summaryrefslogtreecommitdiff
path: root/Master/texmf-dist/source/latex/l3kernel/l3expan.dtx
blob: 9697c45991bd3a7af1fb88f69def81dd9a8bc89a (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
% \iffalse meta-comment
%
%% File: l3expan.dtx Copyright (C) 1990-2017 The LaTeX3 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
%
%    http://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]{l3doc}
\def\ExplFileDate{2017/03/18}
\def\ExplFileVersion{7019}
\begin{document}
  \DocInput{\jobname.dtx}
\end{document}
%</driver>
% \fi
%
% \title{^^A
%   The \pkg{l3expan} package\\ Argument expansion^^A
%   \thanks{This file describes v\ExplFileVersion,
%      last revised \ExplFileDate.}^^A
% }
%
% \author{^^A
%  The \LaTeX3 Project\thanks
%    {^^A
%      E-mail:
%        \href{mailto:latex-team@latex-project.org}
%          {latex-team@latex-project.org}^^A
%    }^^A
% }
%
% \date{Released \ExplFileDate}
%
% \maketitle
%
% \begin{documentation}
%
% This module provides generic methods for expanding \TeX{} arguments in a
% systematic manner. The functions in this module all have prefix |exp|.
%
% Not all possible variations are implemented for every base
% function. Instead only those that are used within the \LaTeX3 kernel
% or otherwise seem to be of general interest are implemented.
% Consult the module description to find out which functions are
% actually defined. The next section explains how to define missing
% variants.
%
% \section{Defining new variants}
% \label{sec:l3expan:defining-variants}
%
% The definition of variant forms for base functions may be necessary
% when writing new functions or when applying a kernel function in a
% situation that we haven't thought of before.
%
% Internally preprocessing of arguments is done with functions from the
% |\exp_| module.  They all look alike, an example would be
% \cs{exp_args:NNo}. This function has three arguments, the first and the
% second are a single tokens, while the third argument should be given
% in braces. Applying \cs{exp_args:NNo} will expand the content of third
% argument once before any expansion of the first and second arguments.
% If \cs{seq_gpush:No} was not defined it could be coded in the following way:
% \begin{verbatim}
%   \exp_args:NNo \seq_gpush:Nn
%      \g_file_name_stack
%      \l_tmpa_tl
% \end{verbatim}
% In other words, the first argument to \cs{exp_args:NNo} is the base
% function and the other arguments are preprocessed and then passed to
% this base function. In the example the first argument to the base
% function should be a single token which is left unchanged while the
% second argument is expanded once. From this example we can also see
% how the variants are defined. They just expand into the appropriate
% |\exp_| function followed by the desired base function, \emph{e.g.}
% \begin{quote}
%   |\cs_generate_variant:Nn \seq_gpush:Nn { No } |
% \end{quote}
% results in the definition of |\seq_gpush:No|
% \begin{quote}
%   |\cs_new:Npn \seq_gpush:No { \exp_args:NNo \seq_gpush:Nn }|
% \end{quote}
% Providing variants in this way in style files is uncritical as the
% \cs{cs_generate_variant:Nn} function will only create new definitions if
% there is not already one available. Therefore adding
% such definition to later releases of the kernel will not make such
% style files obsolete.
%
% The steps above may be automated by using the function
% \cs{cs_generate_variant:Nn}, described next.
%
% \section{Methods for defining variants}
% \label{sec:l3expan:variants-method}
%
% ^^A Bruno: Should we put a table of variant types (N, c, n, V, v, o, f, x) for reference here?
%
% ^^A Bruno: Should \cs_generate_variant:Nn \foo:n { c } and \foo:N { o } really be ok?
%
% \begin{function}[updated = 2015-08-06]{\cs_generate_variant:Nn}
%   \begin{syntax}
%     \cs{cs_generate_variant:Nn} \meta{parent control sequence} \Arg{variant argument specifiers}
%   \end{syntax}
%   This function is used to define argument-specifier variants of the
%   \meta{parent control sequence} for \LaTeX3 code-level macros. The
%   \meta{parent control sequence} is first separated into the
%   \meta{base name} and \meta{original argument specifier}. The
%   comma-separated list of \meta{variant argument specifiers} is
%   then used to define variants of the
%   \meta{original argument specifier} where these are not already
%   defined. For each \meta{variant} given, a function is created
%   which will expand its arguments as detailed and pass them
%   to the \meta{parent control sequence}. So for example
%   \begin{verbatim}
%     \cs_set:Npn \foo:Nn #1#2 { code here }
%     \cs_generate_variant:Nn \foo:Nn { c }
%   \end{verbatim}
%   will create a new function |\foo:cn| which will expand its first
%   argument into a control sequence name and pass the result to
%   |\foo:Nn|. Similarly
%   \begin{verbatim}
%     \cs_generate_variant:Nn \foo:Nn { NV , cV }
%   \end{verbatim}
%   would generate the functions |\foo:NV| and |\foo:cV| in the same
%   way. The \cs{cs_generate_variant:Nn} function can only be applied if
%   the \meta{parent control sequence} is already defined. Only |n|~and
%   |N| arguments can be changed to other types. If the \meta{parent
%   control sequence} is protected or if the \meta{variant} involves
%   |x|~arguments, then the \meta{variant control sequence} will also be
%   protected.  The \meta{variant} is created globally, as is any
%   \cs[no-index]{exp_args:N\meta{variant}} function needed to carry out the
%   expansion.
% \end{function}
%
% \section{Introducing the variants}
%
% The available internal functions for argument expansion come in two
% flavours, some of them are faster then others. Therefore
% (when speed is important) it is usually
% best to follow the following guidelines when defining new functions
% that are supposed to come with variant forms:
% \begin{itemize}
%   \item
%     Arguments that might need expansion should come first in the list of
%     arguments to make processing faster.
%   \item
%     Arguments that should consist of single tokens should come first.
%   \item
%     Arguments that need full expansion (\emph{i.e.}, are denoted
%     with |x|) should be avoided if possible as they can not be
%     processed expandably, \emph{i.e.}, functions of this type will
%     not work correctly in arguments that are themselves subject to |x|
%     expansion.
%   \item
%     In general, unless in the last position, multi-token arguments
%     |n|, |f|, and |o| will need special processing when more than
%     one argument is being expanded. This special processing is not fast.
%     Therefore it is best to use the optimized functions, namely
%     those that contain only |N|, |c|, |V|, and |v|, and, in the last
%     position, |o|, |f|, with possible trailing |N| or |n|, which are
%     not expanded.
% \end{itemize}
%
% The |V| type returns the value of a register, which can be one of
% |tl|, |int|, |skip|, |dim|, |toks|, or built-in \TeX{}
% registers. The |v| type is the same except it first creates a
% control sequence out of its argument before returning the
% value.
%
% In general, the programmer should not need to be concerned with
% expansion control. When simply using the content of a variable,
% functions with a |V| specifier should be used. For those referred to by
% (cs)name, the |v| specifier is available for the same purpose. Only when
% specific expansion steps are needed, such as when using delimited
% arguments, should the lower-level functions with |o| specifiers be employed.
%
% The |f| type is so special that it deserves an example.  It is
% typically used in contexts where only expandable commands are allowed.
% Then |x|-expansion cannot be used, and |f|-expansion provides an
% alternative that expands as much as can be done in such contexts.  For
% instance, say that we want to evaluate the integer expression $3 + 4$
% and pass the result $7$ as an argument to an expandable function
% |\example:n|.  For this, one should define a variant using
% \cs{cs_generate_variant:Nn} |\example:n| |{| |f| |}|, then do
% \begin{quote}
%   |\example:f { \int_eval:n { 3 + 4 } }|
% \end{quote}
% Note that |x|-expansion would also expand \cs{int_eval:n} fully to its
% result~$7$, but the variant |\example:x| cannot be expandable.  Note
% also that |o|-expansion would not expand \cs{int_eval:n} fully to its
% result since that function requires several expansions.  Besides the
% fact that |x|-expansion is protected rather than expandable, another
% difference between |f|-expansion and |x|-expansion is that
% |f|-expansion expands tokens from the beginning and stops as soon as a
% non-expandable token is encountered, while |x|-expansion continues
% expanding further tokens.  Thus, for instance
% \begin{quote}
%   |\example:f { \int_eval:n { 1 + 2 } , \int_eval:n { 3 + 4 } }|
% \end{quote}
% will result in the call |\example:n { 3 , \int_eval:n { 3 + 4 } }|
% while using |\example:x| instead results in |\example:n { 3 , 7 }|
% at the cost of being protected.
% If you use this type of expansion in conditional processing then
% you should stick to using |TF|  type functions only as it does not
% try to finish any |\if... \fi:| itself!
%
% If is important to note that both \texttt{f}- and \texttt{o}-type
% expansion are concerned with the expansion of tokens from left to
% right in their arguments. In particular, \texttt{o}-type expansion
% applies to the first \emph{token} in the argument it receives: it
% is conceptually similar to
% \begin{verbatim}
%   \exp_after:wN <base function> \exp_after:wN { <argument> }
% \end{verbatim}
% At the same time, \texttt{f}-type expansion stops at the emph{first}
% non-expandable token. This means for example that both
% \begin{verbatim}
%    \tl_set:No \l_tmpa_tl { { \g_tmpb_tl } }
% \end{verbatim}
% and
% \begin{verbatim}
%    \tl_set:Nf \l_tmpa_tl { { \g_tmpb_tl } }
% \end{verbatim}
% leave |\g_tmpb_tl| unchanged: |{| is the first token in the
% argument and is non-expandable.
%
% \section{Manipulating the first argument}
%
% These functions are described in detail: expansion of multiple tokens follows
% the same rules but is described in a shorter fashion.
%
% \begin{function}[EXP]{\exp_args:No}
%   \begin{syntax}
%     \cs{exp_args:No} \meta{function} \Arg{tokens} ...
%   \end{syntax}
%   This function absorbs two arguments (the \meta{function} name and
%   the \meta{tokens}). The \meta{tokens} are expanded once, and the result
%   is inserted in braces into the input stream \emph{after} reinsertion
%   of the \meta{function}. Thus the \meta{function} may take more than
%   one argument: all others will be left unchanged.
% \end{function}
%
% \begin{function}[EXP]{\exp_args:Nc, \exp_args:cc}
%   \begin{syntax}
%     \cs{exp_args:Nc} \meta{function} \Arg{tokens}
%   \end{syntax}
%   This function absorbs two arguments (the \meta{function} name and
%   the \meta{tokens}). The \meta{tokens} are expanded until only characters
%   remain, and are then turned into a control sequence. (An internal error
%   will occur if such a conversion is not possible). The result
%   is inserted into the input stream \emph{after} reinsertion
%   of the \meta{function}. Thus the \meta{function} may take more than
%   one argument: all others will be left unchanged.
%
%   The |:cc| variant constructs the \meta{function} name in the same
%   manner as described for the \meta{tokens}.
% \end{function}
%
% \begin{function}[EXP]{\exp_args:NV}
%   \begin{syntax}
%     \cs{exp_args:NV} \meta{function} \meta{variable}
%   \end{syntax}
%   This function absorbs two arguments (the names of the \meta{function} and
%   the \meta{variable}). The content of the \meta{variable} are recovered
%   and placed inside braces into the input stream \emph{after} reinsertion
%   of the \meta{function}. Thus the \meta{function} may take more than
%   one argument: all others will be left unchanged.
% \end{function}
%
% \begin{function}[EXP]{\exp_args:Nv}
%   \begin{syntax}
%     \cs{exp_args:Nv} \meta{function} \Arg{tokens}
%   \end{syntax}
%   This function absorbs two arguments (the \meta{function} name and
%   the \meta{tokens}). The \meta{tokens} are expanded until only characters
%   remain, and are then turned into a control sequence. (An internal error
%   will occur if such a conversion is not possible). This control sequence
%   should
%   be the name of a \meta{variable}.   The content of the \meta{variable} are
%   recovered and placed inside braces into the input stream \emph{after}
%   reinsertion of the \meta{function}. Thus the \meta{function} may take more
%   than one argument: all others will be left unchanged.
% \end{function}
%
% \begin{function}[EXP]{\exp_args:Nf}
%   \begin{syntax}
%     \cs{exp_args:Nf} \meta{function} \Arg{tokens}
%   \end{syntax}
%   This function absorbs two arguments (the \meta{function} name and
%   the \meta{tokens}). The \meta{tokens} are fully expanded until the
%   first non-expandable token or space is found, and the result
%   is inserted in braces into the input stream \emph{after} reinsertion
%   of the \meta{function}. Thus the \meta{function} may take more than
%   one argument: all others will be left unchanged.
% \end{function}
%
% \begin{function}{\exp_args:Nx}
%   \begin{syntax}
%     \cs{exp_args:Nx} \meta{function} \Arg{tokens}
%   \end{syntax}
%   This function absorbs two arguments (the \meta{function} name and
%   the \meta{tokens}) and exhaustively expands the \meta{tokens}
%   second. The result is inserted in braces into the input stream
%   \emph{after} reinsertion of the \meta{function}.
%   Thus the \meta{function} may take more
%   than one argument: all others will be left unchanged.
% \end{function}
%
% \section{Manipulating two arguments}
%
% \begin{function}[EXP]
%   {
%     \exp_args:NNo,
%     \exp_args:NNc,
%     \exp_args:NNv,
%     \exp_args:NNV,
%     \exp_args:NNf,
%     \exp_args:Nco,
%     \exp_args:Ncf,
%     \exp_args:Ncc,
%     \exp_args:NVV
%   }
%   \begin{syntax}
%     \cs{exp_args:NNc} \meta{token_1} \meta{token_2} \Arg{tokens}
%   \end{syntax}
%   These optimized functions absorb three arguments and expand the second and
%   third as detailed by their argument specifier. The first argument
%   of the function is then the next item on the input stream, followed
%   by the expansion of the second and third arguments.
% \end{function}
%
% \begin{function}[EXP, updated = 2012-01-14]
%   {
%     \exp_args:Nno,
%     \exp_args:NnV,
%     \exp_args:Nnf,
%     \exp_args:Noo,
%     \exp_args:Nof,
%     \exp_args:Noc,
%     \exp_args:Nff,
%     \exp_args:Nfo,
%     \exp_args:Nnc
%   }
%   \begin{syntax}
%     \cs{exp_args:Noo} \meta{token} \Arg{tokens_1} \Arg{tokens_2}
%   \end{syntax}
%   These functions absorb three arguments and expand the second and
%   third as detailed by their argument specifier. The first argument
%   of the function is then the next item on the input stream, followed
%   by the expansion of the second and third arguments.
%   These functions need special (slower) processing.
% \end{function}
%
% \begin{function}
%   {
%     \exp_args:NNx,
%     \exp_args:Nnx,
%     \exp_args:Ncx,
%     \exp_args:Nox,
%     \exp_args:Nxo,
%     \exp_args:Nxx
%   }
%   \begin{syntax}
%     \cs{exp_args:NNx} \meta{token_1} \meta{token_2} \Arg{tokens}
%   \end{syntax}
%   These functions absorb three arguments and expand the second and
%   third as detailed by their argument specifier. The first argument
%   of the function is then the next item on the input stream, followed
%   by the expansion of the second and third arguments. These functions
%   are not expandable.
% \end{function}
%
% \section{Manipulating three arguments}
%
% \begin{function}[EXP]
%   {
%     \exp_args:NNNo,
%     \exp_args:NNNV,
%     \exp_args:Nccc,
%     \exp_args:NcNc,
%     \exp_args:NcNo,
%     \exp_args:Ncco
%   }
%   \begin{syntax}
%     \cs{exp_args:NNNo} \meta{token_1} \meta{token_2} \meta{token_3} \Arg{tokens}
%   \end{syntax}
%   These optimized functions absorb four arguments and expand the second, third
%   and fourth as detailed by their argument specifier. The first
%   argument of the function is then the next item on the input stream,
%   followed by the expansion of the second argument, \emph{etc}.
% \end{function}
%
% \begin{function}[EXP]
%   {
%     \exp_args:NNoo,
%     \exp_args:NNno,
%     \exp_args:Nnno,
%     \exp_args:Nnnc,
%     \exp_args:Nooo,
%   }
%   \begin{syntax}
%     \cs{exp_args:NNoo} \meta{token_1} \meta{token_2} \Arg{token_3} \Arg{tokens}
%   \end{syntax}
%   These functions absorb four arguments and expand the second, third
%   and fourth as detailed by their argument specifier. The first
%   argument of the function is then the next item on the input stream,
%   followed by the expansion of the second argument, \emph{etc}.
%   These functions need special (slower) processing.
% \end{function}
%
% \begin{function}[added = 2015-08-12]
%   {
%     \exp_args:NNNx,
%     \exp_args:NNnx,
%     \exp_args:NNox,
%     \exp_args:Nnnx,
%     \exp_args:Nnox,
%     \exp_args:Noox,
%     \exp_args:Ncnx,
%     \exp_args:Nccx
%   }
%   \begin{syntax}
%     \cs{exp_args:NNnx} \meta{token_1} \meta{token_2} \Arg{tokens_1} \Arg{tokens_2}
%   \end{syntax}
%   These functions absorb four arguments and expand the second, third
%   and fourth as detailed by their argument specifier. The first
%   argument of the function is then the next item on the input stream,
%   followed by the expansion of the second argument, \emph{etc.}
% \end{function}
%
% \section{Unbraced expansion}
%
% \begin{function}[EXP, updated = 2012-02-12]
%   {
%     \exp_last_unbraced:NV,
%     \exp_last_unbraced:Nf,
%     \exp_last_unbraced:No,
%     \exp_last_unbraced:Nv,
%     \exp_last_unbraced:Nco,
%     \exp_last_unbraced:NcV,
%     \exp_last_unbraced:NNV,
%     \exp_last_unbraced:NNo,
%     \exp_last_unbraced:Nno,
%     \exp_last_unbraced:Noo,
%     \exp_last_unbraced:Nfo,
%     \exp_last_unbraced:NNNV,
%     \exp_last_unbraced:NNNo,
%     \exp_last_unbraced:NnNo
%     }
%   \begin{syntax}
%     \cs{exp_last_unbraced:Nno} \meta{token} \meta{tokens_1} \meta{tokens_2}
%   \end{syntax}
%   These functions absorb the number of arguments given by their
%   specification, carry out the expansion
%   indicated and leave the results in the input stream, with the
%   last argument not surrounded by the usual braces.
%   Of these, the \texttt{:Nno}, \texttt{:Noo}, and \texttt{:Nfo}
%   variants need special (slower) processing.
%   \begin{texnote}
%     As an optimization, the last argument is unbraced by some
%     of those functions before expansion. This can cause problems
%     if the argument is empty: for instance,
%     \cs{exp_last_unbraced:Nf} |\foo_bar:w| |{ }| \cs{q_stop}
%     leads to an infinite loop, as the quark is \texttt{f}-expanded.
%   \end{texnote}
% \end{function}
%
% \begin{function}{\exp_last_unbraced:Nx}
%   \begin{syntax}
%     \cs{exp_last_unbraced:Nx} \meta{function} \Arg{tokens}
%   \end{syntax}
%   This functions fully expands the \meta{tokens} and leaves
%   the result in the input stream after reinsertion of \meta{function}.
%   This function is not expandable.
% \end{function}
%
% \begin{function}[EXP]{\exp_last_two_unbraced:Noo}
%   \begin{syntax}
%     \cs{exp_last_two_unbraced:Noo} \meta{token} \meta{tokens_1} \Arg{tokens_2}
%   \end{syntax}
%   This function absorbs three arguments and expand the second and third
%   once. The first argument of the function is then the next item on the
%   input stream, followed by the expansion of the second and third arguments,
%   which are not wrapped in braces.
%   This function needs special (slower) processing.
% \end{function}
%
% \begin{function}[EXP]{\exp_after:wN}
%   \begin{syntax}
%     \cs{exp_after:wN} \meta{token_1} \meta{token_2}
%   \end{syntax}
%   Carries out a single expansion of \meta{token_2} (which may consume
%   arguments) prior to the expansion
%   of \meta{token_1}. If \meta{token_2} is a \TeX{} primitive, it will
%   be executed rather than expanded, while if \meta{token_2} has not
%   expansion (for example, if it is a character) then it will be left
%   unchanged. It is important to notice that \meta{token_1} may be
%   \emph{any} single token, including group-opening and -closing
%   tokens (|{| or |}| assuming normal \TeX{} category codes). Unless
%   specifically required, expansion should be carried out using an
%   appropriate argument specifier variant or the appropriate
%   \cs{exp_arg:N} function.
%   \begin{texnote}
%     This is the \TeX{} primitive \tn{expandafter} renamed.
%   \end{texnote}
% \end{function}
%
%
% \section{Preventing expansion}
%
% Despite the fact that the following functions are all about preventing
% expansion, they're designed to be used in an expandable context and hence
% are all marked as being `expandable' since they themselves will not appear
% after the expansion has completed.
%
% \begin{function}[EXP]{\exp_not:N}
%   \begin{syntax}
%     \cs{exp_not:N} \meta{token}
%   \end{syntax}
%   Prevents expansion of the \meta{token} in a context where it would otherwise
%   be expanded, for example an |x|-type argument.
%   \begin{texnote}
%     This is the \TeX{} \tn{noexpand} primitive.
%   \end{texnote}
% \end{function}
%
% \begin{function}[EXP]{\exp_not:c}
%   \begin{syntax}
%     \cs{exp_not:c} \Arg{tokens}
%   \end{syntax}
%   Expands the \meta{tokens} until only unexpandable content remains, and then
%   converts this into a control sequence. Further expansion of this control
%   sequence is then inhibited.
% \end{function}
%
% \begin{function}[EXP]{\exp_not:n}
%   \begin{syntax}
%     \cs{exp_not:n} \Arg{tokens}
%   \end{syntax}
%   Prevents expansion of the \meta{tokens} in a context where they
%   would otherwise be expanded, for example an |x|-type argument.
%   \begin{texnote}
%     This is the \eTeX{} \tn{unexpanded} primitive.  Hence its argument
%     \emph{must} be surrounded by braces.
%   \end{texnote}
% \end{function}
%
% \begin{function}[EXP]{\exp_not:V}
%   \begin{syntax}
%     \cs{exp_not:V} \meta{variable}
%   \end{syntax}
%   Recovers the content of the \meta{variable}, then prevents expansion
%   of this material in a context where it would otherwise
%   be expanded, for example an |x|-type argument.
% \end{function}
%
% \begin{function}[EXP]{\exp_not:v}
%   \begin{syntax}
%     \cs{exp_not:v} \Arg{tokens}
%   \end{syntax}
%   Expands the \meta{tokens} until only unexpandable content remains, and then
%   converts this into a control sequence (which should be a \meta{variable}
%   name). The content of the \meta{variable} is recovered, and further
%   expansion is prevented in a context where it would otherwise
%   be expanded, for example an |x|-type argument.
% \end{function}
%
% \begin{function}[EXP]{\exp_not:o}
%   \begin{syntax}
%     \cs{exp_not:o} \Arg{tokens}
%   \end{syntax}
%   Expands the \meta{tokens} once, then prevents any further expansion in a
%   context where they would otherwise
%   be expanded, for example an |x|-type argument.
% \end{function}
%
% \begin{function}[EXP]{\exp_not:f}
%   \begin{syntax}
%     \cs{exp_not:f} \Arg{tokens}
%   \end{syntax}
%   Expands \meta{tokens} fully until the first unexpandable token
%   is found. Expansion then stops, and the result of the expansion
%   (including any tokens which were not expanded) is protected from
%   further expansion.
% \end{function}
%
% \begin{function}[updated = 2011-06-03, EXP]{\exp_stop_f:}
%   \begin{syntax}
%     |\foo_bar:f| \{ \meta{tokens} \cs{exp_stop_f:} \meta{more tokens} \}
%   \end{syntax}
%   This function terminates an \texttt{f}-type expansion. Thus if
%   a function |\foo_bar:f| starts an \texttt{f}-type expansion
%   and all of \meta{tokens} are expandable \cs{exp_stop_f:} will
%   terminate the expansion of tokens even if \meta{more tokens}
%   are also expandable. The function itself is an implicit space
%   token. Inside an \texttt{x}-type expansion, it will retain its
%   form, but when typeset it produces the underlying space (\verb*| |).
% \end{function}
%
%
% \section{Controlled expansion}
%
% The \pkg{expl3} language makes all efforts to hide the complexity of
% \TeX{} expansion from the programmer by providing concepts that
% evaluate/expand arguments of functions prior to calling the \enquote{base}
% functions. Thus, instead of using many \tn{expandafter} calls and
% other trickery it is usually a matter of choosing the right variant
% of a function to achieve a desired result.
%
% Of course, deep down \TeX{} is using expansion as always and there
% are cases where a programmer needs to control that expansion
% directly; typical situations are basic data manipulation tools. This
% section documents the functions for that level. You will find these
% commands used throughout the kernel code, but we hope that outside
% the kernel there will be little need to resort to them. Instead the
% argument manipulation methods document above should usually be sufficient.
%
% While \cs{exp_after:wN} expands one token (out of order) it is
% sometimes necessary to expand several tokens in one go. The next set
% of commands provide this functionality. Be aware that it is
% absolutely required that the programmer has full control over the
% tokens to be expanded, i.e., it is not possible to use these
% functions to expand unknown input as part of
% \meta{expandable-tokens} as that will break badly if unexpandable
% tokens are encountered in that place!
%
% \begin{function}[added=2015-08-23,EXP]
%       {
%         \exp:w ,
%         \exp_end:
%       }
%   \begin{syntax}
%     \cs{exp:w} \meta{expandable-tokens} \cs{exp_end:} \\
%   \end{syntax}
%   Expands \meta{expandable-tokens} until reaching \cs{exp_end:} at
%   which point expansion stops.
%   The full expansion of \meta{expandable-tokens} has to be empty.
%   If any token in \meta{expandable-tokens} or any token generated by
%   expanding the tokens therein is not expandable the expansion will end
%   prematurely and as a result \cs{exp_end:} will be misinterpreted
%   later on.\footnotemark
%
%   In typical use cases the \cs{exp_end:} will be  hidden somewhere
%   in the replacement text of \meta{expandable-tokens} rather than
%   being on the same expansion level than \cs{exp:w}, e.g., you may
%   see code such as
%\begin{verbatim}
%    \exp:w \@@_case:NnTF #1 {#2} { } { }
%\end{verbatim}
%   where somewhere during the expansion of |\@@_case:NnTF| the
%   \cs{exp_end:} gets generated.
% \end{function}
% \footnotetext{Due to the implementation you might get the character
%   in position 0 in the current font (typically \enquote{\texttt{`}})
%   in the output without any error message!}
%
% \begin{function}[added=2015-08-23, EXP, label = \exp_end_continue_f:w]
%       {
%         \exp:w ,
%         \exp_end_continue_f:w
%       }
%   \begin{syntax}
%     \cs{exp:w} \meta{expandable-tokens} \cs{exp_end_continue_f:w} \meta{further-tokens}
%   \end{syntax}
%   Expands \meta{expandable-tokens} until reaching \cs{exp_end_continue_f:w} at
%   which point expansion continues as an f-type expansion expanding
%   \meta{further-tokens} until an unexpandable token is encountered (or
%   the f-type expansion is explicitly terminated by
%   \cs{exp_stop_f:}). As with all f-type expansions a space ending
%   the expansion will get removed.
%
%   The full expansion of \meta{expandable-tokens} has to be empty.
%   If any token in \meta{expandable-tokens} or any token generated by
%   expanding the tokens therein is not expandable the expansion will end
%   prematurely and as a result \cs{exp_end_continue_f:w} will be misinterpreted
%   later on.\footnotemark
%
%
%   In typical use cases  \meta{expandable-tokens} contains no tokens at all,
%   e.g., you will see code such as
%\begin{verbatim}
%    \exp_after:wN { \exp:w \exp_end_continue_f:w #2 }
%\end{verbatim}
%   where the \cs{exp_after:wN} triggers an f-expansion of the tokens
%   in |#2|. For technical reasons this has to happen using two tokens
%   (if they would be hidden inside another command \cs{exp_after:wN}
%   would only expand the command but not trigger any additional
%    f-expansion).
%
%   You might wonder why there are two different approaches available,
%   after all the effect of
%   \begin{quote}
%     \cs{exp:w} \meta{expandable-tokens} \cs{exp_end:}
%   \end{quote}
%   can be alternatively achieved through an f-type expansion by using
%   \cs{exp_stop_f:}, i.e.
%   \begin{quote}
%     \cs{exp:w} \cs{exp_end_continue_f:w} \meta{expandable-tokens} \cs{exp_stop_f:}
%   \end{quote}
%   The reason is simply that the first approach is slightly faster
%   (one less token to parse and less expansion internally)
%   so in places where such performance really matters and where we
%   want to explicitly stop the expansion at a defined point the first
%   form is preferable.
% \end{function}
% \footnotetext{In this particular case you may get a character into
%    the output as well as an error message.}
%
% \begin{function}[added=2015-08-23, EXP, label = \exp_end_continue_f:nw]
%       {
%         \exp:w ,
%         \exp_end_continue_f:nw
%       }
%   \begin{syntax}
%     \cs{exp:w} \meta{expandable-tokens} \cs{exp_end_continue_f:nw} \meta{further-tokens}
%   \end{syntax}
%   The difference to \cs{exp_end_continue_f:w} is that we first we pick
%   up an argument which is then returned to the input stream. If
%   \meta{further-tokens} starts with a brace group then the braces
%   are removed. If on the other hand it starts with space tokens then
%   these space tokens are removed while searching for the
%   argument. Thus such space tokens will not terminate the f-type expansion.
% \end{function}
%
% \section{Internal functions and variables}
%
% \begin{variable}{\l__exp_internal_tl}
%   The |\exp_| module has its private variables to temporarily store
%   results of the argument expansion. This is done to avoid interference
%   with other functions using temporary variables.
% \end{variable}
%
% \begin{function}{\::n, \::N, \::p, \::c, \::o, \::f, \::x, \::v, \::V, \:::}
%   \begin{syntax}
%     |\cs_set:Npn \exp_args:Ncof { \::c \::o \::f \::: }|
%   \end{syntax}
%   Internal forms for the base expansion types. These names do \emph{not}
%   conform to the general \LaTeX3 approach as this makes them more readily
%   visible in the log and so forth.
% \end{function}
%
% \end{documentation}
%
% \begin{implementation}
%
% \section{\pkg{l3expan} implementation}
%
%    \begin{macrocode}
%<*initex|package>
%    \end{macrocode}
%
%    \begin{macrocode}
%<@@=exp>
%    \end{macrocode}
%
% \begin{macro}{\exp_after:wN}
% \begin{macro}{\exp_not:N}
% \begin{macro}{\exp_not:n}
%   These are defined in \pkg{l3basics}.
% \end{macro}
% \end{macro}
% \end{macro}
%
% \subsection{General expansion}
%
% In this section a general mechanism for defining functions to handle
% argument handling is defined.  These general expansion functions are
% expandable unless |x| is used.  (Any version of |x| is going to have
% to use one of the \LaTeX3 names for \cs{cs_set:Npx} at some
% point, and so is never going to be expandable.)
%
% The definition of expansion functions with this technique happens
% in section~\ref{sec:l3expan:gendef}.
% In section~\ref{sec:l3expan:handtune} some common cases are coded by a more direct
% method for efficiency, typically using calls to \cs{exp_after:wN}.
%
% \begin{variable}{\l_@@_internal_tl}
%   This scratch token list variable is defined in \pkg{l3basics}, as it
%   is needed \enquote{early}.  This is just a reminder that is the
%   case!
% \end{variable}
%
% This code uses internal functions with names that start with |\::|
% to perform the expansions. All macros are |long| as this turned out
% to be desirable since the tokens undergoing expansion may be
% arbitrary user input.
%
% An argument manipulator |\::|\meta{Z} always has signature |#1\:::#2#3|
% where |#1| holds the remaining argument manipulations to be performed,
% \cs{:::} serves as an end marker for the list of manipulations, |#2|
% is the carried over result of the previous expansion steps and |#3| is
% the argument about to be processed.
% One exception to this rule is \cs{::p}, which has to grab an argument
% delimited by a left brace.
%
% \begin{macro}[aux, EXP]{\@@_arg_next:nnn}
% \begin{macro}[aux, EXP]{\@@_arg_next:Nnn}
%   |#1| is the result of an expansion step, |#2| is the remaining
%   argument manipulations and |#3| is the current result of the
%   expansion chain.  This auxiliary function moves |#1| back after
%   |#3| in the input stream and checks if any expansion is left to
%   be done by calling |#2|. In by far the most cases we will require
%   to add a set of braces to the result of an argument manipulation
%   so it is more effective to do it directly here. Actually, so far
%   only the |c| of the final argument manipulation variants does not
%   require a set of braces.
%    \begin{macrocode}
\cs_new:Npn \@@_arg_next:nnn #1#2#3 { #2 \::: { #3 {#1} } }
\cs_new:Npn \@@_arg_next:Nnn #1#2#3 { #2 \::: { #3 #1 } }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}[int]{\:::}
%   The end marker is just another name for the identity function.
%    \begin{macrocode}
\cs_new:Npn \::: #1 {#1}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}[int, EXP]{\::n}
%   This function is used to skip an argument that doesn't need to
%   be expanded.
%    \begin{macrocode}
\cs_new:Npn \::n #1 \::: #2#3 { #1 \::: { #2 {#3} } }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}[int, EXP]{\::N}
%   This function is used to skip an argument that consists of a
%   single token and doesn't need to be expanded.
%    \begin{macrocode}
\cs_new:Npn \::N #1 \::: #2#3 { #1 \::: {#2#3} }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}[int, EXP]{\::p}
%   This function is used to skip an argument that is delimited by a
%   left brace and doesn't need to be expanded.  It should not be
%   wrapped in braces in the result.
%    \begin{macrocode}
\cs_new:Npn \::p #1 \::: #2#3# { #1 \::: {#2#3} }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}[int, EXP]{\::c}
%   This function is used to skip an argument that is turned into
%   a control sequence without expansion.
%    \begin{macrocode}
\cs_new:Npn \::c #1 \::: #2#3
  { \exp_after:wN \@@_arg_next:Nnn \cs:w #3 \cs_end: {#1} {#2} }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}[int, EXP]{\::o}
%   This function is used to expand an argument once.
%    \begin{macrocode}
\cs_new:Npn \::o #1 \::: #2#3
  { \exp_after:wN \@@_arg_next:nnn \exp_after:wN {#3} {#1} {#2} }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}[int, EXP]{\::f}
% \begin{macro}{\exp_stop_f:}
%   This function is used to expand a token list until the first
%   unexpandable token is found. This is achieved through \cs{exp:w}
%   \cs{exp_end_continue_f:w} that
%   expands everything in its way following it. This
%   scanning procedure is terminated once the expansion hits
%   something non-expandable or a space. We introduce \cs{exp_stop_f:}
%   to mark such an end of expansion marker.
%   In the example shown earlier the scanning was stopped
%   once \TeX{} had fully expanded |\cs_set_eq:Nc \aaa { b \l_tmpa_tl b }|
%   into |\cs_set_eq:NN \aaa = \blurb| which then turned out to contain
%   the non-expandable token \cs{cs_set_eq:NN}.  Since the expansion of
%   \cs{exp:w} \cs{exp_end_continue_f:w} is \meta{null}, we wind up with a fully
%   expanded list, only \TeX{} has not tried to execute any of the
%   non-expandable tokens. This is what differentiates this function
%   from the |x| argument type.
%    \begin{macrocode}
\cs_new:Npn \::f #1 \::: #2#3
  {
    \exp_after:wN \@@_arg_next:nnn
      \exp_after:wN { \exp:w \exp_end_continue_f:w #3 }
      {#1} {#2}
  }
\use:nn { \cs_new_eq:NN \exp_stop_f: } { ~ }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}[int]{\::x}
%   This function is used to expand an argument fully.
%    \begin{macrocode}
\cs_new_protected:Npn \::x #1 \::: #2#3
  {
    \cs_set_nopar:Npx \l_@@_internal_tl { {#3} }
    \exp_after:wN \@@_arg_next:nnn \l_@@_internal_tl {#1} {#2}
  }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}[int, EXP]{\::v}
% \begin{macro}[int, EXP]{\::V}
%   These functions return the value of a register, i.e., one of
%   |tl|, |clist|, |int|, |skip|, |dim| and |muskip|. The |V| version
%   expects a single token whereas |v| like |c| creates a csname from
%   its argument given in braces and then evaluates it as if it was a
%   |V|. The \cs{exp:w} sets off an expansion
%   similar to an |f| type expansion, which we will terminate using
%   \cs{exp_end:}. The argument is returned in braces.
%    \begin{macrocode}
\cs_new:Npn \::V #1 \::: #2#3
  {
    \exp_after:wN \@@_arg_next:nnn
      \exp_after:wN { \exp:w \@@_eval_register:N #3 }
      {#1} {#2}
}
\cs_new:Npn \::v # 1\::: #2#3
  {
    \exp_after:wN \@@_arg_next:nnn
      \exp_after:wN { \exp:w \@@_eval_register:c {#3} }
      {#1} {#2}
  }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}[int, EXP]{\@@_eval_register:N, \@@_eval_register:c}
% \begin{macro}[aux, EXP]{\@@_eval_error_msg:w}
%   This function evaluates a register. Now a register might exist as
%   one of two things: A parameter-less macro or a built-in \TeX{}
%   register such as |\count|. For the \TeX{} registers we have to
%   utilize a \tn{the} whereas for the macros we merely have to
%   expand them once. The trick is to find out when to use
%   \tn{the} and when not to. What we do here is try to find out
%   whether the token will expand to something else when hit with
%   \cs{exp_after:wN}. The technique is to compare the meaning of the
%   register in question when it has been prefixed with \cs{exp_not:N}
%   and the register itself. If it is a macro, the prefixed
%   \cs{exp_not:N} will temporarily turn it into the primitive
%   \cs{scan_stop:}.
%    \begin{macrocode}
\cs_new:Npn \@@_eval_register:N #1
  {
    \exp_after:wN \if_meaning:w \exp_not:N #1 #1
%    \end{macrocode}
%   If the token was not a macro it may be a malformed variable from a
%   |c| expansion in which case it is equal to the primitive
%   \cs{scan_stop:}. In that case we throw an error. We could let \TeX{}
%   do it for us but that would result in the rather obscure
%   \begin{quote}
%     |! You can't use `\relax' after \the.|
%   \end{quote}
%   which while quite true doesn't give many hints as to what actually
%   went wrong. We provide something more sensible.
%    \begin{macrocode}
      \if_meaning:w \scan_stop: #1
        \@@_eval_error_msg:w
      \fi:
%    \end{macrocode}
%   The next bit requires some explanation. The function must be
%   initiated by \cs{exp:w} and we want to
%   terminate this expansion chain by inserting the \cs{exp_end:}
%    token.
%   However, we have to expand the register |#1| before we do
%   that. If it is a \TeX{} register, we need to execute the sequence
%   |\exp_after:wN \exp_end: \tex_the:D #1| and if it is a macro we
%   need to execute |\exp_after:wN \exp_end: #1|. We therefore issue
%   the longer of the two sequences and if the register is a macro, we
%   remove the \cs{tex_the:D}.
%    \begin{macrocode}
    \else:
      \exp_after:wN \use_i_ii:nnn
    \fi:
    \exp_after:wN \exp_end: \tex_the:D #1
  }
\cs_new:Npn \@@_eval_register:c #1
  { \exp_after:wN \@@_eval_register:N \cs:w #1 \cs_end: }
%    \end{macrocode}
%   Clean up nicely, then call the undefined control sequence. The
%   result is an error message looking like this:
%   \begin{verbatim}
%     ! Undefined control sequence.
%     <argument> \LaTeX3 error:
%                               Erroneous variable used!
%     l.55 \tl_set:Nv \l_tmpa_tl {undefined_tl}
%   \end{verbatim}
%    \begin{macrocode}
\cs_new:Npn \@@_eval_error_msg:w #1 \tex_the:D #2
  {
      \fi:
    \fi:
    \__msg_kernel_expandable_error:nnn { kernel } { bad-variable } {#2}
    \exp_end:
  }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \subsection{Hand-tuned definitions}
% \label{sec:l3expan:handtune}
%
% One of the most important features of these functions is that they
% are fully expandable and therefore allow to prefix them with
% \cs{tex_global:D} for example.
%
% \begin{macro}[EXP]{\exp_args:No}
% \begin{macro}[EXP]{\exp_args:NNo}
% \begin{macro}[EXP]{\exp_args:NNNo}
%   Those lovely runs of expansion!
%    \begin{macrocode}
\cs_new:Npn \exp_args:No #1#2 { \exp_after:wN #1 \exp_after:wN {#2} }
\cs_new:Npn \exp_args:NNo #1#2#3
  { \exp_after:wN #1 \exp_after:wN #2 \exp_after:wN {#3} }
\cs_new:Npn \exp_args:NNNo #1#2#3#4
  { \exp_after:wN #1 \exp_after:wN#2 \exp_after:wN #3 \exp_after:wN {#4} }
%    \end{macrocode}
% \end{macro}
% \end{macro}
% \end{macro}
%
% \begin{macro}[EXP]{\exp_args:Nc, \exp_args:cc}
%   In \pkg{l3basics}.
% \end{macro}
%
% \begin{macro}[EXP]{\exp_args:NNc, \exp_args:Ncc, \exp_args:Nccc}
%   Here are the functions that turn their argument into csnames but are
%   expandable.
%    \begin{macrocode}
\cs_new:Npn \exp_args:NNc #1#2#3
  { \exp_after:wN #1 \exp_after:wN #2 \cs:w # 3\cs_end: }
\cs_new:Npn \exp_args:Ncc #1#2#3
  { \exp_after:wN #1 \cs:w #2 \exp_after:wN \cs_end: \cs:w #3 \cs_end: }
\cs_new:Npn \exp_args:Nccc #1#2#3#4
  {
    \exp_after:wN #1
      \cs:w #2 \exp_after:wN \cs_end:
      \cs:w #3 \exp_after:wN \cs_end:
      \cs:w #4 \cs_end:
  }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}[EXP]{\exp_args:Nf, \exp_args:NV, \exp_args:Nv}
%    \begin{macrocode}
\cs_new:Npn \exp_args:Nf #1#2
  { \exp_after:wN #1 \exp_after:wN { \exp:w \exp_end_continue_f:w #2 } }
\cs_new:Npn \exp_args:Nv #1#2
  {
    \exp_after:wN #1 \exp_after:wN
      { \exp:w \@@_eval_register:c {#2} }
  }
\cs_new:Npn \exp_args:NV #1#2
  {
    \exp_after:wN #1 \exp_after:wN
      { \exp:w \@@_eval_register:N #2 }
  }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}[EXP]
%   {
%     \exp_args:NNV, \exp_args:NNv, \exp_args:NNf,
%     \exp_args:NVV,
%     \exp_args:Ncf, \exp_args:Nco
%   }
%   Some more hand-tuned function with three arguments.
%   If we forced that an |o| argument always has braces,
%   we could implement \cs{exp_args:Nco} with less tokens
%   and only two arguments.
%    \begin{macrocode}
\cs_new:Npn \exp_args:NNf #1#2#3
  {
    \exp_after:wN #1
    \exp_after:wN #2
    \exp_after:wN { \exp:w \exp_end_continue_f:w #3 }
  }
\cs_new:Npn \exp_args:NNv #1#2#3
  {
    \exp_after:wN #1
    \exp_after:wN #2
    \exp_after:wN { \exp:w \@@_eval_register:c {#3} }
  }
\cs_new:Npn \exp_args:NNV #1#2#3
  {
    \exp_after:wN #1
    \exp_after:wN #2
    \exp_after:wN { \exp:w \@@_eval_register:N #3 }
  }
\cs_new:Npn \exp_args:Nco #1#2#3
  {
    \exp_after:wN #1
    \cs:w #2 \exp_after:wN \cs_end:
    \exp_after:wN {#3}
  }
\cs_new:Npn \exp_args:Ncf #1#2#3
  {
    \exp_after:wN #1
    \cs:w #2 \exp_after:wN \cs_end:
    \exp_after:wN { \exp:w \exp_end_continue_f:w #3 }
  }
\cs_new:Npn \exp_args:NVV #1#2#3
  {
    \exp_after:wN #1
    \exp_after:wN { \exp:w \exp_after:wN
      \@@_eval_register:N \exp_after:wN #2 \exp_after:wN }
    \exp_after:wN { \exp:w \@@_eval_register:N #3 }
  }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}[EXP]
%   {
%     \exp_args:Ncco, \exp_args:NcNc, \exp_args:NcNo,
%     \exp_args:NNNV
%   }
%   A few more that we can hand-tune.
%    \begin{macrocode}
\cs_new:Npn \exp_args:NNNV #1#2#3#4
  {
    \exp_after:wN #1
    \exp_after:wN #2
    \exp_after:wN #3
    \exp_after:wN { \exp:w \@@_eval_register:N #4 }
  }
\cs_new:Npn \exp_args:NcNc #1#2#3#4
  {
    \exp_after:wN #1
    \cs:w #2 \exp_after:wN \cs_end:
    \exp_after:wN #3
    \cs:w #4 \cs_end:
  }
\cs_new:Npn \exp_args:NcNo #1#2#3#4
  {
    \exp_after:wN #1
    \cs:w #2 \exp_after:wN \cs_end:
    \exp_after:wN #3
    \exp_after:wN {#4}
  }
\cs_new:Npn \exp_args:Ncco #1#2#3#4
  {
    \exp_after:wN #1
    \cs:w #2 \exp_after:wN \cs_end:
    \cs:w #3 \exp_after:wN \cs_end:
    \exp_after:wN {#4}
  }
%    \end{macrocode}
% \end{macro}
%
% \subsection{Definitions with the automated technique}
% \label{sec:l3expan:gendef}
%
% Some of these could be done more efficiently, but the complexity of
% coding then becomes an issue. Notice that the auto-generated functions
% are all not long: they don't actually take any arguments themselves.
%
% \begin{macro}{\exp_args:Nx}
%    \begin{macrocode}
\cs_new_protected:Npn \exp_args:Nx { \::x \::: }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}[EXP]
%   {
%     \exp_args:Nnc, \exp_args:Nfo, \exp_args:Nff, \exp_args:Nnf,
%     \exp_args:Nno, \exp_args:NnV, \exp_args:Noo, \exp_args:Nof,
%     \exp_args:Noc
%   }
% \begin{macro}
%   {
%     \exp_args:NNx, \exp_args:Ncx, \exp_args:Nnx,
%     \exp_args:Nox, \exp_args:Nxo, \exp_args:Nxx,
%   }
%   Here are the actual function definitions, using the helper functions
%   above.
%    \begin{macrocode}
\cs_new:Npn \exp_args:Nnc { \::n \::c \::: }
\cs_new:Npn \exp_args:Nfo { \::f \::o \::: }
\cs_new:Npn \exp_args:Nff { \::f \::f \::: }
\cs_new:Npn \exp_args:Nnf { \::n \::f \::: }
\cs_new:Npn \exp_args:Nno { \::n \::o \::: }
\cs_new:Npn \exp_args:NnV { \::n \::V \::: }
\cs_new:Npn \exp_args:Noo { \::o \::o \::: }
\cs_new:Npn \exp_args:Nof { \::o \::f \::: }
\cs_new:Npn \exp_args:Noc { \::o \::c \::: }
\cs_new_protected:Npn \exp_args:NNx { \::N \::x \::: }
\cs_new_protected:Npn \exp_args:Ncx { \::c \::x \::: }
\cs_new_protected:Npn \exp_args:Nnx { \::n \::x \::: }
\cs_new_protected:Npn \exp_args:Nox { \::o \::x \::: }
\cs_new_protected:Npn \exp_args:Nxo { \::x \::o \::: }
\cs_new_protected:Npn \exp_args:Nxx { \::x \::x \::: }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}[EXP]
%   {
%     \exp_args:NNno, \exp_args:NNoo,
%     \exp_args:Nnnc, \exp_args:Nnno, \exp_args:Nooo,
%   }
% \begin{macro}
%   {
%     \exp_args:NNNx,
%     \exp_args:NNnx, \exp_args:NNox, \exp_args:Nnnx, \exp_args:Nnox,
%     \exp_args:Nccx, \exp_args:Ncnx, \exp_args:Noox,
%   }
%    \begin{macrocode}
\cs_new:Npn \exp_args:NNno { \::N \::n \::o \::: }
\cs_new:Npn \exp_args:NNoo { \::N \::o \::o \::: }
\cs_new:Npn \exp_args:Nnnc { \::n \::n \::c \::: }
\cs_new:Npn \exp_args:Nnno { \::n \::n \::o \::: }
\cs_new:Npn \exp_args:Nooo { \::o \::o \::o \::: }
\cs_new_protected:Npn \exp_args:NNNx { \::N \::N \::x \::: }
\cs_new_protected:Npn \exp_args:NNnx { \::N \::n \::x \::: }
\cs_new_protected:Npn \exp_args:NNox { \::N \::o \::x \::: }
\cs_new_protected:Npn \exp_args:Nnnx { \::n \::n \::x \::: }
\cs_new_protected:Npn \exp_args:Nnox { \::n \::o \::x \::: }
\cs_new_protected:Npn \exp_args:Nccx { \::c \::c \::x \::: }
\cs_new_protected:Npn \exp_args:Ncnx { \::c \::n \::x \::: }
\cs_new_protected:Npn \exp_args:Noox { \::o \::o \::x \::: }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \subsection{Last-unbraced versions}
%
% \begin{macro}[aux, EXP]{\@@_arg_last_unbraced:nn}
% \begin{macro}[aux, EXP]{\::f_unbraced}
% \begin{macro}[aux, EXP]{\::o_unbraced}
% \begin{macro}[aux, EXP]{\::V_unbraced}
% \begin{macro}[aux, EXP]{\::v_unbraced}
% \begin{macro}[aux, EXP]{\::x_unbraced}
%   There are a few places where the last argument needs to be available
%   unbraced. First some helper macros.
%    \begin{macrocode}
\cs_new:Npn \@@_arg_last_unbraced:nn #1#2 { #2#1 }
\cs_new:Npn \::f_unbraced \::: #1#2
  {
    \exp_after:wN \@@_arg_last_unbraced:nn
      \exp_after:wN { \exp:w \exp_end_continue_f:w #2 } {#1}
  }
\cs_new:Npn \::o_unbraced \::: #1#2
  { \exp_after:wN \@@_arg_last_unbraced:nn \exp_after:wN {#2} {#1} }
\cs_new:Npn \::V_unbraced \::: #1#2
  {
    \exp_after:wN \@@_arg_last_unbraced:nn
      \exp_after:wN { \exp:w \@@_eval_register:N #2 } {#1}
  }
\cs_new:Npn \::v_unbraced \::: #1#2
  {
    \exp_after:wN \@@_arg_last_unbraced:nn
      \exp_after:wN { \exp:w \@@_eval_register:c {#2} } {#1}
  }
\cs_new_protected:Npn \::x_unbraced \::: #1#2
  {
    \cs_set_nopar:Npx \l_@@_internal_tl { \exp_not:n {#1} #2 }
    \l_@@_internal_tl
  }
%    \end{macrocode}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
%
% \begin{macro}[EXP]{\exp_last_unbraced:NV}
% \begin{macro}[EXP]{\exp_last_unbraced:Nv}
% \begin{macro}[EXP]{\exp_last_unbraced:Nf}
% \begin{macro}[EXP]{\exp_last_unbraced:No}
% \begin{macro}[EXP]{\exp_last_unbraced:Nco}
% \begin{macro}[EXP]{\exp_last_unbraced:NcV}
% \begin{macro}[EXP]{\exp_last_unbraced:NNV}
% \begin{macro}[EXP]{\exp_last_unbraced:NNo}
% \begin{macro}[EXP]{\exp_last_unbraced:NNNV}
% \begin{macro}[EXP]{\exp_last_unbraced:NNNo}
% \begin{macro}[EXP]{\exp_last_unbraced:Nno}
% \begin{macro}[EXP]{\exp_last_unbraced:Noo}
% \begin{macro}[EXP]{\exp_last_unbraced:Nfo}
% \begin{macro}[EXP]{\exp_last_unbraced:NnNo}
% \begin{macro}{\exp_last_unbraced:Nx}
%   Now the business end: most of these are hand-tuned for speed, but the
%   general system is in place.
%    \begin{macrocode}
\cs_new:Npn \exp_last_unbraced:NV #1#2
  { \exp_after:wN #1 \exp:w \@@_eval_register:N #2 }
\cs_new:Npn \exp_last_unbraced:Nv #1#2
  { \exp_after:wN #1 \exp:w \@@_eval_register:c {#2} }
\cs_new:Npn \exp_last_unbraced:No #1#2 { \exp_after:wN #1 #2 }
\cs_new:Npn \exp_last_unbraced:Nf #1#2
  { \exp_after:wN #1 \exp:w \exp_end_continue_f:w #2 }
\cs_new:Npn \exp_last_unbraced:Nco #1#2#3
  { \exp_after:wN #1 \cs:w #2 \exp_after:wN \cs_end: #3 }
\cs_new:Npn \exp_last_unbraced:NcV #1#2#3
  {
    \exp_after:wN #1
    \cs:w #2 \exp_after:wN \cs_end:
    \exp:w \@@_eval_register:N #3
  }
\cs_new:Npn \exp_last_unbraced:NNV #1#2#3
  {
    \exp_after:wN #1
    \exp_after:wN #2
    \exp:w \@@_eval_register:N #3
  }
\cs_new:Npn \exp_last_unbraced:NNo #1#2#3
  { \exp_after:wN #1 \exp_after:wN #2 #3 }
\cs_new:Npn \exp_last_unbraced:NNNV #1#2#3#4
  {
    \exp_after:wN #1
    \exp_after:wN #2
    \exp_after:wN #3
    \exp:w \@@_eval_register:N #4
  }
\cs_new:Npn \exp_last_unbraced:NNNo #1#2#3#4
  { \exp_after:wN #1 \exp_after:wN #2 \exp_after:wN #3 #4 }
\cs_new:Npn \exp_last_unbraced:Nno { \::n \::o_unbraced \::: }
\cs_new:Npn \exp_last_unbraced:Noo { \::o \::o_unbraced \::: }
\cs_new:Npn \exp_last_unbraced:Nfo { \::f \::o_unbraced \::: }
\cs_new:Npn \exp_last_unbraced:NnNo { \::n \::N \::o_unbraced \::: }
\cs_new_protected:Npn \exp_last_unbraced:Nx { \::x_unbraced \::: }
%    \end{macrocode}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
%
% \begin{macro}[EXP]{\exp_last_two_unbraced:Noo}
% \begin{macro}[EXP, aux]{\@@_last_two_unbraced:noN}
%   If |#2| is a single token then this can be implemented as
%   \begin{verbatim}
%     \cs_new:Npn \exp_last_two_unbraced:Noo #1 #2 #3
%      { \exp_after:wN \exp_after:wN \exp_after:wN #1 \exp_after:wN #2 #3 }
%   \end{verbatim}
%   However, for robustness this is not suitable. Instead, a bit of a
%   shuffle is used to ensure that |#2| can be multiple tokens.
%    \begin{macrocode}
\cs_new:Npn \exp_last_two_unbraced:Noo #1#2#3
  { \exp_after:wN \@@_last_two_unbraced:noN \exp_after:wN {#3} {#2} #1 }
\cs_new:Npn \@@_last_two_unbraced:noN #1#2#3
   { \exp_after:wN #3 #2 #1 }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \subsection{Preventing expansion}
%
% \begin{macro}[EXP]{\exp_not:o}
% \begin{macro}[EXP]{\exp_not:c}
% \begin{macro}[EXP]{\exp_not:f}
% \begin{macro}[EXP]{\exp_not:V}
% \begin{macro}[EXP]{\exp_not:v}
%    \begin{macrocode}
\cs_new:Npn \exp_not:o #1 { \etex_unexpanded:D \exp_after:wN {#1} }
\cs_new:Npn \exp_not:c #1 { \exp_after:wN \exp_not:N \cs:w #1 \cs_end: }
\cs_new:Npn \exp_not:f #1
  { \etex_unexpanded:D \exp_after:wN { \exp:w \exp_end_continue_f:w #1 } }
\cs_new:Npn \exp_not:V #1
  {
    \etex_unexpanded:D \exp_after:wN
      { \exp:w \@@_eval_register:N #1 }
  }
\cs_new:Npn \exp_not:v #1
  {
    \etex_unexpanded:D \exp_after:wN
      { \exp:w \@@_eval_register:c {#1} }
  }
%    \end{macrocode}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
%
% \subsection{Controlled expansion}
%
% \begin{macro}{\exp:w}
% \begin{macro}{\exp_end:}
% \begin{macro}{\exp_end_continue_f:w}
% \begin{macro}{\exp_end_continue_f:nw}
%
%    To trigger a sequence of \enquote{arbitrary} many expansions we need a
%    method to invoke \TeX's expansion mechanism in such a way that a)
%    we are able to stop it in a controlled manner and b) that the
%    result of what triggered the expansion in the first place is
%    null, i.e., that we do not get any unwanted side effects. There
%    aren't that many possibilities in \TeX{}; in fact the one
%    explained below might well be the only one (as normally the
%    result of expansion is not null).
%
%    The trick here is to make use of the fact that
%    \cs{tex_romannumeral:D} expands the tokens following it when
%    looking for a number and that its expansion is null if that
%    number turns out to be zero or negative. So we use that to start
%    the expansion sequence.
%    \begin{macrocode}
%\cs_new_eq:NN \exp:w    \tex_romannumeral:D
%    \end{macrocode}
%    So to stop the expansion sequence in a controlled way all we need
%    to provide is a constant integer zero as part of expanded tokens. As this is
%    an integer constant it will immediately stop
%    \cs{tex_romannumerl:D}'s search for a number.
%    \begin{macrocode}
%\int_const:Nn \exp_end: { 0 }
%    \end{macrocode}
%    (Note that according to our specification all tokens we expand
%    initiated by \cs{exp:w} are supposed to be expandable (as well as
%    their replacement text in the expansion) so we will not encounter
%    a \enquote{number} that actually result in a roman numeral being
%    generated. Or if we do then the programmer made a mistake.)
%
%    If on the other hand we want to stop the initial expansion
%    sequence but continue with an f-type expansion we provide the
%    alphabetic constant |`^^@| that also represents |0| but this time
%    \TeX's syntax for a \meta{number} will continue searching for an
%    optional space (and it will continue expansion doing that) ---
%    see \TeX{}book page~269 for details.
%    \begin{macrocode}
\tex_catcode:D `\^^@=13
\cs_new_protected:Npn \exp_end_continue_f:w {`^^@}
%    \end{macrocode}
%    If the above definition ever appears outside its proper context
%    the active character |^^@| will be executed so we turn this into an
%    error.\footnote{Need to get a real error message.}
%    \begin{macrocode}
\cs_new:Npn ^^@{\expansionERROR}
%    \end{macrocode}
%
%    \begin{macrocode}
\cs_new:Npn \exp_end_continue_f:nw #1 { `^^@ #1 }
\tex_catcode:D `\^^@=15
%    \end{macrocode}
% \end{macro}
% \end{macro}
% \end{macro}
% \end{macro}
%
%
%
% \subsection{Defining function variants}
%
%    \begin{macrocode}
%<@@=cs>
%    \end{macrocode}
%
% \begin{macro}{\cs_generate_variant:Nn}
%   \begin{arguments}
%     \item Base form of a function; \emph{e.g.},~\cs{tl_set:Nn}
%     \item One or more variant argument specifiers; e.g., |{Nx,c,cx}|
%   \end{arguments}
%   After making sure that the base form exists, test whether it is
%   protected or not and define \cs{@@_tmp:w} as either
%   \cs{cs_new:Npx} or \cs{cs_new_protected:Npx}, which is
%   then used to define all the variants (except those involving
%   \texttt{x}-expansion, always protected).  Split up the original base
%   function only once, to grab its name and signature.  Then we wish to
%   iterate through the comma list of variant argument specifiers, which
%   we first convert to a string: the reason is explained later.
%    \begin{macrocode}
\cs_new_protected:Npn \cs_generate_variant:Nn #1#2
  {
    \__chk_if_exist_cs:N #1
    \@@_generate_variant:N #1
    \exp_after:wN \@@_split_function:NN
    \exp_after:wN #1
    \exp_after:wN \@@_generate_variant:nnNN
    \exp_after:wN #1
    \tl_to_str:n {#2} , \scan_stop: , \q_recursion_stop
  }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}[aux]{\@@_generate_variant:N}
% \begin{macro}[aux]{\@@_generate_variant:ww, \@@_generate_variant:wwNw}
%   The goal here is to pick up protected parent functions.  There are
%   four cases: the parent function can be a primitive or a macro, and
%   can be expandable or not.  For non-expandable primitives, all
%   variants should be protected; skipping the \cs{else:} branch is safe
%   because all primitive \TeX{} conditionals are expandable.
%
%   The other case where variants should be protected is when the parent
%   function is a protected macro: then |protected| appears in the
%   meaning before the fist occurrence of |macro|.  The |ww| auxiliary
%   removes everything in the meaning string after the first |ma|.  We
%   use |ma| rather than the full |macro| because the meaning of the
%   \tn{firstmark} primitive (and four others) can contain an arbitrary
%   string after a leading |firstmark:|.  Then, look for |pr| in the
%   part we extracted: no need to look for anything longer: the only
%   strings we can have are an empty string, \verb*|\long |,
%   \verb*|\protected |, \verb*|\protected\long |, |\first|, |\top|,
%   |\bot|, |\splittop|, or |\splitbot|, with |\| replaced by the
%   appropriate escape character.  If |pr| appears in the part before
%   |ma|, the first \cs{q_mark} is taken as an argument of the |wwNw|
%   auxiliary, and |#3| is \cs{cs_new_protected:Npx}, otherwise it
%   is \cs{cs_new:Npx}.
%    \begin{macrocode}
\cs_new_protected:Npx \@@_generate_variant:N #1
  {
    \exp_not:N \exp_after:wN \exp_not:N \if_meaning:w
      \exp_not:N \exp_not:N #1 #1
      \cs_set_eq:NN \exp_not:N \@@_tmp:w \cs_new_protected:Npx
    \exp_not:N \else:
      \exp_not:N \exp_after:wN \exp_not:N \@@_generate_variant:ww
        \exp_not:N \token_to_meaning:N #1 \tl_to_str:n { ma }
          \exp_not:N \q_mark
        \exp_not:N \q_mark \cs_new_protected:Npx
        \tl_to_str:n { pr }
        \exp_not:N \q_mark \cs_new:Npx
        \exp_not:N \q_stop
    \exp_not:N \fi:
  }
\use:x
  {
    \cs_new_protected:Npn \exp_not:N \@@_generate_variant:ww
      ##1 \tl_to_str:n { ma } ##2 \exp_not:N \q_mark
  }
  { \@@_generate_variant:wwNw #1 }
\use:x
  {
    \cs_new_protected:Npn \exp_not:N \@@_generate_variant:wwNw
      ##1 \tl_to_str:n { pr } ##2 \exp_not:N \q_mark
      ##3 ##4 \exp_not:N \q_stop
  }
  { \cs_set_eq:NN \@@_tmp:w #3 }
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}[aux]{\@@_generate_variant:nnNN}
%   \begin{arguments}
%     \item Base name.
%     \item Base signature.
%     \item Boolean.
%     \item Base function.
%   \end{arguments}
%   If the boolean is \cs{c_false_bool}, the base function has no colon
%   and we abort with an error; otherwise, set off a loop through the
%   desired variant forms. The original function is retained as |#4| for
%   efficiency.
%    \begin{macrocode}
\cs_new_protected:Npn \@@_generate_variant:nnNN #1#2#3#4
  {
    \if_meaning:w \c_false_bool #3
      \__msg_kernel_error:nnx { kernel } { missing-colon }
        { \token_to_str:c {#1} }
      \exp_after:wN \use_none_delimit_by_q_recursion_stop:w
    \fi:
    \@@_generate_variant:Nnnw #4 {#1}{#2}
  }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}[aux]{\@@_generate_variant:Nnnw}
%   \begin{arguments}
%     \item Base function.
%     \item Base name.
%     \item Base signature.
%     \item Beginning of variant signature.
%   \end{arguments}
%   First check whether to terminate the loop over variant forms.  Then,
%   for each variant form, construct a new function name using the
%   original base name, the variant signature consisting of $l$ letters
%   and the last $k-l$ letters of the base signature (of length $k$).
%   For example, for a base function \cs{prop_put:Nnn} which needs a
%   |cV| variant form, we want the new signature to be |cVn|.
%
%   There are further subtleties:
%   \begin{itemize}
%     \item In \cs{cs_generate_variant:Nn} |\foo:nnTF| |{xxTF}|, it
%       would be better to define |\foo:xxTF| using |\exp_args:Nxx|,
%       rather than a hypothetical |\exp_args:NxxTF|.  Thus, we wish to
%       trim a common trailing part from the base signature and the
%       variant signature.
%     \item In \cs{cs_generate_variant:Nn} |\foo:on| |{ox}|, the
%       function |\foo:ox| should be defined using |\exp_args:Nnx|, not
%       |\exp_args:Nox|, to avoid double |o| expansion.
%     \item Lastly, \cs{cs_generate_variant:Nn} |\foo:on| |{xn}| should
%       trigger an error, because we do not have a means to replace
%       |o|-expansion by |x|-expansion.
%     \end{itemize}
%     All this boils down to a few rules.  Only |n| and |N|-type
%     arguments can be replaced by \cs{cs_generate_variant:Nn}.  Other
%     argument types are allowed to be passed unchanged from the base
%     form to the variant: in the process they are changed to |n|
%     (except for two cases: |N| and |p|-type arguments).  A common
%     trailing part is ignored.
%
%     We compare the base and variant signatures one character at a time
%     within |x|-expansion.  The result is given to
%     \cs{@@_generate_variant:wwNN} in the form \meta{processed variant
%       signature} \cs{q_mark} \meta{errors} \cs{q_stop} \meta{base
%       function} \meta{new function}.  If all went well, \meta{errors}
%     is empty; otherwise, it is a kernel error message, followed by
%     some clean-up code (\cs{use_none:nnnn}).
%
%     Note the space after |#3| and after the following brace group.
%     Those are ignored by \TeX{} when fetching the last argument for
%     \cs{@@_generate_variant_loop:nNwN}, but can be used as a delimiter
%     for \cs{@@_generate_variant_loop_end:nwwwNNnn}.
%    \begin{macrocode}
\cs_new_protected:Npn \@@_generate_variant:Nnnw #1#2#3#4 ,
  {
    \if_meaning:w \scan_stop: #4
      \exp_after:wN \use_none_delimit_by_q_recursion_stop:w
    \fi:
    \use:x
      {
        \exp_not:N \@@_generate_variant:wwNN
        \@@_generate_variant_loop:nNwN { }
          #4
          \@@_generate_variant_loop_end:nwwwNNnn
          \q_mark
          #3 ~
          { ~ { } \fi: \@@_generate_variant_loop_long:wNNnn } ~
          { }
          \q_stop
        \exp_not:N #1 {#2} {#4}
      }
    \@@_generate_variant:Nnnw #1 {#2} {#3}
  }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}[aux, EXP]
%   {
%     \@@_generate_variant_loop:nNwN,
%     \@@_generate_variant_loop_same:w,
%     \@@_generate_variant_loop_end:nwwwNNnn,
%     \@@_generate_variant_loop_long:wNNnn,
%     \@@_generate_variant_loop_invalid:NNwNNnn,
%   }
%   \begin{arguments}
%     \item Last few (consecutive) letters common between the base
%       and variant (in fact, \cs{@@_generate_variant_same:N}
%       \meta{letter} for each letter).
%     \item Next variant letter.
%     \item Remainder of variant form.
%     \item Next base letter.
%   \end{arguments}
%   The first argument is populated by
%   \cs{@@_generate_variant_loop_same:w} when a variant letter and a
%   base letter match.  It is flushed into the input stream whenever the
%   two letters are different: if the loop ends before, the argument is
%   dropped, which means that trailing common letters are ignored.
%
%   The case where the two letters are different is only allowed with a
%   base letter of |N| or |n|.  Otherwise, call
%   \cs{@@_generate_variant_loop_invalid:NNwNNnn} to remove the end of
%   the loop, get arguments at the end of the loop, and place an
%   appropriate error message as a second argument of
%   \cs{@@_generate_variant:wwNN}.  If the letters are distinct and
%   the base letter is indeed |n| or |N|, leave in the input stream
%   whatever argument was collected, and the next variant letter |#2|,
%   then loop by calling \cs{@@_generate_variant_loop:nNwN}.
%
%   The loop can stop in three ways.
%   \begin{itemize}
%     \item If the end of the variant form is encountered first, |#2| is
%       \cs{@@_generate_variant_loop_end:nwwwNNnn} (expanded by the
%       conditional \cs{if:w}), which inserts some tokens to end the
%       conditional; grabs the \meta{base name} as |#7|, the
%       \meta{variant signature} |#8|, the \meta{next base letter} |#1|
%       and the part |#3| of the base signature that wasn't read yet;
%       and combines those into the \meta{new function} to be defined.
%     \item If the end of the base form is encountered first, |#4| is
%       |~{}\fi:| which ends the conditional (with an empty expansion),
%       followed by \cs{@@_generate_variant_loop_long:wNNnn}, which
%       places an error as the second argument of
%       \cs{@@_generate_variant:wwNN}.
%     \item The loop can be interrupted early if the requested expansion
%       is unavailable, namely when the variant and base letters differ
%       and the base is neither |n| nor |N|.  Again, an error is placed
%       as the second argument of \cs{@@_generate_variant:wwNN}.
%   \end{itemize}
%   Note that if the variant form has the same length as the base form,
%   |#2| is as described in the first point, and |#4| as described in
%   the second point above.  The \cs{@@_generate_variant_loop_end:nwwwNNnn}
%   breaking function takes the empty brace group in |#4| as its first
%   argument: this empty brace group produces the correct signature for
%   the full variant.
%    \begin{macrocode}
\cs_new:Npn \@@_generate_variant_loop:nNwN #1#2#3 \q_mark #4
  {
    \if:w #2 #4
      \exp_after:wN \@@_generate_variant_loop_same:w
    \else:
      \if:w N #4 \else:
        \if:w n #4 \else:
          \@@_generate_variant_loop_invalid:NNwNNnn #4#2
        \fi:
      \fi:
    \fi:
    #1
    \prg_do_nothing:
    #2
    \@@_generate_variant_loop:nNwN { } #3 \q_mark
  }
\cs_new:Npn \@@_generate_variant_loop_same:w
    #1 \prg_do_nothing: #2#3#4
  {
    #3 { #1 \@@_generate_variant_same:N #2 }
  }
\cs_new:Npn \@@_generate_variant_loop_end:nwwwNNnn
    #1#2 \q_mark #3 ~ #4 \q_stop #5#6#7#8
  {
    \scan_stop: \scan_stop: \fi:
    \exp_not:N \q_mark
    \exp_not:N \q_stop
    \exp_not:N #6
    \exp_not:c { #7 : #8 #1 #3 }
  }
\cs_new:Npn \@@_generate_variant_loop_long:wNNnn #1 \q_stop #2#3#4#5
  {
    \exp_not:n
      {
        \q_mark
        \__msg_kernel_error:nnxx { kernel } { variant-too-long }
          {#5} { \token_to_str:N #3 }
        \use_none:nnnn
        \q_stop
        #3
        #3
      }
  }
\cs_new:Npn \@@_generate_variant_loop_invalid:NNwNNnn
    #1#2 \fi: \fi: \fi: #3 \q_stop #4#5#6#7
  {
    \fi: \fi: \fi:
    \exp_not:n
      {
        \q_mark
        \__msg_kernel_error:nnxxxx { kernel } { invalid-variant }
          {#7} { \token_to_str:N #5 } {#1} {#2}
        \use_none:nnnn
        \q_stop
        #5
        #5
      }
  }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}[aux, rEXP]{\@@_generate_variant_same:N}
%   When the base and variant letters are identical, don't do any
%   expansion.  For most argument types, we can use the |n|-type
%   no-expansion, but the |N| and |p| types require a slightly different
%   behaviour with respect to braces.
%    \begin{macrocode}
\cs_new:Npn \@@_generate_variant_same:N #1
  {
    \if:w N #1
      N
    \else:
      \if:w p #1
        p
      \else:
        n
      \fi:
    \fi:
  }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}[aux]{\@@_generate_variant:wwNN}
%   If the variant form has already been defined, log its existence.
%   Otherwise, make sure that the |\exp_args:N #3| form is defined, and
%   if it contains |x|, change \cs{@@_tmp:w} locally to
%   \cs{cs_new_protected:Npx}.  Then define the variant by
%   combining the |\exp_args:N #3| variant and the base function.
%    \begin{macrocode}
\cs_new_protected:Npn \@@_generate_variant:wwNN
    #1 \q_mark #2 \q_stop #3#4
  {
    #2
    \cs_if_free:NTF #4
      {
        \group_begin:
          \@@_generate_internal_variant:n {#1}
          \@@_tmp:w #4 { \exp_not:c { exp_args:N #1 } \exp_not:N #3 }
        \group_end:
      }
      {
        \__chk_log:x
          {
            Variant~\token_to_str:N #4~%
            already~defined;~ not~ changing~ it~ \msg_line_context:
          }
      }
  }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}[int]{\@@_generate_internal_variant:n}
% \begin{macro}[aux]{\@@_generate_internal_variant:wwnw}
% \begin{macro}[aux, rEXP]{\@@_generate_internal_variant_loop:n}
%   Test if |\exp_args:N #1| is already defined and if not define it via
%   the |\::| commands using the chars in |#1|.  If |#1| contains an |x|
%   (this is the place where having converted the original comma-list
%   argument to a string is very important), the result should be
%   protected, and the next variant to be defined using that internal
%   variant should be protected.
%    \begin{macrocode}
\cs_new_protected:Npx \@@_generate_internal_variant:n #1
  {
    \exp_not:N \@@_generate_internal_variant:wwnNwnn
      #1 \exp_not:N \q_mark
        { \cs_set_eq:NN \exp_not:N \@@_tmp:w \cs_new_protected:Npx }
        \cs_new_protected:cpx
      \token_to_str:N x \exp_not:N \q_mark
        { }
        \cs_new:cpx
    \exp_not:N \q_stop
      { exp_args:N #1 }
      {
        \exp_not:N \@@_generate_internal_variant_loop:n #1
          { : \exp_not:N \use_i:nn }
      }
  }
\use:x
  {
    \cs_new_protected:Npn \exp_not:N \@@_generate_internal_variant:wwnNwnn
        ##1 \token_to_str:N x ##2 \exp_not:N \q_mark
        ##3 ##4 ##5 \exp_not:N \q_stop ##6 ##7
  }
  {
    #3
    \cs_if_free:cT {#6} { #4 {#6} {#7} }
  }
%    \end{macrocode}
%   This command grabs char by char outputting |\::#1| (not expanded
%   further).  We avoid tests by putting a trailing |: \use_i:nn|, which
%   leaves \cs{cs_end:} and removes the looping macro.  The colon is in
%   fact also turned into \cs{:::} so that the required structure for
%   |\exp_args:N...| commands is correctly terminated.
%    \begin{macrocode}
\cs_new:Npn \@@_generate_internal_variant_loop:n #1
  {
    \exp_after:wN \exp_not:N \cs:w :: #1 \cs_end:
    \@@_generate_internal_variant_loop:n
  }
%    \end{macrocode}
% \end{macro}
% \end{macro}
% \end{macro}
%
%    \begin{macrocode}
%</initex|package>
%    \end{macrocode}
%
% \end{implementation}
%
% \PrintIndex