summaryrefslogtreecommitdiff
path: root/macros/latex/contrib/dialogl/dialogl.dtx
blob: 105a2494dd48ec3fb04dc86b0a16e76421e3257f (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
%%% ====================================================================
%%%  @LaTeX-style-file{
%%%     filename        = "dialogl.dtx",
%%%     version         = "1.99a",
%%%     date            = "2013/01/24",
%%%     author          = "Michael Downes",
%%%     copyright       = "This file is part of the dialogl package, released
%%%                        under the LPPL; see dialogl.ins for details."
%%%     keywords        = "TeX, dialog",
%%%     supported       = "no",
%%%     abstract        = "This file provides macros for writing
%%%       messages and menus on screen, and reading user responses. It
%%%       can be used with LaTeX as a documentstyle option, or in
%%%       other forms of TeX by a standard \input statement.",
%%%  }
%%% ====================================================================
%
% \iffalse
%<*driver>
\input{dia-driv.tex}
%</driver>
% \fi
%
% \section{History}
%    This file, \fn{dialog.sty}, was born out of a utility called
%    \fn{listout.tex} that I wrote for my personal use. The purpose of
%    \fn{listout.tex} was to facilitate printing out plain text
%    files\Dash electronic mail, program source files in various
%    programming languages, and, foremost, \tex/ macro files and log
%    files. An important part of my \tex/ programming practice is to
%    print out a macro file on paper and read it through, marking
%    corrections along the way, then use the marked copy as a script
%    for editing the file. (For one thing, this allows me to analyze
%    and mark corrections while riding the bus to work, or
%    sitting out in the back yard to supervise the kids.) The output I
%    normally desired was two `pages' per sheet of U.S. letter-size
%    paper printed landscape, in order to conserve paper.
%
%    Once created, \fn{listout.tex} quickly became my favorite means
%    of printing out plain text files, not to mention an indispensable
%    tool in my debugging toolbox: I turn on \cw{tracingmacros} and
%    \cw{tracingcommands}, then print out the resulting log file so
%    that I can see several hundred lines of the log at once (by
%    spreading out two or three pages on my desk with 100+ lines per
%    page); then I trace through, cross things out, label other
%    things, draw arrows, and so forth.
%
%    I soon added a filename prompting loop to make it convenient to
%    print multiple files in a single run. In the process of
%    perfecting this simple prompting routine\Dash over two or three
%    years\Dash and adding the ability to optionally specify things like
%    number of columns at run time, eventually I wrote so much
%    dialog-related macro code that it became clear this code should
%    be moved out of \fn{listout.tex} into its own module. The result
%    was \fn{dialog.sty}.
%
%    Before getting into the macro definitions and technical
%    commentary, here are descriptions from the user's
%    perspective of the functions defined in this file.
%
% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% \section{Message-sending functions}
%
%\begin{usage}
%\mesj{<text>}
%\end{usage}
%    Sends the message verbatim: category 12 for all special
%    characters except braces, tab characters, and carriage returns:
%
%^^V { } ^^I ^^M
%
%    Naturally, the catcode changes are effective only if \cw{mesj} is
%    used directly, not inside a macro argument or definition
%    replacement text.
%
%    Multiple spaces in the argument of \cw{mesj} print as multiple
%    spaces on screen. A tab character produces always eight spaces;
%    `smart' handling of tabs is more complicated than I would care to
%    attempt.
%
%    Line breaks in the argument of \cw{mesj} will produce line breaks
%    on screen. That is, you don't need to enter a special sequence
%    such as \ctrl{J}\qc{\%} to get line breaks. See the technical
%    commentary for \cw{mesjsetup} for details. Even though curly
%    braces are left with their normal catcodes, they can be printed
%    in a message without any problem, if they occur in balanced
%    pairs. If not, the message should be sent using \cw{xmesj}
%    instead of \cw{mesj}.
%
%    Because of its careful handling of the message text, \cw{mesj} is
%    extremely easy to use. The only thing you have to worry about is
%    having properly matched braces. Beyond that, you simply type
%    everything exactly as you want it to appear on screen.
%
%\begin{usage}
%\xmesj{<text>}
%\end{usage}
%    This is like \cw{mesj} but expands embedded control sequences
%    instead of printing them verbatim. All special characters have
%    category 12 except backslash, percent, braces, tab, return, and
%    ampersand:
%
%^^V \ % { } ^^I ^^M &
%
%    The first four have normal \tex/ catcodes to make it possible to
%    use most normal \tex/ commands, and comments, in the message
%    text. \ctrl{I} and \ctrl{M} are catcode 13 and behave as
%    described for \cw{mesj}. The \qc{\&} is a special convenience, an
%    abbreviation for \cw{noexpand}, to use for controlling expansion
%    inside the message text.
%
%    Doubled backslash \cs{\\} in the argument will produce a single
%    category 12 backslash character\Dash thus, \verb'\\xxx' can be used
%    instead of \cw{string}\cw{xxx} or \cw{noexpand}\cw{xxx} (notice
%    that this works even for outer things like \cw{bye} or
%    \cw{newif}). Similarly \cs{\%}, \cs{\{}, \cs{\}} and \cs{\&}
%    produce the corresponding single characters.
%
%    Category 12 space means that you cannot write something like
%
%^^V \ifvmode h\else v\fi rule
%
%    in the argument of \cw{xmesj} without getting a space after the
%    \cw{ifvmode}, \cw{else}, and \cw{fi}.%^^A
%^^A
%\footnote{Well, actually, you could replace each space by
%\qc{\%}\meta{newline} to get rid of it. But that makes the message
%text harder to read for the programmer.}
%^^A
%    Since occasionally this may be troublesome, \cs{\.} is defined
%    inside the argument of \cw{xmesj} to be a `control word
%    terminator': If the expansion of \cw{foo} is \verb'abc', then
%    \verb'\foo\.xyz' produces \verb'abcxyz' on screen (as opposed to
%    \verb'\foo xyz' which would produce \verb'abc xyz'). Thus the
%    above conditional could be written as
%
%^^V \ifvmode\.h\else\.v\fi\.rule
%
%    Even though the catcode changes done by \cw{xmesj} setup have no
%    effect if \cw{xmesj} is used inside an argument or definition
%    replacement text, I find it convenient occasionally to use
%    \cw{xmesj} in those contexts, in order to get other aspects of the
%    \cw{xmesj} setup. For instance, if you need to embed a message
%    that contains a percent sign inside a definition, you can write
%
%^^V \def\foo{...
%^^V   \xmesj{... this is a percent
%^^V     sign: \% (sans backslash) ...}
%^^V ...}
%
%    To further support such uses of \cw{xmesj}, the following changes
%    are also done by \cw{xmesj} setup: the backslash-space control
%    symbol {\tt\bslash\char32} is made equivalent to \cw{space};
%    \cs{\^^J} and \cs{\^^M} are defined to produce a \cw{newlinechar};
%    and active tilde \qc{\~} will produce a category-12 tilde.
%
%    Among other things, this setup makes it easier to obtain
%    newlines and multiple spaces in an embedded message. For example,
%    in the following definition the message will have a line break on
%    screen for each backslash at the end of a line, and the third
%    line will be indented four spaces.
%
%^^V \def\bar{...
%^^V   \xmesj{First line\
%^^V     Second line\
%^^V     \ \ \ \ Indented line\
%^^V     Last line}%
%^^V ...}
%
%    The alternative of defining a separate message function
%    \cw{barfoo} with \cw{f[x]mesj} and calling \cw{barfoo} inside of
%    \cw{bar} would allow more natural entry of the newlines and the
%    multiple spaces, but would be slightly more expensive in string
%    pool and hash table usage.
%
%\begin{usage}
%\promptmesj{<text>}
%\promptxmesj{<text>}
%\end{usage}
%    These are like \cw{mesj}, \cw{xmesj} but use \cw{message} rather
%    than \cw{immediate}\cw{write}\verb'16' internally, thus if the
%    following operation is a \cw{read}, the user will see the cursor
%    on screen at the end of the last line, as may be desired when
%    prompting for a short reply, rather than at the beginning of the
%    next line. The character \qc{\!} is preempted internally for
%    newlinechar, for these two functions only,
%    which means that it cannot be actually printed in
%    the message text. Use of a visible character such as \qc{\!},
%    rather than the normal \cw{newlinechar} \ctrl{J}, is necessary
%    for robustness because of the fact that the \cw{message}
%    primitive was unable to use an `invisible' character (outside the
%    range 32--126) for newlines up until \tex/ version 3.1415, which
%    some users do not yet have (at the time of this writing\Dash July
%    1994).
%
%\begin{usage}
%\storemesj\foo{<text>}
%\storexmesj\foo{<text>}
%\end{usage}
%    These functions are similar to \cw{mesj}, \cw{xmesj} but store the
%    given text in the control sequence \cw{foo} instead of immediately
%    sending the message. Standard \tex/ parameter syntax can be used
%    to make \cw{foo} a function with arguments, e.g. after
%
%^^V \storemesj\foo#1{...#1...}
%
%    then you can later write
%
%^^V \message{\foo{\the\hsize}}
%
%    and get the current value of \cw{hsize} into the middle of the
%    message text. Consequently also in the x-version \cw{storexmesj} a
%    category-12 \qc{\#} character can be obtained with \cs{\#}.
%
%\begin{usage}
%\fmesj\foobar#1#2...{...#1...#2...}
%\end{usage}
%    Defines \cw{foobar} as a function that will take the given
%    arguments, sow them into the message text \verb"{...}", and send
%    the message. In the message text all special characters are
%    category 12 except for braces, \qc{\#}, tab, and carriage return.
%
%    If an unmatched brace or a \qc{\#} must be printed in the message
%    text \cw{fxmesj} must be used instead. (\arg{#} could be used to
%    insert a single category-6 \qc{\#} token into the message text, and
%    \tex/ would print it without an error, but both \cw{message} and
%    \cw{write} would print it as two \arg{#} characters, even though
%    it's only a single token internally.)
%
%\begin{usage}
%\fxmesj\foobar#1#2...{...#1...#2...}
%\end{usage}
%    Combination of \cw{xmesj} and \cw{fmesj}. Defines \cw{foobar} like
%    \cw{fmesj}, but with full expansion of the replacement text and
%    with normal category codes for backslash, percent, braces, and hash
%    \qc{\#}. The control symbols \cs{\\} \cs{\%} \cs{\{} \cs{\}}
%    \cs{\&} and \cs{\.} can be used as in \cw{xmesj}, with also \cs{\#}
%    for printing a \qc{\#} character of category 12.
%
% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% \section{Reading functions}
%
%\begin{usage}
%\readline{<default>}\answer
%\end{usage}
%    This reads a line of input from the user into the macro
%    \cw{answer}. (The macro name can be anything chosen by the
%    programmer, not just \cw{answer}.) Before reading, all special
%    characters are deactivated, so that the primitive \cw{read} will
%    not choke if the user happens to enter something like \cw{newif}
%    or {\sc control-l} or \qc{\}}. Depending on the operating
%    system, certain characters\Dash e.g., {\sc control-c}, {\sc
%    control-z}, {\sc control-d}, {\sc control-h}\Dash might have
%    special effects instead of being entered into the replacement
%    text of \cw{answer}, regardless of the catcode changes. To
%    take the most obvious example, under most
%    operating systems, typing {\sc control-h} (the Rubout or
%    Backward-Delete key) will delete the previous character from the
%    user's response, instead of entering an \ascii/ character 8
%    into \cw{answer}.
%
%    There is one significant exception from the catcode changes that
%    are done for \cw{readline}: spaces and tabs retain their normal
%    catcode of 10, so that multiple spaces in an answer will be
%    reduced to a single space, and macros with normal space-delimited
%    arguments will work when applied to the answer. (I can't think of
%    any likely scenario where category 12 for spaces would be
%    useful.) Also, the catcode of \ctrl{M} is set to 9 (ignore) so
%    that an empty line\Dash meaning that the user just pressed the
%    carriage return/enter key\Dash will result in an empty \cw{answer}.
%    If the answer is empty, the given default string will be
%    substituted. The default string can be empty.
%
%\begin{usage}
%\xreadline{<default>}\answer
%\end{usage}
%    Like \cw{readline} but the answer is read as executable tokens;
%    the usual catcodes of the \tex/ special characters remain in
%    effect while reading the answer. A few common outer things
%    (\cw{bye}, \cs{\+}, \cw{newif}, \ctrl{L}, among others) are
%    neutralized before the \cw{read} is done, but the user can still
%    cause problems by entering some other outer control sequence or
%    unbalanced braces. I doubt there's any bulletproof solution, if
%    the tokens are to remain executable, short of the usual last
%    resort: reading the answer using \cw{readline}, writing it to a
%    file, then inputting the file.
%
%\begin{usage}
%\readchar{<default>}\answer
%\end{usage}
%    This is like \cw{readline} but it reduces the answer to its first
%    character. \meta{default} is either a single character or empty.
%
%\begin{usage}
%\readChar{<default>}\answer
%\end{usage}
%    This is like \cw{readchar} and also uppercases the answer.
%
%\begin{usage}
%\changecase\uppercase\answer
%\end{usage}
%    The function \cw{changecase} redefines its second argument, which
%    must be a macro, to contain the same text as before, but
%    uppercased or lowercased according to the first argument. Thus
%    \cw{readChar}\verb'{Q}'\cw{answer} is equivalent to
%
%^^V \readchar{q}\answer
%^^V \changecase\uppercase\answer
%
%    It might sometimes be desirable to force lower case before using
%    a file name given by the user, for example.
%
% %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% \section{Checking functions}
%
%\begin{usage}
%\checkinteger\reply\tempcount
%\end{usage}
%    To read in and check an answer that is supposed to be an integer,
%    use \cw{readline}{}\cw{reply} and then apply \cw{checkinteger} to
%    the \cw{reply}, supplying a count register, not necessarily named
%    \cw{tempcount}, wherein \cw{checkinteger} will leave the validated
%    integer. If \cw{reply} does not contain a valid integer the
%    returned value will be \verb'-'\cw{maxdimen}.
%
%    At the present time only decimal digits are handled;
%    some valid \tex/ numbers such as \verb'"AB', \verb'`\@',
%    \cw{number}\cw{prevgraf}, or a count register name, will not be
%    recognized by \cw{checkinteger}. There seems to be no bulletproof
%    way to allow these possibilities.
%
%    Tests that hide \cw{checkinteger} under the hood, such as a
%    \cw{nonnegativeinteger} test, are not provided because as often as
%    not the number being prompted for will have to be tested to see if
%    it falls inside a more specific range, such as 0\dash 255 for an
%    8-bit number or 1\dash 31 for a date, and it seems common sense to
%    omit overhead if it would usually be redundant. It's easy enough to
%    define such a test for yourself, if you want one.
%
%\begin{usage}
%\checkdimen\reply\tempdim
%\end{usage}
%    Analog of \cw{checkinteger} for dimension values. If \cw{reply}
%    does not contain a valid dimension the value returned in
%    \cw{tempdim} will be \verb'-'\cw{maxdimen}.
%
%    Only explicit dimensions with decimal digits, optional decimal
%    point and more decimal digits, followed by explicit units
%    \verb'pt' \verb'cm' \verb'in' or whatever are checked for; some
%    valid \tex/ dimensions such as \cw{parindent},
%    \verb'.3'\cw{baselineskip}, or \cw{fontdimen}\verb'5'\cw{font}
%    will not be recognized by \cw{checkdimen}.
%
% \subsection*{What good is all this?}
%    What good is all this stuff, practically speaking?\Dash you may
%    ask. Well, a typical application might be something like: At the
%    beginning of a document, prompt interactively to find out if the
%    user wants to print on A4 or US letter-size paper, or change the
%    top or left margin. Such a query could be done like this:
%
%^^V \promptxmesj{
%^^V Do you want to print on A4 or US letter paper?
%^^V Enter u or U for US letter, anything else for A4: }
%^^V \readChar{A}\reply % default = A4
%^^V \if U\reply \textheight=11in \textwidth=8.5in
%^^V \else \textheight=297mm \textwidth=210mm \fi
%^^V %    Subtract space for 1-inch margins
%^^V \addtolength{\textheight}{-2in}
%^^V \addtolength{\textwidth}{-2in}
%^^V
%^^V \promptxmesj{
%^^V Left margin setting? [Return = keep current value,
%^^V \the\oddsidemargin]: }
%^^V \readline{\the\oddsidemargin}\reply
%^^V \checkdimen\reply{\dimen0}
%^^V \ifdim\dimen0>-\maxdimen
%^^V   \setlength\oddsidemargin{\dimen0}%
%^^V   \xmesj{OK, using new left margin of %
%^^V \the\oddsidemargin.}
%^^V \else
%^^V   \xmesj{Sorry, I don't understand %
%^^V that reply: `\reply'.\
%^^V Using default value: \the\oddsidemargin.}
%^^V \fi
%
%    Although \latex/'s \cw{typeout} and \cw{typein} functions can
%    be used for this same task, they are rather more awkward, and
%    checking the margin value for validity would be quite difficult.
%
% \StopEventually{}
%
% \section{Implementation}
%    Standard package identification:
%    \begin{macrocode}
%<*2e>
\NeedsTeXFormat{LaTeX2e}
\ProvidesPackage{dialog}[1994/11/08 v0.9y]
%</2e>
%    \end{macrocode}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% \section{Preliminaries}
%
%    \begin{macrocode}
%<*2e>
\RequirePackage{grabhedr}
%</2e>
%    \end{macrocode}
%
%    If \fn{grabhedr.sty} is not already loaded, load it now. The
%    \cw{trap.input} function is explained in \fn{grabhedr.doc}.
%    \begin{macrocode}
%<*209>
\csname trap.input\endcsname
\input grabhedr.sty \relax
\fileversiondate{dialog.sty}{0.9y}{1994/11/08}%
%</209>
%    \end{macrocode}
%
%    The functions \cw{localcatcodes} and \cw{restore\-catcodes} are
%    defined in \fn{grabhedr.sty}. We use them to save and restore
%    catcodes of any special characters needed in this file whose
%    current catcodes might not be what we want them to be. Saving and
%    restoring catcode of at-sign \qc{\@} makes this file work equally
%    well as a \latex/ documentstyle option or as a simple input file in
%    other contexts. The double quote character \qc{\"} might be active
%    for German and other languages. Saving and restoring tilde \qc{\~},
%    hash \qc{\#}, caret \qc{\^}, and left quote \qc{\`} catcodes is
%    normally redundant but reduces the number of assumptions we must
%    rely on. (The following catcodes are assumed: \qc{\\} 0, \qc{\{} 1,
%    \qc{\}} 2, \qc{\%} 14, {\tt a}\dash {\tt z} {\tt A}\dash {\tt Z}
%    11, {\tt 0}\dash {\tt 9} \qc{\.} \qc{\-} 12. Also note that
%    \cw{endlinechar} is assumed to have a non-null value.)
%    \begin{macrocode}
%% The line break is significant here:
\localcatcodes{\@{11}\ {10}\
{5}\~{13}\"{12}\#{6}\^{7}\`{12}}
%    \end{macrocode}
%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% \section{Definitions}
%
% \begin{macro}{\otherchars}
%
%    For deactivating characters with special catcodes during
%    \cw{readline} we use, instead of \cw{dospecials}, a more
%    bulletproof (albeit slower) combination of \cw{otherchars},
%    \cw{controlchars}, and \cw{highchars} that covers all characters in
%    the range 0\dash 255 except letters and digits. Handling the
%    characters above 127 triples the overhead done for each read
%    operation or message definition but seems mandatory for maximum
%    robustness.^^A
%^^A%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% \footnote{If you are using \fn{dialog.sty} functions on a slow
% computer, you might want to try setting \cw{highchars} = empty to
% see if that helps the speed.}
%^^A%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%
%    \cw{otherchars} includes the thirty-three nonalphanumeric visible
%    characters (counting space as visible). It is intended as an
%    executable list like \cw{dospecials} but, as an exercise in
%    memory conservation, it is constructed without the \cw{do}s. For
%    the usual application of changing catcodes, the list can still be
%    executed nicely as shown below. Also, if we arrange to
%    make sure that each character token gets category 12, it's not
%    necessary to use control symbols such as \cs{\%} in place of
%    those few special characters that would otherwise be difficult to
%    place inside of a definition. This avoids a problem that would
%    otherwise arise if we included \cs{\+} in the list and tried to
%    process the list with a typical definition of \cs{do}: \cs{\+} is
%    `outer' in plain \tex/ and would cause an error message when
%    \cw{do} attempted to read it as an argument. (As a matter of fact
%    the catcode changes below show a different way around that
%    problem, but a list of category-12 character tokens is a fun
%    thing to have around anyway.)
%
%    \begin{macrocode}
\begingroup
%    \end{macrocode}
%    First we start a group to localize \cw{catcode} changes. Then we
%    change all relevant catcodes to 12 except for backslash, open
%    brace, and close brace, which can be handled by judicious
%    application of \cw{escapechar}, \cw{string}, \cw{edef}, and
%    \cw{xdef}. By defining \cw{do} in a slightly backward way, so that
%    it doesn't take an argument, we don't need to worry about the
%    presence of \cs{\+} in the list of control symbols. Notice the
%    absence of \cs{\`} from the list of control symbols; it was already
%    catcoded to 12 in the \cw{localcatcodes} declaration at the
%    beginning of this file\Dash otherwise it would be troublesome to
%    make the definition of \cw{do} bulletproof (consider the
%    possibilities that \qc{\`} might have catcode 0, 5, 9, or 14).
%
%    \begin{macrocode}
\def\do{12 \catcode`}
\catcode`\~\do\!\do\@\do\#\do\$\do\^\do\&
\do\*\do\(\do\)\do\-\do\_\do\=\do\[\do\]
\do\;\do\:\do\'\do\"\do\<\do\>\do\,\do\.
\do\/\do\?\do\|12\relax
%    \end{macrocode}
%    To handle backslash and braces, we define \cs{\\},
%    \cs{\{}, and \cs{\}} to produce the corresponding category-12
%    character tokens. Setting \cw{escapechar} to $-1$ means that
%    \cw{string} will omit the leading backslash that it would
%    otherwise produce when applied to a control sequence.
%    \begin{macrocode}
\escapechar -1
\edef\\{\string\\}
\edef\{{\string\{}\edef\}{\string\}}
%    \end{macrocode}
%    Space and percent are done last. Then, with almost all the
%    special characters now category 12, it's rather easy to define
%    \cw{otherchars}.
%    \begin{macrocode}
\catcode`\ =12\catcode`\%=12
\xdef\otherchars
{ !"#$%&'()*+,-./:;<=>?[\\]^_`\{|\}~}
\endgroup %              ^     ^  ^
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\controlchars}
%
%    \cw{controlchars} is another list for the control characters
%    \ascii/ 0\dash 31 and 127. The construction of this list is similar
%    to the construction of \cw{otherchars}. We need to turn off
%    \cw{endlinechar} because the catcode of \ctrl{M} is going to be
%    changed. The \ctrl{L} inside the \cw{gdef} is not a problem (as it
%    might have been, due to the usual outerness of \ctrl{L}) because
%    the catcode is changed from 13 to 12 before that point.
%
%    \begin{macrocode}
\begingroup
\endlinechar = -1
\def\do{12 \catcode`}
\catcode`\^^@\do\^^A\do\^^B\do\^^C
\do\^^D\do\^^E\do\^^F\do\^^G\do\^^H\do\^^I
\do\^^J\do\^^K\do\^^L\do\^^M\do\^^N\do\^^O
\do\^^P\do\^^Q\do\^^R\do\^^S\do\^^T\do\^^U
\do\^^V\do\^^W\do\^^X\do\^^Y\do\^^Z\do\^^[
\do\^^\\do\^^]\do\^^^\do\^^_\do\^^? 12\relax
%
\gdef\controlchars{^^@^^A^^B^^C^^D^^E^^F^^G
  ^^H^^I^^J^^K^^L^^M^^N^^O^^P^^Q^^R^^S^^T
  ^^U^^V^^W^^X^^Y^^Z^^[^^\^^]^^^^^_^^?}
\endgroup
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\highchars}
%    And finally, the list \cw{highchars} contains characters
%    128\dash 255, the ones that have the eighth bit set.
%
%    \begin{macrocode}
\begingroup
\def\do{12 \catcode`}
\catcode`\^^80\do\^^81\do\^^82\do\^^83\do\^^84
\do\^^85\do\^^86\do\^^87\do\^^88\do\^^89\do\^^8a
\do\^^8b\do\^^8c\do\^^8d\do\^^8e\do\^^8f
\do\^^90\do\^^91\do\^^92\do\^^93\do\^^94\do\^^95
%    \end{macrocode}
%\verbdots\iffalse
\do\^^96\do\^^97\do\^^98\do\^^99\do\^^9a\do\^^9b
\do\^^9c\do\^^9d\do\^^9e\do\^^9f
\do\^^a0\do\^^a1\do\^^a2\do\^^a3\do\^^a4\do\^^a5
\do\^^a6\do\^^a7\do\^^a8\do\^^a9\do\^^aa\do\^^ab
\do\^^ac\do\^^ad\do\^^ae\do\^^af
\do\^^b0\do\^^b1\do\^^b2\do\^^b3\do\^^b4\do\^^b5
\do\^^b6\do\^^b7\do\^^b8\do\^^b9\do\^^ba\do\^^bb
\do\^^bc\do\^^bd\do\^^be\do\^^bf
\do\^^c0\do\^^c1\do\^^c2\do\^^c3\do\^^c4\do\^^c5
\do\^^c6\do\^^c7\do\^^c8\do\^^c9\do\^^ca\do\^^cb
\do\^^cc\do\^^cd\do\^^ce\do\^^cf
\do\^^d0\do\^^d1\do\^^d2\do\^^d3\do\^^d4\do\^^d5
\do\^^d6\do\^^d7\do\^^d8\do\^^d9\do\^^da\do\^^db
\do\^^dc\do\^^dd\do\^^de\do\^^df
\do\^^e0\do\^^e1\do\^^e2\do\^^e3\do\^^e4\do\^^e5
\do\^^e6\do\^^e7\do\^^e8\do\^^e9\do\^^ea\do\^^eb
\do\^^ec\do\^^ed\do\^^ee\do\^^ef
\do\^^f0\do\^^f1\do\^^f2\do\^^f3\do\^^f4\do\^^f5
\do\^^f6\do\^^f7\do\^^f8\do\^^f9\do\^^fa\do\^^fb
%\fi
%    \begin{macrocode}
\do\^^fc\do\^^fd\do\^^fe\do\^^ff 12\relax
%
\gdef\highchars{%
^^80^^81^^82^^83^^84^^85^^86^^87^^88%
^^89^^8a^^8b^^8c^^8d^^8e^^8f%
^^90^^91^^92^^93^^94^^95^^96^^97^^98%
%    \end{macrocode}
%\verbdots\iffalse
^^99^^9a^^9b^^9c^^9d^^9e^^9f%
^^a0^^a1^^a2^^a3^^a4^^a5^^a6^^a7^^a8%
^^a9^^aa^^ab^^ac^^ad^^ae^^af%
^^b0^^b1^^b2^^b3^^b4^^b5^^b6^^b7^^b8%
^^b9^^ba^^bb^^bc^^bd^^be^^bf%
^^c0^^c1^^c2^^c3^^c4^^c5^^c6^^c7^^c8%
^^c9^^ca^^cb^^cc^^cd^^ce^^cf%
^^d0^^d1^^d2^^d3^^d4^^d5^^d6^^d7^^d8%
^^d9^^da^^db^^dc^^dd^^de^^df%
^^e0^^e1^^e2^^e3^^e4^^e5^^e6^^e7^^e8%
^^e9^^ea^^eb^^ec^^ed^^ee^^ef%
^^f0^^f1^^f2^^f3^^f4^^f5^^f6^^f7^^f8%
%\fi
%    \begin{macrocode}
^^f9^^fa^^fb^^fc^^fd^^fe^^ff}
\endgroup
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\actively}
%
%    The function \cw{actively} makes a given character active and
%    carries out the assignment given as the first argument. The
%    assignment can be embedded in the replacement text of a macro
%    definition without requiring any special setup to produce an active
%    character in the replacement text. The argument should be a control
%    symbol, e.g. \cw{@} or \cs{\#} or \cs{\^^M}, rather than a single
%    character. (Except that \qc{\+} is safer than \cs{\+} in
%    \plaintex/.) If the assignment is a definition (\cw{def},
%    \cw{edef}, \cw{gdef}, \cw{xdef}) it is allowed to take arguments in
%    the normal \tex/ way. Prefixes such as \cw{global}, \cw{long}, or
%    \cw{outer} must go inside the first argument rather than before
%    \cw{actively}.
%
%    Usage:
%\begin{usage}
%\actively\def\?{<replacement text>}
%\actively\def\%#1#2{<replacement text>}
%\actively{\global\let}\^^@=\space
%\end{usage}
%    One place where this function can be put to good use is in
%    making \ctrl{M} active in order to get special action at the end
%    of each line of input. The usual way of going about this would be
%    to write
%
%^^V \def\par{something}\obeylines
%
%    \noindent which is a puzzling construction to the \tex/ novice
%    who doesn't know what \cw{obeylines} does with \cw{par}. The same
%    effect could be gotten a little more transparently with
%
%^^V \actively\def\^^M{something}
%
%    In the definition of \cw{actively} we use the unique properties
%    of \cw{lowercase} to create an active character with the right
%    character code, overlapping with a \cw{begingroup} \cw{endgroup}
%    structure that localizes the necessary lc-code change.
%    \begin{macrocode}
\def\actively#1#2{\catcode`#2\active
  \begingroup \lccode`\~=`#2\relax
  \lowercase{\endgroup#1~}}
%    \end{macrocode}
% \end{macro}
%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%
% \begin{macro}{\mesjsetup}
%
%    The \cw{mesjsetup} function starts a group to localize catcode
%    changes. The group will be closed eventually by a separate
%    function that does the actual sending or stores the message text
%    for later retrieval.
%
%    We want to change the catcode of each character in the three
%    lists \cw{otherchars}, \cw{controlchars}, and \cw{highchars} to
%    12. After giving \cw{do} a recursive definition, we apply it to
%    each of the three lists, adding a suitable element at the end of
%    the list to make the recursion stop there. This allows
%    leaving out the \cw{do} tokens from the character lists, without
%    incurring the cost of an if test at each recursion step.
%    \begin{macrocode}
\def\mesjsetup{\begingroup \count@=12
  \def\do##1{\catcode`##1\count@ \do}%
%    \end{macrocode}
%    The abbreviation \cw{xp@} = \cw{expandafter} is from
%    \fn{grabhedr.sty}.
%    \begin{macrocode}
  \xp@\do\otherchars{a11 \@gobbletwo}%
  \xp@\do\controlchars{a11 \@gobbletwo}%
  \xp@\do\highchars{a11 \@gobbletwo}%
%    \end{macrocode}
%    Make the tab character produce eight spaces:
%    \begin{macrocode}
  \actively\edef\^^I{ \space\space\space
    \space\space\space\space}%
%    \end{macrocode}
%    The convenient treatment of newlines in the argument of \cw{mesj}
%    (every line break produces a line break on screen) is achieved by
%    making the \ctrl{M} character active and defining it to produce a
%    category-12 \ctrl{J} character. Although for \cw{mesj} it would
%    have sufficed to make \ctrl{M} category 12 and locally set
%    \cw{newlinechar} = \ctrl{M} while sending the message, it turns out
%    to be useful for other functions to have the \ctrl{M} character
%    active, so that it can be remapped to an arbitrary function for
%    handling new lines (e.g., perhaps adding extra spaces at the
%    beginning of each line). And if \cw{mesj} treats \ctrl{M} the same,
%    we can arrange for it to share the setup routines needed for the
%    other functions.
%    \begin{macrocode}
  \endlinechar=`\^^M\actively\let\^^M=\relax
  \catcode`\{=1 \catcode`\}=2 }
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\sendmesj}
%
%    In \cw{sendmesj} we go to a little extra trouble to make sure
%    \ctrl{M} produces a newline character, no matter what the value of
%    \cw{newlinechar} might be in the surrounding environment. The
%    impending \cw{endgroup} will restore \cw{newlinechar} to its
%    previous value. One reason for using \ctrl{J} (instead of, say,
%    \ctrl{M} directly) is to allow e.g. \cw{mesj}\verb'{xxx^^Jxxxx}' to
%    be written inside a definition, as is sometimes convenient. This
%    would be difficult with \ctrl{M} instead of \ctrl{J} because of
%    catcodes.
%
%    \begin{macrocode}
\def\sendmesj{\newlinechar`\^^J%
  \actively\def\^^M{^^J}%
  \immediate\write\sixt@@n{\mesjtext}\endgroup}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\mesj}
%
%    Given the support functions defined above, the definition of
%    \cw{mesj} is easy: Use \cw{mesjsetup} to clear all special
%    catcodes, then set up \cw{sendmesj} to be triggered by the next
%    assignment, then read the following balanced-braces group into
%    \cw{mesjtext}. As soon as the definition is completed, \tex/ will
%    execute \cw{sendmesj}, which will send the text and close the
%    group that was started in \cw{mesjsetup} to localize the catcode
%    changes.
%
%    \begin{macrocode}
\def\mesj{\mesjsetup \afterassignment\sendmesj
  \def\mesjtext}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\sendprompt}
%
%    The \cw{sendprompt} function is just like \cw{sendmesj} except
%    that it uses \cw{message} instead of \cw{write}, as might be
%    desired when prompting for user input, so that the on-screen
%    cursor stays on the same line as the prompt instead of hopping
%    down to the beginning of the next line. In order for newlines to
%    work with \cw{message} we must use a visible character instead of
%    \ctrl{J}. When everyone has \tex/ version 3.1415 or later this
%    will no longer be true. The choice of \qc{\!} might be construed
%    (if you wish) as editorial comment that \qc{\!} should not be
%    shouted at the user in a prompt.
%
%    \begin{macrocode}
\def\sendprompt{%
  \newlinechar`\!\relax \actively\def\^^M{!}%
  \message{\mesjtext}\endgroup}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\promptmesj}
%
%    This function is like \cw{mesj} but uses \cw{sendprompt} instead of
%    \cw{sendmesj}.
%
%    \begin{macrocode}
\def\promptmesj{\mesjsetup
  \afterassignment\sendprompt \def\mesjtext}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\storemesj}
%
%    Arg \arg{1} of \cw{storemesj} is the control sequence under which
%    the message text is to be stored.
%    \begin{macrocode}
\def\storemesj#1{\mesjsetup
  \catcode`\#=6 % to allow arguments if needed
  \afterassignment\endgroup
  \long\gdef#1}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\fmesj}
%
%    While \cw{storemesj}\cw{foo}\verb"{...}" is more or less the
%    same as \cw{def}\cw{foo}\verb"{...}" with special catcode
%    changes, \cw{fmesj}\cw{foo}\verb"{...}" corresponds to
%    \cw{def}\cw{foo}\qc{\{}\cw{mesj}\verb"{...}}", that is, after
%    \cw{fmesj}\cw{foo} the function \cw{foo} can be executed directly
%    to send the message. Thus \cw{storemesj} is typically used for
%    storing pieces of messages, while \cw{fmesj} is used for storing
%    entire messages.
%
%    To read the parameter text \arg{2}, we use the peculiar
%    \verb'#{' feature of \tex/ to read everything up to the opening
%    brace.
%    \begin{macrocode}
\def\fmesj#1#2#{\mesjsetup
  \catcode`\#=6 % restore to normal
%    \end{macrocode}
%    The parameter text \arg{2} must be stored in a token register
%    rather than a macro to avoid problems with \qc{\#} characters.
%    The \cw{long} prefix is just to admit the (unlikely) possibility
%    of using \cw{fmesj} to define something such as an error message
%    saying `You can't use \arg{1} here' where one of the
%    possibilities for \arg{1} is \cw{string}\cw{par}.
%    \begin{macrocode}
  \toks@{\long\gdef#1#2}%
%    \end{macrocode}
%    Define \cw{@tempa} to put together the first two arguments and
%    [pseudo]argument \arg{3} and make the definition of \arg{1}.
%    \begin{macrocode}
  \def\@tempa{%
    \edef\@tempa{%
      \the\toks@{%
%    \end{macrocode}
%    The abbreviation \cw{nx@} = \cw{noexpand} is from
%    \fn{grabhedr.sty}.
%    \begin{macrocode}
      \begingroup\def\nx@\mesjtext{\the\toks2 }%
        \nx@\sendmesj}%
    }%
    \@tempa
    \endgroup % Turn off the \mesjsetup catcodes
  }%
  \afterassignment\@tempa
  \toks2=}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\xmesjsetup}
%
%    \cw{xmesjsetup} is like \cw{mesjsetup} except it prepares to allow
%    control sequence tokens and normal comments in the message text.
%    For \tex/nicians' convenience certain other features are
%    thrown in.
%
%    Here, unlike the setup for \cw{xreadline}, I don't bother to
%    remove the outerness of \cw{bye}, \cw{newif}, etc., because I
%    presume the arguments of \cw{xmesj}, \cw{fxmesj},
%    \cw{storexmesj}, \cw{fxmenu}, etc.\ are more likely to be written
%    by a \tex/nician than by an average end user, whereas
%    \cw{xreadline} is designed to handle arbitrary input from
%    arbitrary users.
%
%    \begin{macrocode}
\def\xmesjsetup{\mesjsetup
%    \end{macrocode}
%    Throw in pseudo braces just in case we are inside an \cw{halign}
%    with \cs{\\} let equal to \cw{cr} at the time when \cw{xmesjsetup}
%    is called. (As might happen in \amstex/.)
%    \begin{macrocode}
  \iffalse{\fi
  \catcode`\\=0 \catcode`\%=14
%    \end{macrocode}
%    Define {\def~{\spacefactor1500 \space}%
%    \cs{\%}~\cs{\\}~\cs{\{}~\cs{\}}~\cs{\&}} to produce the
%    corresponding single characters, category 12. The \cw{lowercase}
%    trick here allows these definitions to be nonglobal.
%    \begin{macrocode}
  \begingroup \lccode`\0=`\\\lccode`\1=`\{%
  \lccode`\2=`\}\lccode`\3=`\%%
  \lowercase{\endgroup \def\\{0}\def\{{1}%
    \def\}{2}\def\%{3}}%
  \iffalse}\fi
  \edef\&{\string &}%
%    \end{macrocode}
%    Let \qc{\&} \qc{\=} \cw{noexpand} for expansion control inside the
%    argument text; let active \ctrl{M} \qc{\=} \cw{relax} so that
%    newlines will remain inert during the expansion.
%    \begin{macrocode}
  \actively\let\&=\noexpand
  \actively\let\^^M=\relax
%    \end{macrocode}
%    Define \cs{\.} to be a noop, for terminating a control word when
%    it is followed by letters and no space is wanted.
%    \begin{macrocode}
  \def\.{}%
%    \end{macrocode}
%    Support for use of \cw{xmesj} inside a definition replacement
%    text or macro argument: control-space
%    \verb*"\ " = \cw{space}, tilde \qc{\~} prints as itself,
%    \cs{\^^M} (i.e., a lone backslash at the end of a line) will
%    produce a newline, also \cs{\^^J}, while finally \cw{par} \qc{\=}
%    blank line translates to two newlines.
%    \begin{macrocode}
  \def\ { }\edef~{\string ~}%
%    \end{macrocode}
%    Define \cs{\^^M} to produce an active \ctrl{M} character, which (we
%    hope) will be suitably defined to produce a newline or whatever.
%    \begin{macrocode}
  \begingroup \lccode`\~=`\^^M%
    \lowercase{\endgroup \def\^^M{~}}%
  \let\^^J\^^M \def\par{\^^M\^^M}%
}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\xmesj}
%
%    \cw{xmesj} uses \cw{xmesjsetup} and \cw{edef}.
%    \begin{macrocode}
\def\xmesj{\xmesjsetup \afterassignment\sendmesj
  \edef\mesjtext}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\promptxmesj}
%
%    \cw{promptxmesj} is analogous to \cw{promptmesj}, but with
%    expansion.
%    \begin{macrocode}
\def\promptxmesj{\xmesjsetup
  \afterassignment\sendprompt \edef\mesjtext}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\storexmesj}
%
%    And \cw{storexmesj} is like \cw{storemesj}, with expansion. Since
%    we allow arguments for the function being defined, we also must
%    define \cs{\#} to produce a single category-12 \qc{\#} character so
%    that there will be a way to print \qc{\#} in the message text.
%    \begin{macrocode}
\def\storexmesj#1#2#{\xmesjsetup
  \catcode`\#=6 % to allow arguments if needed
  \edef\#{\string##}%
  \afterassignment\endgroup
  \long\xdef#1#2}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\fxmesj}
%
%    And \cw{fxmesj} is the expansive analog of \cw{fmesj}.
%    \begin{macrocode}
\def\fxmesj#1#2#{\xmesjsetup
  \catcode`\#=6 % restore to normal
  \edef\#{\string##}%
  \toks@{\long\xdef#1#2}%
  \def\@tempa{%
    \edef\@tempa{%
      \the\toks@{\begingroup
      \def\nx@\nx@\nx@\mesjtext{\the\toks\tw@}%
      \nx@\nx@\nx@\sendmesj}}%
    \@tempa % execute the constructed xdef
    \endgroup % restore normal catcodes
  }%
  \afterassignment\@tempa
  \toks\tw@=}
%    \end{macrocode}
% \end{macro}
%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
% \section{Reading functions}
%
% \begin{macro}{\readline}
%    The \cw{readline} function gets one line of input from the user.
%    Arguments are: \arg{1} default to be used if the user response is
%    empty (i.e., if the user just pressed the return/enter key),
%    \arg{2} macro to receive the input.
%
%    \begin{macrocode}
\def\readline#1#2{%
  \begingroup \count@ 12 %
  \def\do##1{\catcode`##1\count@ \do}%
  \xp@\do\otherchars{a11 \@gobbletwo}%
  \xp@\do\controlchars{a11 \@gobbletwo}%
  \xp@\do\highchars{a11 \@gobbletwo}%
%    \end{macrocode}
%    Make spaces and tabs normal instead of category 12.
%    \begin{macrocode}
  \catcode`\ =10 \catcode`\^^I=10 %
  \catcode`\^^M=9 % ignore
%    \end{macrocode}
%    Reset end-of-line char to normal, just in case.
%    \begin{macrocode}
  \endlinechar`\^^M
%    \end{macrocode}
%    We go to a little trouble to avoid \cw{gdef}-ing \arg{2}, in order
%    to prevent save stack buildup if the user of \cw{readline} carries
%    on unaware doing local redefinitions of \arg{2} after the initial
%    read.
%    \begin{macrocode}
  \read\m@ne to#2%
  \edef#2{\def\nx@#2{#2}}%
  \xp@\endgroup #2%
  \ifx\@empty#2\def#2{#1}\fi
}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\xreadline}
%
%    \cw{xreadline} is like \cw{readline} except that it leaves almost
%    all catcodes unchanged so that the return value is executable
%    tokens instead of strictly character tokens of category 11 or 12.
%    \begin{macrocode}
\def\xreadline#1#2{%
  \begingroup
%    \end{macrocode}
%    Render some outer control sequences innocuous.
%    \begin{macrocode}
    \xp@\let\csname bye\endcsname\relax
    \xp@\let\csname newif\endcsname\relax
    \xp@\let\csname newcount\endcsname\relax
    \xp@\let\csname newdimen\endcsname\relax
    \xp@\let\csname newskip\endcsname\relax
    \xp@\let\csname newmuskip\endcsname\relax
    \xp@\let\csname newtoks\endcsname\relax
    \xp@\let\csname newbox\endcsname\relax
    \xp@\let\csname newinsert\endcsname\relax
    \xp@\let\csname +\endcsname\relax
    \actively\let\^^L\relax
  \catcode`\^^M=9 % ignore
  \endlinechar`\^^M% reset to normal
  \read\m@ne to#2%
  \toks@\xp@{#2}%
  \edef\@tempa{\def\nx@#2{\the\toks@}}%
  \xp@\endgroup \@tempa
  \ifx\@empty#2\def#2{#1}\fi
}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\readchar}
%
%    \cw{readchar} reduces the user response to a single character.
%    \begin{macrocode}
\def\readchar#1#2{%
  \readline{#1}#2%
%    \end{macrocode}
%    If the user's response and the default response are both empty,
%    we need something after \arg{1} to keep \cw{@car} from running
%    away, so we add an empty pair of braces.
%    \begin{macrocode}
  \edef#2{\xp@\@car#2#1{}\@nil}%
}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\readChar}
%
%    \cw{readChar} reduces the user response to a single uppercase
%    character. (This is useful to simplify testing the response
%    later with \cw{if}.)
%
%    \begin{macrocode}
\def\readChar#1#2{%
  \readline{#1}#2%
  \changecase\uppercase#2%
%    \end{macrocode}
%    Reduce \arg{2} to its first character, or the first character of
%    \arg{1}, if \arg{2} is empty. The extra braces \verb'{}' are to
%    prevent a runaway argument error from \cw{@car} if \arg{2} and
%    \arg{1} are both empty.
%    \begin{macrocode}
  \edef#2{\xp@\@car #2#1{}\@nil}%
}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\changecase}
%    The function \cw{changecase} uppercases or lowercases the
%    replacement text of its second argument, which
%    must be a macro. The first argument should be \cw{uppercase} or
%    \cw{lowercase}.
%    \begin{macrocode}
\def\changecase#1#2{\@casetoks\xp@{#2}%
  \edef#2{#1{\def\nx@#2{\the\@casetoks}}}#2}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\@casetoks}
%    We allocate a token register just for the use of \cw{changecase}
%    because it might be used at a low level internally where we don't
%    want to interfere with other uses of the scratch token registers
%    0\dash 9.
%    \begin{macrocode}
\newtoks\@casetoks
%    \end{macrocode}
% \end{macro}
%
%    A common task in reading user input is to verify, when an answer
%    of a certain kind was requested, that the response has indeed the
%    desired form\Dash for example, if a nonnegative integer is required
%    for subsequent processing, it behooves us to verify that we have
%    a nonnegative integer in hand before doing anything that might
%    lead to inconvenient error messages. However, it's not easy to
%    decide how best to handle such verification. One possibility
%    might be to have a function
%
%^^V \readnonnegativeinteger\foo
%
%    to do all the work of going out and fetching a number from the
%    user and leaving it in the macro \cw{foo}. Another possibility
%    would be to read the response using \cw{readline} and then apply
%    a separate function that can be used in combination with \cw{if},
%    for example
%
%^^V   \readline{}\reply
%^^V   \if\validnumber\reply ... \else ... \fi
%
%    For maximum flexibility, a slightly lower-level approach is
%    chosen here. The target syntax is
%
%^^V   \readline{}\reply
%^^V   \checkinteger\reply\tempcount
%
%    where \cw{tempcount} will be set to \verb'-'\cw{maxdimen} if
%    \cw{reply} does {\em not\/} contain a valid integer. (Negative
%    integers are allowed, as long as they are greater than
%    \verb'-'\cw{maxdimen}.)  Then the function that calls
%    \cw{checkinteger} is free to make additional checks on the range
%    of the reply and give error messages tailored to the
%    circumstances. And the handling of an empty \cw{reply} can be
%    arbitrarily customized, something that would tend to be
%    inconvenient for the first method mentioned.
%
%    The first and second approaches can be built on top of the third
%    if desired, e.g.\ (for the second approach)
%
%^^V \def\validnumber#1{TT\fi
%^^V   \checkinteger#1\tempcount%
%^^V   \ifnum\tempcount>-\maxdimen }
%
%    The curious \verb'TT\fi'\verbdots\cw{ifnum} construction is from
%    \TeXhax{} 1989, no.~20 and no.~38 (a suggestion of D. E. Knuth in
%    reply to a query by S. von Bechtolsheim).
%
% \begin{macro}{\checkinteger}
%
%    The arguments of \cw{checkinteger}'s are: \arg{2}, a count register
%    to hold the result; \arg{1}, a macro holding zero or more arbitrary
%    characters of category 11 or 12.
%
%    \begin{macrocode}
\def\checkinteger#1#2{\let\scansign@\@empty
  \def\scanresult@{#2}%
  \xp@\scanint#1x\endscan}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\scanint}
%
%    To validate a number, the function \cw{scanint} must first scan
%    away leading \qc{\+} or \qc{\-} signs (keeping track in
%    \cw{scansign@}), then look at the first token after that: if it's
%    a digit, fine, scan that digit and any succeeding digits into the
%    given count register (\cw{scanresult@}), ending with \cw{endscan}
%    to get rid of any following garbage tokens that might just
%    possibly show up.
%
%    Typical usage of \cw{scanint} includes initializing \cw{scansign@}
%    to empty, as in the definition of \cw{checkinteger}.
%\begin{usage}
%\let\scansign@\@empty
%\def\scanresult@{\tempcount}%
%\xp@\scanint\reply x\endscan
%\end{usage}
%    Assumption: \cw{reply} is either empty or contains only category
%    11 or 12 characters (which it will if you used \cw{readline}!). If
%    a separate check is done earlier to trap the case where \cw{reply}
%    is empty\Dash for example, by using a nonempty default for
%    \cw{readline}\Dash then the \verb'x' before \cw{endscan} is
%    superfluous.
%
%    Arg \arg{1} = next character from the string being tested.
%    The test whether \arg{1} is a decimal digit is similar in spirit
%    to the test \verb'\if!#1!' to see if an argument is empty
%    (\TB, Appendix~D, p.~376).
%    \begin{macrocode}
\def\scanint#1{%
  \ifodd 0#11 %
%    \end{macrocode}
%    Is \arg{1} a decimal digit? If so read all digits into
%    \cw{scanresult@} with the sign prefix.
%    \begin{macrocode}
    \def\@tempa{\afterassignment\endscan
      \scanresult@=\scansign@#1}%
  \else
    \if -#1\relax
%    \end{macrocode}
%    Here we flipflop the sign; watch closely.
%    \begin{macrocode}
      \edef\scansign@{%
        \ifx\@empty\scansign@ -\fi}%
      \def\@tempa{\scanint}%
    \else
%    \end{macrocode}
%    A plus sign can just be ignored.
%    \begin{macrocode}
      \if +#1\relax
        \def\@tempa{\scanint}%
      \else % not a valid number
        \def\@tempa{%
          \scanresult@=-\maxdimen\endscan}%
  \fi\fi\fi
  \@tempa
}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\endscan}
%    The \cw{endscan} function just gobbles any remaining garbage. It
%    uses its own name as the argument delimiter.
%    \begin{macrocode}
\def\endscan#1\endscan{}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\dimenfirstpart}
% \begin{macro}{\dimentoks}
%    \cw{dimenfirstpart}, a count register, receives the digits, if
%    any, preceding the decimal point. \cw{dimentoks}, a token
%    register, receives any digits after the decimal point.
%    \begin{macrocode}
\newcount\dimenfirstpart
\newtoks\dimentoks
%    \end{macrocode}
% \end{macro}
% \end{macro}
%
% \begin{macro}{\scandimen}
%
%    \cw{scandimen} is similar to \cw{scanint} but has to call some
%    auxiliary functions to scan the various subcomponents of a
%    dimension (leading digits, decimal point, fractional part, and
%    units, with optional {\tt true}, in addition to the sign). The
%    minimum requirements of \tex/'s syntax for dimensions are a digit
%    or decimal point \qc{\+} the units; all the other components are
%    optional (\TB, Exercise~10.3, p.~58).
%
%    When scanning for the digits of a fractional part, we can't throw
%    away leading zeros; therefore we don't read the fractional part
%    into a count register as we did for the digits before the decimal
%    point; instead we read the digits one by one and store them in
%    \cw{dimentoks}.
%
%    The function that calls \cw{scandimen} should initialize
%    \cw{scansign@} to \cw{@empty}, \cw{dimenfirstpart} to \cw{z@},
%    \cw{dimentoks} to empty \verb'{}', and \cw{dimentrue@} to
%    \cw{@empty}.
%
%    Test values: {\tt 0pt}, {\tt 1.1in}, {\tt -2cm}, {\tt .3mm}, {\tt
%    0.4dd}, {\tt 5.cc}, {\tt.10000000009pc}, \cw{hsize}, {\tt em}.
%
%    \begin{macrocode}
\def\scandimen#1{%
  \ifodd 0#11
    \def\@tempa{\def\@tempa{\scandimenb}%
      \afterassignment\@tempa
      \dimenfirstpart#1}%
  \else
%    \end{macrocode}
%
%    The following test resolves to true if \arg{1} is either a period
%    or a comma (both recognized by \tex/ as decimal point
%    characters).
%
%    \begin{macrocode}
    \if \if,#1.\else#1\fi.%
      \def\@tempa{\scandimenc}%
    \else
      \if -#1% then flipflop the sign
        \edef\scansign@{%
          \ifx\@empty\scansign@ -\fi}%
        \def\@tempa{\scandimen}%
      \else
        \if +#1% then ignore it
          \def\@tempa{\scandimen}%
        \else % not a valid dimen
          \def\@tempa{%
            \scanresult@=-\maxdimen\endscan}%
  \fi\fi\fi\fi
  \@tempa
}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\scandimenb}
%
%    Scan for an optional decimal point.
%
%    \begin{macrocode}
\def\scandimenb#1{%
  \if \if,#1.\else#1\fi.%
    \def\@tempa{\scandimenc}%
  \else
%    \end{macrocode}
%    If the decimal point is absent, we need to put back \arg{2} and
%    rescan it to see if it is the first letter of the units.
%    \begin{macrocode}
    \def\@tempa{\scanunitsa#1}%
  \fi
  \@tempa
}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\scandimenc}
%
%    Scan for the fractional part: digits after the decimal point.
%
%    \begin{macrocode}
\def\scandimenc#1{%
%    \end{macrocode}
%    If \arg{1} is a digit, add it to \cw{dimentoks}.
%    \begin{macrocode}
  \ifodd 0#11 \dimentoks\xp@{%
      \the\dimentoks#1}%
    \def\@tempa{\scandimenc}%
  \else
%    \end{macrocode}
%    Otherwise rescan \arg{1}, presumably the first letter of the
%    units.
%    \begin{macrocode}
  \def\@tempa{\scanunitsa#1}%
  \fi
  \@tempa
}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\scanunitsa}
%
%    \begin{macrocode}
\def\scanunitsa#1\endscan{%
%    \end{macrocode}
%    Check for \verb'true' qualifier.
%    \begin{macrocode}
  \def\@tempa##1true##2##3\@tempa{##2}%
%    \end{macrocode}
%
%    The peculiar nature of \cw{lowercase} is evident here as we are
%    able to apply it to only the test part of the conditional without
%    running into brace-matching problems. (Compare the braces in this
%    example to something like \cw{message}\qc{\{}\cw{iffalse}
%    \verb"A}"\cw{else} \verb"B}"\cw{fi}.)
%
%    \begin{macrocode}
  \lowercase{%
    \xp@\ifx\xp@\end
    \@tempa#1true\end\@tempa
  }%
%    \end{macrocode}
%    No \verb'true' was found:
%    \begin{macrocode}
    \let\dimentrue@\@empty
    \def\@tempa{\scanunitsb#1\endscan}%
  \else
    \def\dimentrue@{true}%
    \def\@tempa##1true##2\@tempa{%
      \def\@tempa{##1}%
      \ifx\@tempa\@empty
        \def\@tempa{\scanunitsb##2\endscan}%
      \else
        \def\@tempa{\scanunitsb xx\endscan}%
      \fi}%
    \@tempa#1\@tempa
  \fi
  \@tempa
}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\scanunitsb}
%
%    Scan for the name of the units and complete the assignment of the
%    scanned value to \cw{scanresult@}. Notice that, because of the way
%    \cw{scanunitsb} picks up \arg{1} and \arg{2} as macro arguments,
%    \verb"p t" is allowed as a variation of \verb"pt". Eliminating
%    this permissiveness doesn't seem worth the speed penalty
%    that would be incurred in \cw{scanunitsb}.
%
%    The method for detecting a valid units string is to define the
%    scratch function \cw{@tempa} to apply \tex/'s parameter-matching
%    abilities to a special string that will yield a boolean value
%    of true if and only if the given string is a valid \tex/ unit.
%    \begin{macrocode}
\def\scanunitsb#1#2{%
  \def\@tempa##1#1#2##2##3\@nil{##2}%
  \def\@tempb##1{T\@tempa
    pcTptTcmTccTemTexTinTmmTddTspT##1F\@nil}%
%    \end{macrocode}
%    Force lowercase just in case the units were entered with
%    uppercase letters (accepted by \tex/, so we had better accept
%    uppercase also).
%    \begin{macrocode}
  \lowercase{%
  \if\@tempb{#1#2}%
  }%
   \scanresult@=\scansign@
     \number\dimenfirstpart.\the\dimentoks
     \dimentrue@#1#2\relax
  \else
    \scanresult@=-\maxdimen
  \fi
%    \end{macrocode}
%    Call \cw{endscan} to gobble garbage tokens, if any.
%    \begin{macrocode}
  \endscan
}
%    \end{macrocode}
% \end{macro}
%
% \begin{macro}{\checkdimen}
%
%    Argument \arg{2} must be a dimen register; \arg{1} is expected to
%    be a macro holding zero or more arbitrary characters of category 11
%    or 12.
%    \begin{macrocode}
\def\checkdimen#1#2{%
  \let\scansign@\@empty \def\scanresult@{#2}%
  \let\dimentrue@\@empty
  \dimenfirstpart\z@ \dimentoks{}%
  \xp@\scandimen#1xx\endscan
}
%    \end{macrocode}
% \end{macro}
%
%    Finish up.
%    \begin{macrocode}
\restorecatcodes
\endinput
%    \end{macrocode}
%
% \CheckSum{858}
% \Finale