771
|
1 : #-*- Perl -*-
|
|
2
|
|
3 ### make-mswin-unicode --- generate Unicode-encapsulation code for MS Windows
|
|
4
|
2367
|
5 ## Copyright (C) 2001, 2002, 2004 Ben Wing.
|
771
|
6
|
|
7 ## Author: Ben Wing <ben@xemacs.org>
|
|
8 ## Maintainer: Ben Wing <ben@xemacs.org>
|
|
9 ## Current Version: 1.0, August 24, 2001
|
|
10
|
|
11 ## This file is part of XEmacs.
|
|
12
|
|
13 ## XEmacs is free software; you can redistribute it and/or modify it
|
|
14 ## under the terms of the GNU General Public License as published by
|
|
15 ## the Free Software Foundation; either version 2, or (at your option)
|
|
16 ## any later version.
|
|
17
|
|
18 ## XEmacs is distributed in the hope that it will be useful, but
|
|
19 ## WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
20 ## MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
|
|
21 ## General Public License for more details.
|
|
22
|
|
23 ## You should have received a copy of the GNU General Public License
|
|
24 ## along with XEmacs; see the file COPYING. If not, write to the Free
|
|
25 ## Software Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA
|
|
26 ## 02111-1307, USA.
|
|
27
|
|
28 eval 'exec perl -w -S $0 ${1+"$@"}'
|
|
29 if 0;
|
|
30
|
|
31 use strict;
|
|
32 use File::Basename;
|
|
33 use Getopt::Long;
|
|
34
|
|
35 my ($myName, $myPath) = fileparse ($0);
|
|
36
|
|
37 my $usage="
|
|
38 Usage: $myName [--c-output FILE] [--h-output FILE] [--help] [FILES ...]
|
|
39
|
|
40 The purpose of this script is to auto-generate Unicode-encapsulation
|
|
41 code for MS Windows library functions that come in two versions (ANSI
|
|
42 and Unicode). The MS Windows header files provide a way of
|
|
43 automatically calling the right version, but only at compile-time,
|
|
44 which is *NOT* sufficient for any real-world program. The solution is
|
|
45 run-time Unicode encapsulation, which is not conceptually difficult
|
|
46 but is time-consuming, and is not supported standardly only due to
|
|
47 evil marketing decisions made by Microsoft. See src/intl-win32.c
|
|
48 for more information.
|
|
49
|
800
|
50 In XEmacs, this file is normally run using `nmake -f xemacs.mak
|
|
51 unicode-encapsulate'.
|
|
52
|
771
|
53 This script processes the specified files, looking for commands
|
|
54 indicating library routines to Unicode-encapsulate, as follows:
|
|
55
|
|
56 Portions of the files that should be processed are enclosed in lines
|
|
57 consisting only of the words \"begin-unicode-encapsulation-script\"
|
|
58 and \"end-unicode-encapsulation-script\". More than one section can
|
|
59 occur in a single file. Processed lines begin with a command word,
|
|
60 followed by one or more args (no quotes are necessary for spaces):
|
|
61
|
|
62 file specifies a file to start reading from.
|
|
63 yes indicates a function to be automatically Unicode-encapsulated.
|
|
64 (All parameters either need no special processing or are LPTSTR or
|
|
65 LPCTSTR.)
|
|
66 soon indicates a function that should be automatically Unicode-encapsulated,
|
|
67 but we're not ready to process it yet.
|
|
68 no indicates a function we don't support (it will be #defined to cause
|
|
69 a compile error, with the text after the function included in the
|
|
70 erroneous definition to indicate why we don't support it).
|
|
71 skip indicates a function we support manually; only a comment about this
|
|
72 will be generated.
|
|
73 split indicates a function with a split structure (different versions
|
|
74 for Unicode and ANSI), but where the only difference is in pointer
|
|
75 types, and the actual size does not differ. The structure name
|
|
76 should follow the function name, and it will be automatically
|
|
77 Unicode-encapsulated with appropriate casts.
|
|
78 begin-bracket indicates a #if statement to be inserted here.
|
|
79 end-bracket indicates the corresponding #endif statement.
|
|
80 blank lines and lines beginning with // are ignored.
|
|
81 ";
|
|
82
|
|
83 # ------------------ process command-line options ------------------
|
|
84
|
|
85 my %options;
|
|
86 my @SAVE_ARGV = @ARGV;
|
|
87
|
|
88 $Getopt::Long::ignorecase = 0;
|
|
89 &GetOptions (
|
|
90 \%options,
|
|
91 'c-output=s',
|
|
92 'h-output=s',
|
778
|
93 'includedir=s',
|
771
|
94 'help',
|
|
95 );
|
|
96
|
|
97 die $usage if $options{"help"};
|
|
98
|
|
99 my $in_script;
|
|
100 my $slurp;
|
|
101
|
778
|
102 my ($cout, $hout, $dir) = ($options{"c-output"},
|
|
103 $options{"h-output"},
|
|
104 $options{"includedir"});
|
|
105 if (!$dir)
|
|
106 {
|
|
107 $dir=$ENV{"MSVCDIR"} or die "Environment variable MSVCDIR undefined - run vcvars32.bat from your MSVC installation";
|
|
108 $dir.='/include';
|
|
109 }
|
|
110 die "Can't find MSVC include files in \"$dir\"" unless (-f $dir.'/WINDOWS.H');
|
771
|
111
|
|
112 open (COUT, ">$cout") or die "Can't open C output file $cout: $!";
|
|
113 open (HOUT, ">$hout") or die "Can't open C output file $hout: $!";
|
|
114
|
|
115 select (STDOUT); $| = 1;
|
|
116
|
|
117 print COUT "/* Automatically-generated Unicode-encapsulation file,
|
|
118 using the command
|
|
119
|
|
120 $myPath$myName @SAVE_ARGV
|
|
121
|
|
122 Do not edit. See `$myName'.
|
|
123 */
|
|
124
|
|
125 #include <config.h>
|
|
126 #include \"lisp.h\"
|
|
127
|
|
128 #include \"syswindows.h\"
|
|
129
|
|
130 ";
|
|
131 print HOUT "/* Automatically-generated Unicode-encapsulation header file.
|
|
132 Do not edit. See `$myName'.
|
|
133 */\n\n";
|
|
134
|
|
135 my %files;
|
|
136 my %processed;
|
|
137 my %bracket;
|
|
138
|
|
139 my $current_file;
|
|
140 my @current_bracket;
|
|
141
|
|
142 while (<>)
|
|
143 {
|
|
144 chomp;
|
800
|
145 # remove trailing CR. #### Should not be necessary! Perl should be
|
|
146 # opening these in text mode by default, as the docs claim, and
|
|
147 # automatically remove the CR's.
|
|
148 tr/\r//d;
|
771
|
149
|
|
150 if (/^begin-unicode-encapsulation-script$/)
|
|
151 {
|
|
152 $in_script = 1;
|
|
153 }
|
|
154 elsif (/^end-unicode-encapsulation-script$/)
|
|
155 {
|
|
156 $in_script = 0;
|
|
157 }
|
|
158 elsif ($in_script)
|
|
159 {
|
|
160 next if (m!^//!);
|
|
161 next if (/^[ \t]*$/);
|
778
|
162 if (/(file|yes|soon|no|skip|split|begin-bracket|end-bracket)(?: (.*))?/)
|
771
|
163 {
|
|
164 my ($command, $parms) = ($1, $2);
|
778
|
165 if ($command eq "file")
|
771
|
166 {
|
|
167 $current_file = $parms;
|
|
168 }
|
|
169 elsif ($command eq "begin-bracket")
|
|
170 {
|
|
171 my $current_bracket = $current_bracket[$#current_bracket];
|
|
172 if (defined ($current_bracket))
|
|
173 {
|
|
174 $current_bracket .= "&& $parms";
|
|
175 }
|
|
176 else
|
|
177 {
|
|
178 $current_bracket = "$parms";
|
|
179 }
|
|
180 push @current_bracket, $current_bracket;
|
|
181 }
|
|
182 elsif ($command eq "end-bracket")
|
|
183 {
|
|
184 pop @current_bracket;
|
|
185 }
|
|
186 else
|
|
187 {
|
|
188 my ($fun, $reason) = split /\s+/, $parms, 2;
|
|
189 $files{$current_file}{$fun} = [$command, $reason];
|
|
190 $bracket{$current_file}{$fun} =
|
|
191 $current_bracket[$#current_bracket];
|
|
192 }
|
|
193 }
|
|
194 else
|
|
195 {
|
|
196 print "WARNING: Unknown line $_\n";
|
|
197 }
|
|
198 }
|
|
199 }
|
|
200
|
|
201
|
|
202 foreach my $file (keys %files)
|
|
203 {
|
|
204 $slurp = &FileContents ($file);
|
|
205 print "Processing file $file\n";
|
|
206 print HOUT "\n/* Processing file $file */\n\n";
|
|
207 my $totalspace = 70 - length ("Processing file $file");
|
|
208 $totalspace = 0 if $totalspace < 0;
|
|
209 my $alignspaceleft = $totalspace / 2;
|
|
210 my $alignspaceright = ($totalspace + 1) / 2;
|
|
211 print COUT "
|
|
212 /*----------------------------------------------------------------------*/
|
|
213 /*" . (" " x $alignspaceleft) . "Processing file $file" .
|
|
214 (" " x $alignspaceright) . "*/
|
|
215 /*----------------------------------------------------------------------*/
|
|
216
|
|
217 ";
|
|
218
|
|
219 my ($ws_re, $must_ws_re, $tok_ch) =
|
|
220 ("\\s*", "\\s+", "\\w");
|
|
221 # unfortunately there is no surefire way short of
|
|
222 # parsing all include files for typedefs to
|
|
223 # distinguish types from parameters, and prototypes
|
|
224 # appear in the include files both with and without
|
|
225 # parameters -- the latter kinds appear in a very
|
|
226 # different style and were obviously added later. so
|
|
227 # we rely on the fact that defined types are all
|
|
228 # upper-case, and parameters generally are not, and
|
|
229 # special-case the exceptions.
|
|
230 my $typeword_re =
|
|
231 # note the negative lookahead assertions: the first
|
|
232 # one excludes the words "X" and "Y" from type
|
|
233 # words, since they appear as parameter names in
|
|
234 # CreateWindowEx; the second prevents "void
|
|
235 # *Argument" from being parsed as a type "void *A"
|
|
236 # followed by a parameter "rgument".
|
|
237 "(?:(?!(?:X\\b|Y\\b))(?:unsigned|int|long|short|va_list|[A-Z_0-9]+)(?!${tok_ch}))";
|
|
238 my $typetoken_re = "(?:$typeword_re$ws_re\\**$ws_re)";
|
|
239 my $arg_re = "(?:($typetoken_re+)(${tok_ch}+)?(?: OPTIONAL)?)";
|
|
240 my $fun_re = "(SHSTDAPI_\\(${tok_ch}+\\)|${tok_ch}" . "[A-Za-z_0-9 \t\n\r\f]*?${tok_ch})${ws_re}(${tok_ch}+)W${ws_re}\\(((${ws_re}${arg_re}${ws_re},)*${ws_re}${arg_re}${ws_re})\\);";
|
|
241
|
|
242 # print "regexp: $fun_re\n";
|
|
243 while ($slurp =~ /$fun_re/g)
|
|
244 {
|
|
245 my ($rettype, $fun, $args) = ($1, $2, $3);
|
|
246 $processed{$fun} = 1;
|
|
247 print "Processing: $fun";
|
|
248
|
|
249 my ($command, $reason) = ($files{$file}{$fun}[0], $files{$file}{$fun}[1]);
|
|
250 if (!defined ($command))
|
|
251 {
|
|
252 print " (no command found)\n";
|
|
253 }
|
|
254 else
|
|
255 {
|
|
256 print "\n";
|
|
257 my $bracket = $bracket{$file}{$fun};
|
|
258 if (defined ($bracket))
|
|
259 {
|
|
260 print HOUT "#if $bracket\n";
|
|
261 print COUT "#if $bracket\n\n";
|
|
262 }
|
|
263 if ($command eq "no")
|
|
264 {
|
|
265 if (!defined ($reason))
|
|
266 {
|
|
267 print "WARNING: No reason given for `no' with function $fun\n";
|
|
268 $reason = "";
|
|
269 }
|
|
270
|
|
271 print HOUT "#undef $fun\n";
|
2367
|
272 (my $munged_reason = $reason) =~ s/[^A-Za-z0-9]/_/g;
|
|
273 print HOUT "#define $fun error_$munged_reason\n";
|
771
|
274 print COUT "/* Error if $fun used: $reason */\n\n";
|
|
275 }
|
|
276 elsif ($command eq "skip")
|
|
277 {
|
|
278 if (!defined ($reason))
|
|
279 {
|
|
280 print "WARNING: No reason given for `skip' with function $fun\n";
|
|
281 $reason = "";
|
|
282 }
|
|
283
|
|
284 print HOUT "/* Skipping $fun because $reason */\n";
|
|
285 print COUT "/* Skipping $fun because $reason */\n\n";
|
|
286 }
|
|
287 elsif ($command eq "soon")
|
|
288 {
|
|
289 $reason = "" if !defined ($reason);
|
|
290
|
|
291 print HOUT "/* Not yet: $fun $reason */\n";
|
|
292 print COUT "/* Not yet: $fun $reason */\n\n";
|
|
293 }
|
|
294 else
|
|
295 {
|
|
296 my (@args, %argtype, %ansiarg, %xarg, $split_struct,
|
|
297 $split_rettype);
|
|
298 if ($command eq "split")
|
|
299 {
|
|
300 ($split_struct, $reason) = split /\s+/, $reason, 2;
|
|
301 }
|
|
302 my $argno = 0;
|
|
303 while ($args =~ /$arg_re/g)
|
|
304 {
|
|
305 $argno++;
|
|
306 my ($argtype, $argname) = ($1, $2);
|
|
307 $argtype =~ s/\s*$//;
|
|
308 next if $argtype eq "void" || $argtype eq "VOID";
|
|
309 $argname = "arg$argno" if !defined ($argname);
|
|
310 $argtype{$argname} = $argtype;
|
|
311 $ansiarg{$argname} = $argtype;
|
|
312 $ansiarg{$argname} =~ s/\bLPWSTR\b/LPSTR/;
|
|
313 $ansiarg{$argname} =~ s/\bLPCWSTR\b/LPCSTR/;
|
|
314 $xarg{$argname} = $argtype;
|
|
315 $xarg{$argname} =~ s/\bLPWSTR\b/Extbyte */;
|
|
316 $xarg{$argname} =~ s/\bLPCWSTR\b/const Extbyte */;
|
|
317 if (defined ($split_struct))
|
|
318 {
|
|
319 my $fuck_cperl1 = "\\b${split_struct}W\\b";
|
|
320 my $fuck_cperl2 = "${split_struct}A";
|
|
321 $ansiarg{$argname} =~ s/$fuck_cperl1/$fuck_cperl2/;
|
|
322 }
|
|
323 push @args, $argname;
|
|
324 }
|
|
325 $rettype =~ s/\bSHSTDAPI_\((.*)\)/$1/;
|
|
326 $rettype =~ s/\s*WIN\w*?API\s*//g;
|
|
327 $rettype =~ s/\bAPIENTRY\b//;
|
|
328 $rettype =~ s/\bSHSTDAPI\b/HRESULT/;
|
|
329 if ($rettype =~ /LPC?WSTR/)
|
|
330 {
|
|
331 $split_rettype = 1;
|
|
332 $rettype =~ s/\bLPWSTR\b/Extbyte */;
|
|
333 $rettype =~ s/\bLPCWSTR\b/const Extbyte */;
|
|
334 }
|
800
|
335 print HOUT "#ifdef ERROR_WHEN_NONINTERCEPTED_FUNS_USED\n";
|
|
336 print HOUT "#undef $fun\n";
|
2367
|
337 print HOUT "#define $fun error_use_qxe${fun}_or_${fun}A_and_${fun}W\n";
|
800
|
338 print HOUT "#endif\n";
|
771
|
339 if (defined ($reason))
|
|
340 {
|
|
341 print COUT "/* NOTE: $reason */\n";
|
|
342 }
|
|
343 print COUT "$rettype\nqxe$fun (";
|
|
344 print HOUT "$rettype qxe$fun (";
|
|
345 my $first = 1;
|
|
346 if (!@args)
|
|
347 {
|
|
348 print COUT "void";
|
|
349 print HOUT "void";
|
|
350 }
|
|
351 else
|
|
352 {
|
|
353 foreach my $x (@args)
|
|
354 {
|
|
355 print COUT ", " if !$first;
|
|
356 print HOUT ", " if !$first;
|
|
357 $first = 0;
|
|
358 print COUT "$xarg{$x} $x";
|
|
359 print HOUT "$xarg{$x} $x";
|
|
360 }
|
|
361 }
|
|
362 print HOUT ");\n";
|
|
363 print COUT ")\n{\n if (XEUNICODE_P)\n ";
|
|
364 if ($rettype ne "void" && $rettype ne "VOID")
|
|
365 {
|
|
366 print COUT "return ";
|
|
367 print COUT "($rettype) " if $split_rettype;
|
|
368 }
|
|
369 print COUT "${fun}W (";
|
|
370 $first = 1;
|
|
371 foreach my $x (@args)
|
|
372 {
|
|
373 print COUT ", " if !$first;
|
|
374 $first = 0;
|
|
375 print COUT ($argtype{$x} eq $xarg{$x} ? $x :
|
|
376 "($argtype{$x}) $x");
|
|
377 }
|
|
378 print COUT ");\n else\n ";
|
|
379 if ($rettype ne "void" && $rettype ne "VOID")
|
|
380 {
|
|
381 print COUT "return ";
|
|
382 print COUT "($rettype) " if $split_rettype;
|
|
383 }
|
|
384 print COUT "${fun}A (";
|
|
385 $first = 1;
|
|
386 foreach my $x (@args)
|
|
387 {
|
|
388 print COUT ", " if !$first;
|
|
389 $first = 0;
|
|
390 print COUT ($argtype{$x} eq $ansiarg{$x} ? $x :
|
|
391 "($ansiarg{$x}) $x");
|
|
392 }
|
|
393 print COUT ");\n}\n\n";
|
|
394 }
|
|
395 if (defined ($bracket))
|
|
396 {
|
|
397 print HOUT "#endif /* $bracket */\n";
|
|
398 print COUT "#endif /* $bracket */\n\n";
|
|
399 }
|
800
|
400 print HOUT "\n";
|
771
|
401 }
|
|
402 }
|
|
403 }
|
|
404
|
|
405 foreach my $file (keys %files)
|
|
406 {
|
|
407 foreach my $fun (keys %{$files{$file}})
|
|
408 {
|
|
409 if (!$processed{$fun} && $files{$file}{$fun}[0] =~ /^(yes|soon|split)$/)
|
|
410 {
|
|
411 print "WARNING: Can't locate prototype for $fun\n";
|
|
412 }
|
|
413 }
|
|
414 }
|
|
415
|
|
416
|
|
417 sub FileContents
|
|
418 {
|
|
419 local $/ = undef;
|
778
|
420 open (FILE, "< $dir/$_[0]") or die "$dir/$_[0]: $!";
|
771
|
421 my $retval = scalar <FILE>;
|
|
422 # must hack away CRLF junk.
|
|
423 $retval =~ s/\r\n/\n/g;
|
|
424 return $retval;
|
|
425 }
|