \tl_map_inline:Nn { c }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}{\tl_map_tokens:nn}
% \begin{macro}{\tl_map_tokens:Nn, \tl_map_tokens:cn}
% \begin{macro}{\@@_map_tokens:nn}
%   Much like the function mapping.
%    \begin{macrocode}
\cs_new:Npn \tl_map_tokens:nn #1#2
  {
    \@@_map_tokens:nn {#2} #1
      \q_@@_recursion_tail
    \prg_break_point:Nn \tl_map_break: { }
  }
\cs_new:Npn \tl_map_tokens:Nn
  { \exp_args:No \tl_map_tokens:nn }
\cs_generate_variant:Nn \tl_map_tokens:Nn { c }
\cs_new:Npn \@@_map_tokens:nn #1#2
  {
    \@@_if_recursion_tail_break:nN {#2} \tl_map_break:
    \use:n {#1} {#2}
    \@@_map_tokens:nn {#1}
  }
%    \end{macrocode}
% \end{macro}
% \end{macro}
% \end{macro}
%
% \begin{macro}{\tl_map_variable:nNn}
% \begin{macro}{\tl_map_variable:NNn, \tl_map_variable:cNn}
% \begin{macro}{\@@_map_variable:Nnn}
%   \cs{tl_map_variable:nNn} \meta{token list} \meta{tl~var}
%   \meta{action} assigns \meta{tl~var} to each element and executes
%   \meta{action}.  The assignment to \meta{tl~var} is done after the
%   quark test so that this variable does not get set to a quark.
%    \begin{macrocode}
\cs_new_protected:Npn \tl_map_variable:nNn #1#2#3
  {
    \@@_map_variable:Nnn #2 {#3} #1
      \q_@@_recursion_tail
    \prg_break_point:Nn \tl_map_break: { }
  }
\cs_new_protected:Npn \tl_map_variable:NNn
  { \exp_args:No \tl_map_variable:nNn }
\cs_new_protected:Npn \@@_map_variable:Nnn #1#2#3
  {
    \@@_if_recursion_tail_break:nN {#3} \tl_map_break:
    \tl_set:Nn #1 {#3}
    \use:n {#2}
    \@@_map_variable:Nnn #1 {#2}
  }
\cs_generate_variant:Nn \tl_map_variable:NNn { c }
%    \end{macrocode}
% \end{macro}
% \end{macro}
% \end{macro}
%
% \begin{macro}{\tl_map_break:}
% \begin{macro}{\tl_map_break:n}
%   The break statements use the general \cs{prg_map_break:Nn}.
%    \begin{macrocode}
\cs_new:Npn \tl_map_break:
  { \prg_map_break:Nn \tl_map_break: { } }
\cs_new:Npn \tl_map_break:n
  { \prg_map_break:Nn \tl_map_break: }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \subsection{Using token lists}
%
% \begin{macro}{\tl_to_str:n, \tl_to_str:V}
%   Another name for a primitive: defined in \pkg{l3basics}.
%    \begin{macrocode}
\cs_generate_variant:Nn \tl_to_str:n { V }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\tl_to_str:N, \tl_to_str:c}
%    These functions return the replacement text of a token list as a
%    string.
%    \begin{macrocode}
\cs_new:Npn \tl_to_str:N #1 { \__kernel_tl_to_str:w \exp_after:wN {#1} }
\cs_generate_variant:Nn \tl_to_str:N { c }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\tl_use:N, \tl_use:c}
% Token lists which are simply not defined give a clear \TeX{}
% error here. No such luck for ones equal to \cs{scan_stop:} so
% instead a test is made and if there is an issue an error is forced.
%    \begin{macrocode}
\cs_new:Npn \tl_use:N #1
  {
    \tl_if_exist:NTF #1 {#1}
      {
        \__kernel_msg_expandable_error:nnn
          { kernel } { bad-variable } {#1}
      }
  }
\cs_generate_variant:Nn \tl_use:N { c }
%    \end{macrocode}
% \end{macro}
%
% \subsection{Working with the contents of token lists}
%
% \begin{macro}{\tl_count:n, \tl_count:V, \tl_count:o}
% \begin{macro}{\tl_count:N, \tl_count:c}
% \begin{macro}{\@@_count:n}
%   Count number of elements within a token list or token list
%   variable. Brace groups within the list are read as a single
%   element. Spaces are ignored.
%   \cs{@@_count:n} grabs the element and replaces it by |+1|.
%   The |0| ensures that it works on an empty list.
%    \begin{macrocode}
\cs_new:Npn \tl_count:n #1
  {
    \int_eval:n
      { 0 \tl_map_function:nN {#1} \@@_count:n }
  }
\cs_new:Npn \tl_count:N #1
  {
    \int_eval:n
      { 0 \tl_map_function:NN #1 \@@_count:n }
  }
\cs_new:Npn \@@_count:n #1 { + 1 }
\cs_generate_variant:Nn \tl_count:n { V , o }
\cs_generate_variant:Nn \tl_count:N { c }
%    \end{macrocode}
% \end{macro}
% \end{macro}
% \end{macro}
%
%
% \begin{macro}[EXP]{\tl_count_tokens:n}
% \begin{macro}[EXP]{\@@_act_count_normal:nN,
%     \@@_act_count_group:nn, \@@_act_count_space:n}
%   The token count is computed through an \cs{int_eval:n} construction.
%   Each \texttt{1+} is output to the \emph{left}, into the integer
%   expression, and the sum is ended by the \cs{exp_end:} inserted by
%   \cs{@@_act_end:wn} (which is technically implemented as  \cs{c_zero_int}).
%   Somewhat a hack!
%    \begin{macrocode}
\cs_new:Npn \tl_count_tokens:n #1
  {
    \int_eval:n
      {
        \@@_act:NNNn
          \@@_act_count_normal:N
          \@@_act_count_group:n
          \@@_act_count_space:
          {#1}
      }
  }
\cs_new:Npn \@@_act_count_normal:N #1 { 1 + }
\cs_new:Npn \@@_act_count_space: { 1 + }
\cs_new:Npn \@@_act_count_group:n #1 { 2 + \tl_count_tokens:n {#1} + }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}{\tl_reverse_items:n}
% \begin{macro}{\@@_reverse_items:nwNwn}
% \begin{macro}{\@@_reverse_items:wn}
%    Reversal of a token list is done by taking one item at a time
%    and putting it after \cs{s_@@_stop}.
%    \begin{macrocode}
\cs_new:Npn \tl_reverse_items:n #1
  {
    \@@_reverse_items:nwNwn #1 ?
      \s_@@_mark \@@_reverse_items:nwNwn
      \s_@@_mark \@@_reverse_items:wn
      \s_@@_stop { }
  }
\cs_new:Npn \@@_reverse_items:nwNwn #1 #2 \s_@@_mark #3 #4 \s_@@_stop #5
  {
    #3 #2
      \s_@@_mark \@@_reverse_items:nwNwn
      \s_@@_mark \@@_reverse_items:wn
      \s_@@_stop { {#1} #5 }
  }
\cs_new:Npn \@@_reverse_items:wn #1 \s_@@_stop #2
  { \__kernel_exp_not:w \exp_after:wN { \use_none:nn #2 } }
%    \end{macrocode}
% \end{macro}
% \end{macro}
% \end{macro}
%
% \begin{macro}{\tl_trim_spaces:n, \tl_trim_spaces:o}
% \begin{macro}{\tl_trim_spaces_apply:nN, \tl_trim_spaces_apply:oN}
% \begin{macro}
%   {
%     \tl_trim_spaces:N, \tl_trim_spaces:c,
%     \tl_gtrim_spaces:N, \tl_gtrim_spaces:c
%   }
%   Trimming spaces from around the input is deferred to an internal
%   function whose first argument is the token list to trim, augmented
%   by an initial \cs{@@_trim_mark:}, and whose second argument is a
%   \meta{continuation}, which receives as a braced argument
%   \cs{@@_trim_mark:} \meta{trimmed token list}.  The control sequence
%   \cs{@@_trim_mark:} expands to nothing in a single expansion.  In the case
%   at hand, we take \cs{__kernel_exp_not:w} \cs{exp_after:wN} as our
%   continuation, so that space trimming behaves correctly within an
%   \texttt{x}-type expansion.
%    \begin{macrocode}
\cs_new:Npn \tl_trim_spaces:n #1
  {
    \@@_trim_spaces:nn
      { \@@_trim_mark: #1 }
      { \__kernel_exp_not:w \exp_after:wN }
  }
\cs_generate_variant:Nn \tl_trim_spaces:n { o }
\cs_new:Npn \tl_trim_spaces_apply:nN #1#2
  { \@@_trim_spaces:nn { \@@_trim_mark: #1 } { \exp_args:No #2 } }
\cs_generate_variant:Nn \tl_trim_spaces_apply:nN { o }
\cs_new_protected:Npn \tl_trim_spaces:N #1
  { \__kernel_tl_set:Nx #1 { \exp_args:No \tl_trim_spaces:n {#1} } }
\cs_new_protected:Npn \tl_gtrim_spaces:N #1
  { \__kernel_tl_gset:Nx #1 { \exp_args:No \tl_trim_spaces:n {#1} } }
\cs_generate_variant:Nn \tl_trim_spaces:N  { c }
\cs_generate_variant:Nn \tl_gtrim_spaces:N { c }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}{\@@_trim_spaces:nn}
% \begin{macro}
%   {
%     \@@_trim_spaces_auxi:w, \@@_trim_spaces_auxii:w,
%     \@@_trim_spaces_auxiii:w, \@@_trim_spaces_auxiv:w
%   }
% \begin{macro}{\@@_trim_mark:}
%   Trimming spaces from around the input is done using delimited
%   arguments and quarks, and to get spaces at odd places in the
%   definitions, we nest those in \cs{@@_tmp:w}, which then receives
%   a single space as its argument: |#1| is \verb*+ +.
%   Removing leading spaces is done with \cs{@@_trim_spaces_auxi:w},
%   which loops until \cs{@@_trim_mark:}\verb*+ + matches the end of the token
%   list: then |##1| is the token list and |##3| is
%   \cs{@@_trim_spaces_auxii:w}. This hands the relevant tokens to the
%   loop \cs{@@_trim_spaces_auxiii:w}, responsible for trimming
%   trailing spaces. The end is reached when \verb*+ + \cs{s_@@_nil}
%   matches the one present in the definition of \cs{tl_trim_spaces:n}.
%   Then \cs{@@_trim_spaces_auxiv:w} puts the token list into a group,
%   with a lingering \cs{@@_trim_mark:} at the start (which will expand to
%   nothing in one step of expansion), and feeds this to the
%   \meta{continuation}.
%    \begin{macrocode}
\cs_set_protected:Npn \@@_tmp:w #1
  {
    \cs_new:Npn \@@_trim_spaces:nn ##1
      {
        \@@_trim_spaces_auxi:w
          ##1
          \s_@@_nil
          \@@_trim_mark: #1 { }
          \@@_trim_mark: \@@_trim_spaces_auxii:w
          \@@_trim_spaces_auxiii:w
          #1 \s_@@_nil
          \@@_trim_spaces_auxiv:w
        \s_@@_stop
      }
    \cs_new:Npn
        \@@_trim_spaces_auxi:w ##1 \@@_trim_mark: #1 ##2 \@@_trim_mark: ##3
      {
        ##3
        \@@_trim_spaces_auxi:w
        \@@_trim_mark:
        ##2
        \@@_trim_mark: #1 {##1}
      }
    \cs_new:Npn \@@_trim_spaces_auxii:w
        \@@_trim_spaces_auxi:w \@@_trim_mark: \@@_trim_mark: ##1
      {
        \@@_trim_spaces_auxiii:w
        ##1
      }
    \cs_new:Npn \@@_trim_spaces_auxiii:w ##1 #1 \s_@@_nil ##2
      {
        ##2
        ##1 \s_@@_nil
        \@@_trim_spaces_auxiii:w
      }
    \cs_new:Npn \@@_trim_spaces_auxiv:w ##1 \s_@@_nil ##2 \s_@@_stop ##3
      { ##3 { ##1 } }
    \cs_new:Npn \@@_trim_mark: {}
  }
\@@_tmp:w { ~ }
%    \end{macrocode}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
%
% \begin{macro}
%   {\tl_sort:Nn, \tl_sort:cn, \tl_gsort:Nn, \tl_gsort:cn, \tl_sort:nN}
%   Implemented in \pkg{l3sort}.
% \end{macro}
%
% \subsection{The first token from a token list}
%
% \begin{macro}{\tl_head:N, \tl_head:n, \tl_head:V, \tl_head:v, \tl_head:f}
% \begin{macro}{\@@_head_auxi:nw, \@@_head_auxii:n}
% \begin{macro}{\tl_head:w,\@@_tl_head:w}
% \begin{macro}{\tl_tail:N, \tl_tail:n, \tl_tail:V, \tl_tail:v, \tl_tail:f}
%   Finding the head of a token list expandably always strips braces, which
%   is fine as this is consistent with for example mapping over a list. The
%   empty brace groups in \cs{tl_head:n} ensure that a blank argument gives an
%   empty result. The result is returned within the \tn{unexpanded} primitive.
%   The approach here is to use \cs{if_false:} to allow us to use |}| as
%   the closing delimiter: this is the only safe choice, as any other token
%   would not be able to parse it's own code. If the \tn{expanded} primitive
%   is available it is used to get a fast and safe code variant in which we
%   don't have to ensure that the left-most token is an internal to not break
%   in an |f|-type expansion. If \tn{expanded} isn't available, using a marker,
%   we can see if what we are grabbing is exactly the marker, or there is
%   anything else to deal with. If there is, there is a loop. If not, tidy up
%   and leave the item in the output stream. More detail in
%   \url{http://tex.stackexchange.com/a/70168}.
%    \begin{macrocode}
\cs_if_exist:NTF \tex_expanded:D
  {
    \cs_new:Npn \tl_head:n #1
      {
        \__kernel_exp_not:w \tex_expanded:D
          { { \if_false: { \fi: \@@_head_aux:n #1 { } } } }
      }
    \cs_new:Npn \@@_head_aux:n #1
      {
        \__kernel_exp_not:w {#1}
        \exp_after:wN \use_none:n \exp_after:wN { \if_false: } \fi:
      }
  }
  {
    \cs_new:Npn \tl_head:n #1
      {
        \__kernel_exp_not:w
          \if_false: { \fi: \@@_head_auxi:nw #1 { } \s_@@_stop }
      }
    \cs_new:Npn \@@_head_auxi:nw #1#2 \s_@@_stop
      {
        \exp_after:wN \@@_head_auxii:n \exp_after:wN {
          \if_false: } \fi: {#1}
      }
    \exp_args:Nno \use:n
      { \cs_new:Npn \@@_head_auxii:n #1 }
      {
        \@@_if_empty_if:o { \use_none:n #1 }
          \exp_after:wN \use_ii:nnn
        \fi:
        \use_ii:nn
          {#1}
          { \if_false: { \fi: \@@_head_auxi:nw #1 } }
      }
  }
\cs_generate_variant:Nn \tl_head:n { V , v , f }
\cs_new:Npn \tl_head:w #1#2 \q_stop {#1}
\cs_new:Npn \@@_tl_head:w #1#2 \s_@@_stop {#1}
\cs_new:Npn \tl_head:N { \exp_args:No \tl_head:n }
%    \end{macrocode}
%   To correctly leave the tail of a token list, it's important \emph{not} to
%   absorb any of the tail part as an argument. For example, the simple
%   definition
%   \begin{verbatim}
%     \cs_new:Npn \tl_tail:n #1 { \tl_tail:w #1 \q_stop }
%     \cs_new:Npn \tl_tail:w #1#2 \q_stop
%   \end{verbatim}
%   would give the wrong result for |\tl_tail:n { a { bc } }| (the braces would
%   be stripped). Thus the only safe way to proceed is to first check that
%   there is an item to grab (\emph{i.e.}~that the argument is not blank) and
%   assuming there is to dispose of the first item.  As with \cs{tl_head:n},
%   the result is protected from further expansion by \tn{unexpanded}.
%   While we could optimise the test here, this would leave some tokens
%   \enquote{banned} in the input, which we do not have with this definition.
%    \begin{macrocode}
\exp_args:Nno \use:n { \cs_new:Npn \tl_tail:n #1 }
  {
    \exp_after:wN \__kernel_exp_not:w
      \tl_if_blank:nTF {#1}
        { { } }
        { \exp_after:wN { \use_none:n #1 } }
  }
\cs_generate_variant:Nn \tl_tail:n { V , v , f }
\cs_new:Npn \tl_tail:N { \exp_args:No \tl_tail:n }
%    \end{macrocode}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
%
% \begin{macro}[pTF]{\tl_if_head_eq_meaning:nN}
% \begin{macro}[pTF]{\tl_if_head_eq_charcode:nN}
% \begin{macro}[pTF]{\tl_if_head_eq_charcode:fN}
% \begin{macro}[pTF]{\tl_if_head_eq_catcode:nN}
% \begin{macro}[pTF]{\tl_if_head_eq_catcode:oN}
% \begin{macro}[EXP]{\@@_head_exp_not:w}
% \begin{macro}[EXP]{\@@_if_head_eq_empty_arg:w}
%   Accessing the first token of a token list is tricky in three cases:
%   when it has category code $1$ (begin-group token), when it is an
%   explicit space, with category code $10$ and character code $32$, or
%   when the token list is empty (obviously).
%
%   Forgetting temporarily about this issue we would use the following
%   test in \cs{tl_if_head_eq_charcode:nN}.  Here, \cs{tl_head:w} yields
%   the first token of the token list, then passed to \cs{exp_not:N}.
% \begin{verbatim}
% \if_charcode:w
%     \exp_after:wN \exp_not:N \tl_head:w #1 \q_nil \q_stop
%     \exp_not:N #2
% \end{verbatim}
%   The two first special cases are detected by testing if the token
%   list starts with an \texttt{N}-type token (the extra |?| sends empty
%   token lists to the \texttt{true} branch of this test).  In those
%   cases, the first token is a character, and since we only care about
%   its character code, we can use \cs{str_head:n} to access it (this
%   works even if it is a space character).  An empty argument
%   results in \cs{tl_head:w} leaving two token: |^| and
%   \cs{@@_if_head_eq_empty_arg:w} which will result in the \cs{if_charcode:w}
%   test being false and remove \cs{exp_not:N} and |#2|.
%    \begin{macrocode}
\prg_new_conditional:Npnn \tl_if_head_eq_charcode:nN #1#2 { p , T , F , TF }
  {
    \if_charcode:w
        \tl_if_head_is_N_type:nTF { #1 ? }
          { \@@_head_exp_not:w #1 { ^ \@@_if_head_eq_empty_arg:w } \s_@@_stop }
          { \str_head:n {#1} }
        \exp_not:N #2
      \prg_return_true:
    \else:
      \prg_return_false:
    \fi:
  }
\prg_generate_conditional_variant:Nnn \tl_if_head_eq_charcode:nN
  { f } { p , TF , T , F }
%    \end{macrocode}
%   For \cs{tl_if_head_eq_catcode:nN}, again we detect special cases
%   with a \cs{tl_if_head_is_N_type:n}.  Then we need to test if the
%   first token is a begin-group token or an explicit space token, and
%   produce the relevant token, either \cs{c_group_begin_token} or
%   \cs{c_space_token}.  Again, for an empty argument, a hack is used,
%   removing the token given by the user and leaving two tokens in the input
%   stream which will make the \cs{if_catcode:w} test return \texttt{false}.
%    \begin{macrocode}
\prg_new_conditional:Npnn \tl_if_head_eq_catcode:nN #1 #2 { p , T , F , TF }
  {
    \if_catcode:w
        \tl_if_head_is_N_type:nTF { #1 ? }
          { \@@_head_exp_not:w #1 { ^ \@@_if_head_eq_empty_arg:w } \s_@@_stop }
          {
            \tl_if_head_is_group:nTF {#1}
              \c_group_begin_token
              \c_space_token
          }
        \exp_not:N #2
      \prg_return_true:
    \else:
      \prg_return_false:
    \fi:
  }
\prg_generate_conditional_variant:Nnn \tl_if_head_eq_catcode:nN
  { o } { p , TF , T , F }
%    \end{macrocode}
%   For \cs{tl_if_head_eq_meaning:nN}, again, detect special cases.  In
%   the normal case, use \cs{tl_head:w}, with no \cs{exp_not:N} this
%   time, since \cs{if_meaning:w} causes no expansion.  With an empty
%   argument, the test is \texttt{true}, and \cs{use_none:nnn} removes
%   |#2| and \cs{prg_return_true:} and \cs{else:} (it is safe this way here as
%   in this case \cs{prg_new_conditional:Npnn} didn't optimize these two away).
%   In the special cases, we know that the first token is a character,
%   hence \cs{if_charcode:w} and \cs{if_catcode:w} together are enough.
%   We combine them in some order, hopefully faster than the reverse.
%   Tests are not nested because the arguments may contain unmatched
%   primitive conditionals.
%    \begin{macrocode}
\prg_new_conditional:Npnn \tl_if_head_eq_meaning:nN #1#2 { p , T , F , TF }
  {
    \tl_if_head_is_N_type:nTF { #1 ? }
      \@@_if_head_eq_meaning_normal:nN
      \@@_if_head_eq_meaning_special:nN
    {#1} #2
  }
\cs_new:Npn \@@_if_head_eq_meaning_normal:nN #1 #2
  {
    \exp_after:wN \if_meaning:w
        \@@_tl_head:w #1 { ?? \use_none:nnn } \s_@@_stop #2
      \prg_return_true:
    \else:
      \prg_return_false:
    \fi:
  }
\cs_new:Npn \@@_if_head_eq_meaning_special:nN #1 #2
  {
    \if_charcode:w \str_head:n {#1} \exp_not:N #2
      \exp_after:wN \use_ii:nn
    \else:
      \prg_return_false:
    \fi:
    \use_none:n
    {
      \if_catcode:w \exp_not:N #2
                    \tl_if_head_is_group:nTF {#1}
                      { \c_group_begin_token }
                      { \c_space_token }
        \prg_return_true:
      \else:
        \prg_return_false:
      \fi:
    }
  }
%    \end{macrocode}
%
%   Both \cs{tl_if_head_eq_charcode:nN} and \cs{tl_if_head_eq_catcode:nN} will
%   need to get the first token of their argument and apply \cs{exp_not:N} to
%   it. \cs{@@_head_exp_not:w} does exactly that.
%    \begin{macrocode}
\cs_new:Npn \@@_head_exp_not:w #1 #2 \s_@@_stop
  { \exp_not:N #1 }
%    \end{macrocode}
%
%   If the argument of \cs{tl_if_head_eq_charcode:nN} and
%   \cs{tl_if_head_eq_catcode:nN} was empty \cs{@@_if_head_eq_empty_arg:w} will
%   be left in the input stream. This macro has to remove \cs{exp_not:N} and
%   the following token from the input stream to make sure no unbalanced
%   if-construct is created and leave tokens there which make the two tests
%   return \texttt{false}.
%    \begin{macrocode}
\cs_new:Npn \@@_if_head_eq_empty_arg:w \exp_not:N #1
  { ? }
%    \end{macrocode}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
%
% \begin{macro}[pTF]{\tl_if_head_is_N_type:n}
% \begin{macro}[EXP]
%   {
%     \@@_if_head_is_N_type_auxi:w   ,
%     \@@_if_head_is_N_type_auxii:nn ,
%     \@@_if_head_is_N_type_auxiii:n
%   }
%   A token list can be empty, can start with an explicit space
%   character (catcode 10 and charcode 32), can start with a begin-group
%   token (catcode 1), or start with an \texttt{N}-type argument.  In
%   the first two cases, the line involving \cs{@@_if_head_is_N_type_auxi:w}
%   produces~|f| (and otherwise nothing).  In the third case
%   (begin-group token), the lines involving \cs{token_to_str:N} produce a
%   single closing brace.  The category code test is thus true exactly
%   in the fourth case, which is what we want.  One cannot optimize by
%   moving one of the \cs{scan_stop:} to the beginning: if |#1| contains
%   primitive conditionals, all of its occurrences must be dealt with before
%   the \cs{if:w} tries to skip the \texttt{true} branch of the
%   conditional.
%    \begin{macrocode}
\prg_new_conditional:Npnn \tl_if_head_is_N_type:n #1 { p , T , F , TF }
  {
    \if:w
        \if_false: { \fi: \@@_if_head_is_N_type_auxi:w \prg_do_nothing: #1 ~ }
        { \exp_after:wN { \token_to_str:N #1 } }
        \scan_stop: \scan_stop:
      \prg_return_true:
    \else:
      \prg_return_false:
    \fi:
  }
\exp_args:Nno \use:n { \cs_new:Npn \@@_if_head_is_N_type_auxi:w #1 ~ }
  {
    \tl_if_empty:oTF { #1 }
      { f \exp_after:wN \use_none:nn }
      { \exp_after:wN \@@_if_head_is_N_type_auxii:n }
    \exp_after:wN { \if_false: } \fi:
  }
\cs_new:Npn \@@_if_head_is_N_type_auxii:n #1
  { \exp_after:wN \use_none:n \exp_after:wN }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}[EXP,pTF]{\tl_if_head_is_group:n}
% \begin{macro}[EXP]{\@@_if_head_is_group_fi_false:w}
%   Pass the first token of |#1| through \cs{token_to_str:N}, then check
%   for the brace balance.  The extra \texttt{?} caters for an empty
%   argument.  This could be made faster, but we need all brace tricks
%   to happen in one step of expansion, keeping the token list brace
%   balanced at all times.
%    \begin{macrocode}
\prg_new_conditional:Npnn \tl_if_head_is_group:n #1 { p , T , F , TF }
  {
    \if:w
        \exp_after:wN \use_none:n
          \exp_after:wN { \exp_after:wN { \token_to_str:N #1 ? } }
        \scan_stop: \scan_stop:
      \@@_if_head_is_group_fi_false:w
    \fi:
    \if_true:
      \prg_return_true:
    \else:
      \prg_return_false:
    \fi:
  }
\cs_new:Npn \@@_if_head_is_group_fi_false:w \fi: \if_true: { \fi: \if_false: }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}[EXP,pTF]{\tl_if_head_is_space:n}
% \begin{macro}[EXP]{\@@_if_head_is_space:w}
%   The auxiliary's argument is all that is before the first explicit
%   space in |\prg_do_nothing:#1?~|.  If that is a single~|\prg_do_nothing:| the
%   test yields \texttt{true}.  Otherwise, that is more than one token, and the
%   test yields \texttt{false}.  The work is done within braces (with an
%   |\if_false: { \fi: ... }| construction) both to hide potential
%   alignment tab characters from \TeX{} in a table, and to allow for
%   removing what remains of the token list after its first space.  The use of
%   \cs{if:w} ensures that the result of a single step of expansion directly
%   yields a balanced token list (no trailing closing brace).
%    \begin{macrocode}
\prg_new_conditional:Npnn \tl_if_head_is_space:n #1 { p , T , F , TF }
  {
    \if:w
        \if_false: { \fi: \@@_if_head_is_space:w \prg_do_nothing: #1 ? ~ }
        \scan_stop: \scan_stop:
      \prg_return_true:
    \else:
      \prg_return_false:
    \fi:
  }
\exp_args:Nno \use:n { \cs_new:Npn \@@_if_head_is_space:w #1 ~ }
  {
    \@@_if_empty_if:o {#1} \else: f \fi:
    \exp_after:wN \use_none:n \exp_after:wN { \if_false: } \fi:
  }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \subsection{Token by token changes}
%
% \begin{variable}{\s_@@_act_stop}
%   The \cs[no-index]{@@_act_\ldots{}} functions may be applied to any token list.
%   Hence, we use a private quark, to allow any token, even quarks,
%   in the token list.
%   Only \cs{s_@@_act_stop} may not appear in the token lists manipulated by
%   \cs{@@_act:NNNn} functions.
%    \begin{macrocode}
\scan_new:N \s_@@_act_stop
%    \end{macrocode}
% \end{variable}
%
% \begin{macro}[EXP]{\@@_act:NNNn}
% \begin{macro}[EXP]{\@@_act_output:n, \@@_act_reverse_output:n}
% \begin{macro}[EXP]{\@@_act_loop:w}
% \begin{macro}[EXP]{\@@_act_normal:NwNNN}
% \begin{macro}[EXP]{\@@_act_group:nwNNN}
% \begin{macro}[EXP]{\@@_act_space:wwNNN}
% \begin{macro}[EXP]{\@@_act_end:w}
% \begin{macro}[EXP]
%   {
%     \@@_act_if_head_is_space:nTF,
%     \@@_act_if_head_is_space:w,
%     \@@_act_if_head_is_space_true:w
%   }
% \begin{macro}[EXP]{\@@_use_none_delimit_by_q_act_stop:w}
%   To help control the expansion, \cs{@@_act:NNNn} should always
%   be preceeded by \cs{exp:w} and ends by producing \cs{exp_end:}
%   once the result has been obtained. This way no internal token of it can be
%   accidentally end up in the input stream.
%   Because \cs{s_@@_act_stop} can't appear without braces around it in the
%   argument~|#1| of \cs{@@_act_loop:w}, we can use this marker to set up a fast
%   test for leading spaces.
%    \begin{macrocode}
\cs_set_protected:Npn \@@_tmp:w #1
  {
    \cs_new:Npn \@@_act_if_head_is_space:nTF ##1
      {
        \@@_act_if_head_is_space:w
          \s_@@_act_stop ##1 \s_@@_act_stop \@@_act_if_head_is_space_true:w
          \s_@@_act_stop #1  \s_@@_act_stop \use_ii:nn
      }
    \cs_new:Npn \@@_act_if_head_is_space:w
        ##1 \s_@@_act_stop #1 ##2 \s_@@_act_stop
      {}
    \cs_new:Npn \@@_act_if_head_is_space_true:w
        \s_@@_act_stop #1 \s_@@_act_stop \use_ii:nn ##1 ##2
      {##1}
  }
\@@_tmp:w { ~ }
%    \end{macrocode}
%   (We expand the definition \cs{@@_act_if_head_is_space:nTF} when
%   setting up \cs{@@_act_loop:w}, so we can then undefine the auxiliary.)
%   In the loop, we check how the token list begins and act
%   accordingly. In the \enquote{group} case, we may have
%   reached \cs{s_@@_act_stop}, the end of the list. Then
%   leave \cs{exp_end:} and the result in the input stream,
%   to terminate the expansion of \cs{exp:w}.
%   Otherwise, apply the relevant function to the
%   \enquote{arguments}, |#3|
%   and to the head of the token list. Then repeat the loop.
%   The scheme is the same if the token list starts with an |N|-type
%   or with a space, making sure that
%   \cs{@@_act_space:wwNNN} gobbles the space.
%    \begin{macrocode}
\exp_args:Nnx \use:n { \cs_new:Npn \@@_act_loop:w #1 \s_@@_act_stop }
  {
    \exp_not:o { \@@_act_if_head_is_space:nTF {#1} }
      \exp_not:N \@@_act_space:wwNNN
      {
        \exp_not:o { \tl_if_head_is_group:nTF {#1} }
          \exp_not:N \@@_act_group:nwNNN
          \exp_not:N \@@_act_normal:NwNNN
      }
    \exp_not:n {#1} \s_@@_act_stop
  }
\cs_undefine:N \@@_act_if_head_is_space:nTF
\cs_new:Npn \@@_act_normal:NwNNN #1 #2 \s_@@_act_stop #3
  {
    #3 #1
    \@@_act_loop:w #2 \s_@@_act_stop
    #3
  }
\cs_new:Npn \@@_use_none_delimit_by_s_act_stop:w #1 \s_@@_act_stop { }
\cs_new:Npn \@@_act_end:wn #1 \@@_act_result:n #2
  { \group_align_safe_end: \exp_end: #2 }
\cs_new:Npn \@@_act_group:nwNNN #1 #2 \s_@@_act_stop #3#4#5
  {
    \@@_use_none_delimit_by_s_act_stop:w #1 \@@_act_end:wn \s_@@_act_stop
    #5 {#1}
    \@@_act_loop:w #2 \s_@@_act_stop
    #3 #4 #5
  }
\exp_last_unbraced:NNo
  \cs_new:Npn \@@_act_space:wwNNN \c_space_tl #1 \s_@@_act_stop #2#3
  {
    #3
    \@@_act_loop:w #1 \s_@@_act_stop
    #2 #3
  }
%    \end{macrocode}
%   \cs{@@_act:NNNn} loops over tokens, groups, and spaces in |#4|.
%   |{\s_@@_act_stop}| serves as the end of token list marker, the |?| after it
%   avoids losing outer braces. The result is stored as an argument for the
%   dummy function \cs{@@_act_result:n}.
%    \begin{macrocode}
\cs_new:Npn \@@_act:NNNn #1#2#3#4
  {
    \group_align_safe_begin:
    \@@_act_loop:w #4 { \s_@@_act_stop } ? \s_@@_act_stop
    #1 #3 #2
    \@@_act_result:n { }
  }
%    \end{macrocode}
%   Typically, the output is done to the right of what was already output,
%   using \cs{@@_act_output:n}, but for the \cs{@@_act_reverse} functions,
%   it should be done to the left.
%    \begin{macrocode}
\cs_new:Npn \@@_act_output:n #1 #2 \@@_act_result:n #3
  { #2 \@@_act_result:n { #3 #1 } }
\cs_new:Npn \@@_act_reverse_output:n #1 #2 \@@_act_result:n #3
  { #2 \@@_act_result:n { #1 #3 } }
%    \end{macrocode}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
%
% \begin{macro}[EXP]{\tl_reverse:n, \tl_reverse:o, \tl_reverse:V}
% \begin{macro}[EXP]{\@@_reverse_normal:nN}
% \begin{macro}[EXP]{\@@_reverse_group_preserve:nn}
% \begin{macro}[EXP]{\@@_reverse_space:n}
%   The goal here is to reverse without losing spaces nor braces.
%   This is done using the general internal function \cs{@@_act:NNNn}.
%   Spaces and \enquote{normal} tokens are output on the left of the current
%   output. Grouped tokens are output to the left but without any reversal
%   within the group.
%    \begin{macrocode}
\cs_new:Npn \tl_reverse:n #1
  {
    \__kernel_exp_not:w \exp_after:wN
      {
        \exp:w
        \@@_act:NNNn
          \@@_reverse_normal:N
          \@@_reverse_group_preserve:n
          \@@_reverse_space:
          {#1}
      }
  }
\cs_generate_variant:Nn \tl_reverse:n { o , V }
\cs_new:Npn \@@_reverse_normal:N
  { \@@_act_reverse_output:n }
\cs_new:Npn \@@_reverse_group_preserve:n #1
  { \@@_act_reverse_output:n { {#1} } }
\cs_new:Npn \@@_reverse_space:
  { \@@_act_reverse_output:n { ~ } }
%    \end{macrocode}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
%
% \begin{macro}{\tl_reverse:N, \tl_reverse:c, \tl_greverse:N, \tl_greverse:c}
%   This reverses the list, leaving \cs{exp_stop_f:} in front,
%   which stops the \texttt{f}-expansion.
%    \begin{macrocode}
\cs_new_protected:Npn \tl_reverse:N #1
  { \__kernel_tl_set:Nx #1 { \exp_args:No \tl_reverse:n { #1 } } }
\cs_new_protected:Npn \tl_greverse:N #1
  { \__kernel_tl_gset:Nx #1 { \exp_args:No \tl_reverse:n { #1 } } }
\cs_generate_variant:Nn \tl_reverse:N  { c }
\cs_generate_variant:Nn \tl_greverse:N { c }
%    \end{macrocode}
% \end{macro}
%
% \subsection{Using a single item}
%
% \begin{macro}{\tl_item:nn, \tl_item:Nn, \tl_item:cn}
% \begin{macro}{\@@_item_aux:nn, \@@_item:nn}
%   The idea here is to find the offset of the item from the left, then use
%   a loop to grab the correct item. If the resulting offset is too large,
%   then \cs{@@_if_recursion_tail_break:nN} terminates the loop, and returns
%   nothing at all.
%    \begin{macrocode}
\cs_new:Npn \tl_item:nn #1#2
  {
    \exp_args:Nf \@@_item:nn
      { \exp_args:Nf \@@_item_aux:nn { \int_eval:n {#2} } {#1} }
    #1
    \q_@@_recursion_tail
    \prg_break_point:
  }
\cs_new:Npn \@@_item_aux:nn #1#2
  {
    \int_compare:nNnTF {#1} < 0
      { \int_eval:n { \tl_count:n {#2} + 1 + #1 } }
      {#1}
  }
\cs_new:Npn \@@_item:nn #1#2
  {
    \@@_if_recursion_tail_break:nN {#2} \prg_break:
    \int_compare:nNnTF {#1} = 1
      { \prg_break:n { \exp_not:n {#2} } }
      { \exp_args:Nf \@@_item:nn { \int_eval:n { #1 - 1 } } }
  }
\cs_new:Npn \tl_item:Nn { \exp_args:No \tl_item:nn }
\cs_generate_variant:Nn \tl_item:Nn { c }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}{\tl_rand_item:n, \tl_rand_item:N, \tl_rand_item:c}
%   Importantly \cs{tl_item:nn} only evaluates its argument once.
%    \begin{macrocode}
\cs_new:Npn \tl_rand_item:n #1
  {
    \tl_if_blank:nF {#1}
      { \tl_item:nn {#1} { \int_rand:nn { 1 } { \tl_count:n {#1} } } }
  }
\cs_new:Npn \tl_rand_item:N { \exp_args:No \tl_rand_item:n }
\cs_generate_variant:Nn \tl_rand_item:N { c }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\tl_range:Nnn, \tl_range:cnn, \tl_range:nnn}
% \begin{macro}
%   {
%     \@@_range:Nnnn, \@@_range:nnnNn, \@@_range:nnNn, \@@_range_skip:w,
%     \@@_range:w, \@@_range_skip_spaces:n, \@@_range_collect:nn,
%     \@@_range_collect:ff, \@@_range_collect_space:nw,
%     \@@_range_collect_N:nN, \@@_range_collect_group:nN,
%   }
%   To avoid checking for the end of the token list at every step, start
%   by counting the number $l$ of items and \enquote{normalizing} the
%   bounds, namely clamping them to the interval $[0,l]$ and dealing
%   with negative indices.  More precisely, \cs{@@_range_items:nnNn}
%   receives the number of items to skip at the beginning of the token
%   list, the index of the last item to keep, a function which is either
%   \cs{@@_range:w} or the token list itself.  If nothing should be kept,
%   leave |{}|: this stops the \texttt{f}-expansion of \cs{tl_head:f} and that
%   function produces an empty result.  Otherwise, repeatedly call
%   \cs{@@_range_skip:w} to delete |#1| items from the input stream (the
%   extra brace group avoids an off-by-one shift).  For the braced
%   version \cs{@@_range_braced:w} sets up
%   \cs{@@_range_collect_braced:w} which stores items one by one in an
%   argument after the semicolon. Depending on the first token of the tail,
%   either just move  it (if it is a space) or also decrement the number of
%   items left to find.  Eventually, the result is a brace group followed by
%   the rest of the token list, and \cs{tl_head:f} cleans up and gives the
%   result in \cs{exp_not:n}.
%    \begin{macrocode}
\cs_new:Npn \tl_range:Nnn { \exp_args:No \tl_range:nnn }
\cs_generate_variant:Nn \tl_range:Nnn { c }
\cs_new:Npn \tl_range:nnn { \@@_range:Nnnn \@@_range:w }
\cs_new:Npn \@@_range:Nnnn #1#2#3#4
  {
    \tl_head:f
      {
        \exp_args:Nf \@@_range:nnnNn
          { \tl_count:n {#2} } {#3} {#4} #1 {#2}
      }
  }
\cs_new:Npn \@@_range:nnnNn #1#2#3
  {
    \exp_args:Nff \@@_range:nnNn
      {
        \exp_args:Nf \@@_range_normalize:nn
          { \int_eval:n { #2 - 1 } } {#1}
      }
      {
        \exp_args:Nf \@@_range_normalize:nn
          { \int_eval:n {#3} } {#1}
      }
  }
\cs_new:Npn \@@_range:nnNn #1#2#3#4
  {
    \if_int_compare:w #2 > #1 \exp_stop_f: \else:
      \exp_after:wN { \exp_after:wN }
    \fi:
    \exp_after:wN #3
    \int_value:w \int_eval:n { #2 - #1 } \exp_after:wN ;
    \exp_after:wN { \exp:w \@@_range_skip:w #1 ; { } #4 }
  }
\cs_new:Npn \@@_range_skip:w #1 ; #2
  {
    \if_int_compare:w #1 > 0 \exp_stop_f:
      \exp_after:wN \@@_range_skip:w
      \int_value:w \int_eval:n { #1 - 1 } \exp_after:wN ;
    \else:
      \exp_after:wN \exp_end:
    \fi:
  }
\cs_new:Npn \@@_range:w #1 ; #2
  {
    \exp_args:Nf \@@_range_collect:nn
      { \@@_range_skip_spaces:n {#2} } {#1}
  }
\cs_new:Npn \@@_range_skip_spaces:n #1
  {
    \tl_if_head_is_space:nTF {#1}
      { \exp_args:Nf \@@_range_skip_spaces:n {#1} }
      { { } #1 }
  }
\cs_new:Npn \@@_range_collect:nn #1#2
  {
    \int_compare:nNnTF {#2} = 0
      {#1}
      {
        \exp_args:No \tl_if_head_is_space:nTF { \use_none:n #1 }
          {
            \exp_args:Nf \@@_range_collect:nn
              { \@@_range_collect_space:nw #1 }
              {#2}
          }
          {
            \@@_range_collect:ff
              {
                \exp_args:No \tl_if_head_is_N_type:nTF { \use_none:n #1 }
                  { \@@_range_collect_N:nN }
                  { \@@_range_collect_group:nn }
                #1
              }
              { \int_eval:n { #2 - 1 } }
          }
      }
  }
\cs_new:Npn \@@_range_collect_space:nw #1 ~ { { #1 ~ } }
\cs_new:Npn \@@_range_collect_N:nN #1#2 { { #1 #2 } }
\cs_new:Npn \@@_range_collect_group:nn #1#2 { { #1 {#2} } }
\cs_generate_variant:Nn \@@_range_collect:nn { ff }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}[EXP]{\@@_range_normalize:nn}
%   This function converts an \meta{index} argument into an explicit
%   position in the token list (a result of $0$ denoting \enquote{out of
%   bounds}).  Expects two explicit integer arguments: the \meta{index}
%   |#1| and the string count~|#2|.  If |#1| is negative, replace it by
%   $|#1| + |#2| + 1$, then limit to the range $[0, |#2|]$.
%    \begin{macrocode}
\cs_new:Npn \@@_range_normalize:nn #1#2
  {
    \int_eval:n
      {
        \if_int_compare:w #1 < 0 \exp_stop_f:
          \if_int_compare:w #1 < -#2 \exp_stop_f:
            0
          \else:
            #1 + #2 + 1
          \fi:
        \else:
          \if_int_compare:w #1 < #2 \exp_stop_f:
            #1
          \else:
            #2
          \fi:
        \fi:
      }
  }
%    \end{macrocode}
% \end{macro}
%
% \subsection{Viewing token lists}
%
% \begin{macro}{\tl_show:N, \tl_show:c, \tl_log:N, \tl_log:c, \@@_show:NN}
%   Showing token list variables is done after checking that the
%   variable is defined (see \cs{__kernel_register_show:N}).
%    \begin{macrocode}
\cs_new_protected:Npn \tl_show:N { \@@_show:NN \tl_show:n }
\cs_generate_variant:Nn \tl_show:N { c }
\cs_new_protected:Npn \tl_log:N { \@@_show:NN \tl_log:n }
\cs_generate_variant:Nn \tl_log:N { c }
\cs_new_protected:Npn \@@_show:NN #1#2
  {
    \__kernel_chk_defined:NT #2
      {
        \exp_args:Nf \tl_if_empty:nTF
          { \cs_prefix_spec:N #2 \cs_argument_spec:N #2 }
          {
            \exp_args:Ne #1
              { \token_to_str:N #2 = \__kernel_exp_not:w \exp_after:wN {#2} }
          }
          {
            \__kernel_msg_error:nnxxx { kernel } { bad-type }
              { \token_to_str:N #2 } { \token_to_meaning:N #2 } { tl }
          }
      }
  }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\tl_show:n, \@@_show:n}
% \begin{macro}[EXP]{\@@_show:w}
%   Many |show| functions are based on \cs{tl_show:n}.
%   The argument of \cs{tl_show:n} is line-wrapped using
%   \cs{iow_wrap:nnnN} but with a leading |>~| and trailing period, both
%   removed before passing the wrapped text to the \tn{showtokens}
%   primitive.  This primitive shows the result with a leading |>~| and
%   trailing period.
%
%   The token list \cs{l_@@_internal_a_tl} containing the result
%   of all these manipulations is displayed to the terminal using
%   \cs{tex_showtokens:D} and an odd \cs{exp_after:wN} which expand the
%   closing brace to improve the output slightly.  The calls to
%   \cs{__kernel_iow_with:Nnn} ensure that the \tn{newlinechar} is set to~$10$
%   so that the \cs{iow_newline:} inserted by the line-wrapping code
%   are correctly recognized by \TeX{}, and that \tn{errorcontextlines}
%   is $-1$ to avoid printing irrelevant context.
%    \begin{macrocode}
\cs_new_protected:Npn \tl_show:n #1
  { \iow_wrap:nnnN { >~ \tl_to_str:n {#1} . } { } { } \@@_show:n }
\cs_new_protected:Npn \@@_show:n #1
  {
    \tl_set:Nf \l_@@_internal_a_tl { \@@_show:w #1 \s_@@_stop }
    \__kernel_iow_with:Nnn \tex_newlinechar:D { 10 }
      {
        \__kernel_iow_with:Nnn \tex_errorcontextlines:D { -1 }
          {
            \tex_showtokens:D \exp_after:wN \exp_after:wN \exp_after:wN
              { \exp_after:wN \l_@@_internal_a_tl }
          }
      }
  }
\cs_new:Npn \@@_show:w #1 > #2 . \s_@@_stop {#2}
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}{\tl_log:n}
%   Logging is much easier, simply line-wrap.  The |>~| and trailing
%   period is there to match the output of \cs{tl_show:n}.
%    \begin{macrocode}
\cs_new_protected:Npn \tl_log:n #1
  { \iow_wrap:nnnN { > ~ \tl_to_str:n {#1} . } { } { } \iow_log:n }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\__kernel_chk_tl_type:NnnT}
%   Helper for checking that |#1| has the correct internal structure to
%   be of a certain type.  Make sure that it is defined and that it is a
%   token list, namely a macro with no \tn{long} nor \tn{protected}
%   prefix.  Then compare |#1| to an attempt at reconstructing a valid
%   structure of the given type using |#2| (see implementation of
%   \cs{seq_show:N} for instance).  If that is successful run the
%   requested code~|#4|.
%    \begin{macrocode}
\cs_new_protected:Npn \__kernel_chk_tl_type:NnnT #1#2#3#4
  {
    \__kernel_chk_defined:NT #1
      {
        \exp_args:Nf \tl_if_empty:nTF
          { \cs_prefix_spec:N #1 \cs_argument_spec:N #1 }
          {
            \tl_set:Nx \l_@@_internal_a_tl {#3}
            \tl_if_eq:NNTF #1 \l_@@_internal_a_tl
              {#4}
              {
                \__kernel_msg_error:nnxxxx { kernel } { bad-type }
                  { \token_to_str:N #1 } { \tl_to_str:N #1 }
                  {#2} { \tl_to_str:N \l_@@_internal_a_tl }
              }
          }
          {
            \__kernel_msg_error:nnxxx { kernel } { bad-type }
              { \token_to_str:N #1 } { \token_to_meaning:N #1 } {#2}
          }
      }
  }
%    \end{macrocode}
% \end{macro}
%
% \subsection{Internal scan marks}
%
% \begin{variable}{\s_@@_nil,\s_@@_mark,\s_@@_stop}
%   Internal scan marks. These are defined here at the end because the
%   code for \cs{scan_new:N} depends on some \pkg{l3tl} functions.
%    \begin{macrocode}
\scan_new:N \s_@@_nil
\scan_new:N \s_@@_mark
\scan_new:N \s_@@_stop
%    \end{macrocode}
% \end{variable}
%
% \subsection{Scratch token lists}
%
% \begin{variable}{\g_tmpa_tl, \g_tmpb_tl}
%    Global temporary token list variables.
%    They are supposed to be set and used immediately,
%    with no delay between the definition and the use because you
%    can't count on other macros not to redefine them from under you.
%    \begin{macrocode}
\tl_new:N \g_tmpa_tl
\tl_new:N \g_tmpb_tl
%    \end{macrocode}
% \end{variable}
%
% \begin{variable}{\l_tmpa_tl, \l_tmpb_tl}
%    These are local temporary token list variables. Be sure not to assume
%    that the value you put into them will survive for
%    long---see discussion above.
%    \begin{macrocode}
\tl_new:N \l_tmpa_tl
\tl_new:N \l_tmpb_tl
%    \end{macrocode}
% \end{variable}
%
% We finally clean up a temporary control sequence that we have used at
% various points to set up some definitions.
%    \begin{macrocode}
\cs_undefine:N \@@_tmp:w
%    \end{macrocode}
%
%    \begin{macrocode}
%</package>
%    \end{macrocode}
%
% \end{implementation}
%
% \PrintIndex