gcc/ABOUT-GCC-NLS

   1 Notes on GCC's Native Language Support
   2
   3 GCC's Native Language Support (NLS) is relatively new and
   4 experimental, so NLS is currently disabled by default.  Use
   5 configure's --enable-nls option to enable it.  Eventually, NLS will be
   6 enabled by default, and you'll need --disable-nls to disable it.  You
   7 must enable NLS in order to make a GCC distribution.
   8
   9 By and large, only diagnostic messages have been internationalized.
  10 Some work remains in other areas; for example, GCC does not yet allow
  11 non-ASCII letters in identifiers.
  12
  13 Not all of GCC's diagnostic messages have been internationalized.
  14 Programs like `enquire' and `genattr' are not internationalized, as
  15 their users are GCC maintainers who typically need to be able to read
  16 English anyway; internationalizing them would thus entail needless
  17 work for the human translators.  And no one has yet gotten around to
  18 internationalizing the messages in the C++ compiler, or in the
  19 specialized MIPS-specific programs mips-tdump and mips-tfile.
  20
  21 The GCC library should not contain any messages that need
  22 internationalization, because it operates below the
  23 internationalization library.
  24
  25 Currently, the only language translation supplied is en_UK (British English).
  26
  27 Unlike some other GNU programs, the GCC sources contain few instances
  28 of explicit translation calls like _("string").  Instead, the
  29 diagnostic printing routines automatically translate their arguments.
  30 For example, GCC source code should not contain calls like `error
  31 (_("unterminated comment"))'; it should contain calls like `error
  32 ("unterminated comment")' instead, as it is the `error' function's
  33 responsibility to translate the message before the user sees it.
  34
  35 By convention, any function parameter in the GCC sources whose name
  36 ends in `msgid' is expected to be a message requiring translation.
  37 For example, the `error' function's first parameter is named `msgid'.
  38 GCC's exgettext script uses this convention to determine which
  39 function parameter strings need to be translated.  The exgettext
  40 script also assumes that any occurrence of `%eMSGID}' on a source
  41 line, where MSGID does not contain `%' or `}', corresponds to a
  42 message MSGID that requires translation; this is needed to identify
  43 diagnostics in GCC spec strings.
  44
  45 If you enable NLS and modify source files, you'll need to use a
  46 special version of the GNU gettext package to propagate the
  47 modifications to the translation tables.  Apply the following patch
  48 (use `patch -p0') to GNU gettext 0.10.35, which you can retrieve from:
  49
  50 ftp://alpha.gnu.org/gnu/gettext-0.10.35.tar.gz
  51
  52 This patch has been submitted to the GNU gettext maintainer, so
  53 eventually we shouldn't need this special gettext version.
  54
  55 This patch is free software; you can redistribute it and/or modify
  56 it under the terms of the GNU General Public License as published by
  57 the Free Software Foundation; either version 2, or (at your option)
  58 any later version.
  59
  60 This patch is distributed in the hope that it will be useful,
  61 but WITHOUT ANY WARRANTY; without even the implied warranty of
  62 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  63 GNU General Public License for more details.
  64
  65 You should have received a copy of the GNU General Public License
  66 along with this patch; see the file COPYING.  If not, write to
  67 the Free Software Foundation, 59 Temple Place - Suite 330,
  68 Boston, MA 02111-1307, USA.
  69
  70 1998-07-26  Paul Eggert  <eggert@twinsun.com>
  71
  72         * po/Makefile.in.in (maintainer-clean): Remove cat-id-tbl.c and
  73         stamp-cat-id.
  74
  75 1998-07-24  Paul Eggert  <eggert@twinsun.com>
  76
  77         * po/Makefile.in.in (cat-id-tbl.o): Depend on
  78         $(top_srcdir)/intl/libgettext.h, not ../intl/libgettext.h.
  79
  80 1998-07-20  Paul Eggert  <eggert@twinsun.com>
  81
  82         * po/Makefile.in.in (.po.pox, all-yes, $(srcdir)/cat-id-tbl.c,
  83         $(srcdir)/stamp-cat-id, update-po): Prepend `$(srcdir)/' to
  84         files built in the source directory; this is needed for
  85         VPATH-based make in Solaris 2.6.
  86
  87 1998-07-17  Paul Eggert  <eggert@twinsun.com>
  88
  89         Add support for user-specified argument numbers for keywords.
  90         Extract all strings from a keyword arg, not just the first one.
  91         Handle parenthesized commas inside keyword args correctly.
  92         Warn about nested keywords.
  93
  94         * doc/gettext.texi: Document --keyword=id:argnum.
  95
  96         * src/xgettext.c (scan_c_file):
  97         Warn about nested keywords, e.g. _(_("xxx")).
  98         Warn also about not-yet-implemented but allowed nesting, e.g.
  99         dcgettext(..._("xxx")..., "yyy").
 100         Get all strings in a keyword arg, not just the first one.
 101         Handle parenthesized commas inside keyword args correctly.
 102
 103         * src/xget-lex.h (enum xgettext_token_type_ty):
 104         Replace xgettext_token_type_keyword1 and
 105         xgettext_token_type_keyword2 with just plain
 106         xgettext_token_type_keyword; it now has argnum value.
 107         Add xgettext_token_type_rp.
 108         (struct xgettext_token_ty): Add argnum member.
 109         line_number and file_name are now also set for
 110         xgettext_token_type_keyword.
 111         (xgettext_lex_keyword): Arg is const char *.
 112
 113         * src/xget-lex.c: Include "hash.h".
 114         (enum token_type_ty): Add token_type_rp.
 115         (keywords): Now a hash table.
 116         (phase5_get): Return token_type_rp for ')'.
 117         (xgettext_lex, xgettext_lex_keyword): Add support for keyword argnums.
 118         (xgettext_lex): Return xgettext_token_type_rp for ')'.
 119         Report keyword argnum, line number, and file name back to caller.
 120
 121 1998-07-09  Paul Eggert  <eggert@twinsun.com>
 122
 123         * intl/Makefile.in (uninstall):
 124         Do nothing unless $(PACKAGE) is gettext.
 125
 126 ===================================================================
 127 RCS file: doc/gettext.texi,v
 128 retrieving revision 0.10.35.0
 129 retrieving revision 0.10.35.1
 130 diff -pu -r0.10.35.0 -r0.10.35.1
 131 --- doc/gettext.texi    1998/05/01 05:53:32     0.10.35.0
 132 +++ doc/gettext.texi    1998/07/18 00:25:15     0.10.35.1
 133 @@ -1854,13 +1854,19 @@ List of directories searched for input f
 134  Join messages with existing file.
 135
 136  @item -k @var{word}
 137 -@itemx --keyword[=@var{word}]
 138 -Additonal keyword to be looked for (without @var{word} means not to
 139 +@itemx --keyword[=@var{keywordspec}]
 140 +Additonal keyword to be looked for (without @var{keywordspec} means not to
 141  use default keywords).
 142
 143 -The default keywords, which are always looked for if not explicitly
 144 -disabled, are @code{gettext}, @code{dgettext}, @code{dcgettext} and
 145 -@code{gettext_noop}.
 146 +If @var{keywordspec} is a C identifer @var{id}, @code{xgettext} looks
 147 +for strings in the first argument of each call to the function or macro
 148 +@var{id}.  If @var{keywordspec} is of the form
 149 +@samp{@var{id}:@var{argnum}}, @code{xgettext} looks for strings in the
 150 +@var{argnum}th argument of the call.
 151 +
 152 +The default keyword specifications, which are always looked for if not
 153 +explicitly disabled, are @code{gettext}, @code{dgettext:2},
 154 +@code{dcgettext:2} and @code{gettext_noop}.
 155
 156  @item -m [@var{string}]
 157  @itemx --msgstr-prefix[=@var{string}]
 158 ===================================================================
 159 RCS file: intl/Makefile.in,v
 160 retrieving revision 0.10.35.0
 161 retrieving revision 0.10.35.1
 162 diff -pu -r0.10.35.0 -r0.10.35.1
 163 --- intl/Makefile.in    1998/04/27 21:53:18     0.10.35.0
 164 +++ intl/Makefile.in    1998/07/09 21:39:18     0.10.35.1
 165 @@ -143,10 +143,14 @@ install-data: all
 166  installcheck:
 167
 168  uninstall:
 169 -       dists="$(DISTFILES.common)"; \
 170 -       for file in $$dists; do \
 171 -         rm -f $(gettextsrcdir)/$$file; \
 172 -       done
 173 +       if test "$(PACKAGE)" = "gettext"; then \
 174 +         dists="$(DISTFILES.common)"; \
 175 +         for file in $$dists; do \
 176 +           rm -f $(gettextsrcdir)/$$file; \
 177 +         done
 178 +       else \
 179 +         : ; \
 180 +       fi
 181
 182  info dvi:
 183
 184 ===================================================================
 185 RCS file: src/xget-lex.c,v
 186 retrieving revision 0.10.35.0
 187 retrieving revision 0.10.35.1
 188 diff -pu -r0.10.35.0 -r0.10.35.1
 189 --- src/xget-lex.c      1998/07/09 22:49:48     0.10.35.0
 190 +++ src/xget-lex.c      1998/07/18 00:25:15     0.10.35.1
 191 @@ -33,6 +33,7 @@
 192  #include "error.h"
 193  #include "system.h"
 194  #include "libgettext.h"
 195 +#include "hash.h"
 196  #include "str-list.h"
 197  #include "xget-lex.h"
 198
 199 @@ -83,6 +84,7 @@ enum token_type_ty
 200    token_type_eoln,
 201    token_type_hash,
 202    token_type_lp,
 203 +  token_type_rp,
 204    token_type_comma,
 205    token_type_name,
 206    token_type_number,
 207 @@ -109,7 +111,7 @@ static FILE *fp;
 208  static int trigraphs;
 209  static int cplusplus_comments;
 210  static string_list_ty *comment;
 211 -static string_list_ty *keywords;
 212 +static hash_table keywords;
 213  static int default_keywords = 1;
 214
 215  /* These are for tracking whether comments count as immediately before
 216 @@ -941,6 +943,10 @@ phase5_get (tp)
 217        tp->type = token_type_lp;
 218        return;
 219
 220 +    case ')':
 221 +      tp->type = token_type_rp;
 222 +      return;
 223 +
 224      case ',':
 225        tp->type = token_type_comma;
 226        return;
 227 @@ -1179,6 +1185,7 @@ xgettext_lex (tp)
 228    while (1)
 229      {
 230        token_ty token;
 231 +      void *keyword_value;
 232
 233        phase8_get (&token);
 234        switch (token.type)
 235 @@ -1213,17 +1220,20 @@ xgettext_lex (tp)
 236           if (default_keywords)
 237             {
 238               xgettext_lex_keyword ("gettext");
 239 -             xgettext_lex_keyword ("dgettext");
 240 -             xgettext_lex_keyword ("dcgettext");
 241 +             xgettext_lex_keyword ("dgettext:2");
 242 +             xgettext_lex_keyword ("dcgettext:2");
 243               xgettext_lex_keyword ("gettext_noop");
 244               default_keywords = 0;
 245             }
 246
 247 -         if (string_list_member (keywords, token.string))
 248 -           {
 249 -             tp->type = (strcmp (token.string, "dgettext") == 0
 250 -                         || strcmp (token.string, "dcgettext") == 0)
 251 -               ? xgettext_token_type_keyword2 : xgettext_token_type_keyword1;
 252 +         if (find_entry (&keywords, token.string, strlen (token.string),
 253 +                         &keyword_value)
 254 +             == 0)
 255 +           {
 256 +             tp->type = xgettext_token_type_keyword;
 257 +             tp->argnum = (int) keyword_value;
 258 +             tp->line_number = token.line_number;
 259 +             tp->file_name = logical_file_name;
 260             }
 261           else
 262             tp->type = xgettext_token_type_symbol;
 263 @@ -1236,6 +1246,12 @@ xgettext_lex (tp)
 264           tp->type = xgettext_token_type_lp;
 265           return;
 266
 267 +       case token_type_rp:
 268 +         last_non_comment_line = newline_count;
 269 +
 270 +         tp->type = xgettext_token_type_rp;
 271 +         return;
 272 +
 273         case token_type_comma:
 274           last_non_comment_line = newline_count;
 275
 276 @@ -1263,16 +1279,32 @@ xgettext_lex (tp)
 277
 278  void
 279  xgettext_lex_keyword (name)
 280 -     char *name;
 281 +     const char *name;
 282  {
 283    if (name == NULL)
 284      default_keywords = 0;
 285    else
 286      {
 287 -      if (keywords == NULL)
 288 -       keywords = string_list_alloc ();
 289 +      int argnum;
 290 +      size_t len;
 291 +      const char *sp;
 292 +
 293 +      if (keywords.table == NULL)
 294 +       init_hash (&keywords, 100);
 295 +
 296 +      sp = strchr (name, ':');
 297 +      if (sp)
 298 +       {
 299 +         len = sp - name;
 300 +         argnum = atoi (sp + 1);
 301 +       }
 302 +      else
 303 +       {
 304 +         len = strlen (name);
 305 +         argnum = 1;
 306 +       }
 307
 308 -      string_list_append_unique (keywords, name);
 309 +      insert_entry (&keywords, name, len, (void *) argnum);
 310      }
 311  }
 312
 313 ===================================================================
 314 RCS file: src/xget-lex.h,v
 315 retrieving revision 0.10.35.0
 316 retrieving revision 0.10.35.1
 317 diff -pu -r0.10.35.0 -r0.10.35.1
 318 --- src/xget-lex.h      1998/07/09 22:49:48     0.10.35.0
 319 +++ src/xget-lex.h      1998/07/18 00:25:15     0.10.35.1
 320 @@ -23,9 +23,9 @@ Foundation, Inc., 59 Temple Place - Suit
 321  enum xgettext_token_type_ty
 322  {
 323    xgettext_token_type_eof,
 324 -  xgettext_token_type_keyword1,
 325 -  xgettext_token_type_keyword2,
 326 +  xgettext_token_type_keyword,
 327    xgettext_token_type_lp,
 328 +  xgettext_token_type_rp,
 329    xgettext_token_type_comma,
 330    xgettext_token_type_string_literal,
 331    xgettext_token_type_symbol
 332 @@ -37,8 +37,14 @@ struct xgettext_token_ty
 333  {
 334    xgettext_token_type_ty type;
 335
 336 -  /* These 3 are only set for xgettext_token_type_string_literal.  */
 337 +  /* This 1 is set only for xgettext_token_type_keyword.  */
 338 +  int argnum;
 339 +
 340 +  /* This 1 is set only for xgettext_token_type_string_literal.  */
 341    char *string;
 342 +
 343 +  /* These 2 are set only for xgettext_token_type_keyword and
 344 +     xgettext_token_type_string_literal.  */
 345    int line_number;
 346    char *file_name;
 347  };
 348 @@ -50,7 +56,7 @@ void xgettext_lex PARAMS ((xgettext_toke
 349  const char *xgettext_lex_comment PARAMS ((size_t __n));
 350  void xgettext_lex_comment_reset PARAMS ((void));
 351  /* void xgettext_lex_filepos PARAMS ((char **, int *)); FIXME needed?  */
 352 -void xgettext_lex_keyword PARAMS ((char *__name));
 353 +void xgettext_lex_keyword PARAMS ((const char *__name));
 354  void xgettext_lex_cplusplus PARAMS ((void));
 355  void xgettext_lex_trigraphs PARAMS ((void));
 356
 357 ===================================================================
 358 RCS file: src/xgettext.c,v
 359 retrieving revision 0.10.35.0
 360 retrieving revision 0.10.35.1
 361 diff -pu -r0.10.35.0 -r0.10.35.1
 362 --- src/xgettext.c      1998/07/09 22:49:48     0.10.35.0
 363 +++ src/xgettext.c      1998/07/18 00:25:15     0.10.35.1
 364 @@ -835,6 +835,8 @@ scan_c_file(filename, mlp, is_cpp_file)
 365       int is_cpp_file;
 366  {
 367    int state;
 368 +  int commas_to_skip;  /* defined only when in states 1 and 2 */
 369 +  int paren_nesting;   /* defined only when in state 2 */
 370
 371    /* Inform scanner whether we have C++ files or not.  */
 372    if (is_cpp_file)
 373 @@ -854,63 +856,79 @@ scan_c_file(filename, mlp, is_cpp_file)
 374     {
 375       xgettext_token_ty token;
 376
 377 -     /* A simple state machine is used to do the recognising:
 378 +     /* A state machine is used to do the recognising:
 379          State 0 = waiting for something to happen
 380 -        State 1 = seen one of our keywords with string in first parameter
 381 -        State 2 = was in state 1 and now saw a left paren
 382 -       State 3 = seen one of our keywords with string in second parameter
 383 -       State 4 = was in state 3 and now saw a left paren
 384 -       State 5 = waiting for comma after being in state 4
 385 -       State 6 = saw comma after being in state 5  */
 386 +        State 1 = seen one of our keywords
 387 +        State 2 = waiting for part of an argument */
 388       xgettext_lex (&token);
 389       switch (token.type)
 390         {
 391 -       case xgettext_token_type_keyword1:
 392 +       case xgettext_token_type_keyword:
 393 +        if (!extract_all && state == 2)
 394 +          {
 395 +            if (commas_to_skip == 0)
 396 +              {
 397 +                error (0, 0,
 398 +                       _("%s:%d: warning: keyword nested in keyword arg"),
 399 +                       token.file_name, token.line_number);
 400 +                continue;
 401 +              }
 402 +
 403 +            /* Here we should nest properly, but this would require a
 404 +               potentially unbounded stack.  We haven't run across an
 405 +               example that needs this functionality yet.  For now,
 406 +               we punt and forget the outer keyword.  */
 407 +            error (0, 0,
 408 +                   _("%s:%d: warning: keyword between outer keyword and its arg"),
 409 +                   token.file_name, token.line_number);
 410 +          }
 411 +        commas_to_skip = token.argnum - 1;
 412          state = 1;
 413          continue;
 414
 415 -       case xgettext_token_type_keyword2:
 416 -        state = 3;
 417 -        continue;
 418 -
 419         case xgettext_token_type_lp:
 420          switch (state)
 421            {
 422            case 1:
 423 +            paren_nesting = 0;
 424              state = 2;
 425              break;
 426 -          case 3:
 427 -            state = 4;
 428 +          case 2:
 429 +            paren_nesting++;
 430              break;
 431 -          default:
 432 -            state = 0;
 433            }
 434          continue;
 435
 436 +       case xgettext_token_type_rp:
 437 +        if (state == 2 && paren_nesting != 0)
 438 +          paren_nesting--;
 439 +        else
 440 +          state = 0;
 441 +        continue;
 442 +
 443         case xgettext_token_type_comma:
 444 -        state = state == 5 ? 6 : 0;
 445 +        if (state == 2 && commas_to_skip != 0)
 446 +          commas_to_skip -= paren_nesting == 0;
 447 +        else
 448 +          state = 0;
 449          continue;
 450
 451         case xgettext_token_type_string_literal:
 452 -        if (extract_all || state == 2 || state == 6)
 453 -          {
 454 -            remember_a_message (mlp, &token);
 455 -            state = 0;
 456 -          }
 457 +        if (extract_all || (state == 2 && commas_to_skip == 0))
 458 +          remember_a_message (mlp, &token);
 459          else
 460            {
 461              free (token.string);
 462 -            state = (state == 4 || state == 5) ? 5 : 0;
 463 +            state = state == 2 ? 2 : 0;
 464            }
 465          continue;
 466
 467         case xgettext_token_type_symbol:
 468 -        state = (state == 4 || state == 5) ? 5 : 0;
 469 +        state = state == 2 ? 2 : 0;
 470          continue;
 471
 472         default:
 473 -        state = 0;
 474 -        continue;
 475 +        abort ();
 476
 477         case xgettext_token_type_eof:
 478          break;
 479 ===================================================================
 480 RCS file: po/Makefile.in.in,v
 481 retrieving revision 0.10.35.0
 482 retrieving revision 0.10.35.5
 483 diff -u -r0.10.35.0 -r0.10.35.5
 484 --- po/Makefile.in.in   1998/07/20 20:20:38     0.10.35.0
 485 +++ po/Makefile.in.in   1998/07/26 09:07:52     0.10.35.5
 486 @@ -62,7 +62,7 @@
 487         $(COMPILE) $<
 488
 489  .po.pox:
 490 -       $(MAKE) $(PACKAGE).pot
 491 +       $(MAKE) $(srcdir)/$(PACKAGE).pot
 492         $(MSGMERGE) $< $(srcdir)/$(PACKAGE).pot -o $*.pox
 493
 494  .po.mo:
 495 @@ -79,7 +79,7 @@
 496
 497  all: all-@USE_NLS@
 498
 499 -all-yes: cat-id-tbl.c $(CATALOGS)
 500 +all-yes: $(srcdir)/cat-id-tbl.c $(CATALOGS)
 501  all-no:
 502
 503  $(srcdir)/$(PACKAGE).pot: $(POTFILES)
 504 @@ -90,8 +90,8 @@
 505            || ( rm -f $(srcdir)/$(PACKAGE).pot \
 506                 && mv $(PACKAGE).po $(srcdir)/$(PACKAGE).pot )
 507
 508 -$(srcdir)/cat-id-tbl.c: stamp-cat-id; @:
 509 -$(srcdir)/stamp-cat-id: $(PACKAGE).pot
 510 +$(srcdir)/cat-id-tbl.c: $(srcdir)/stamp-cat-id; @:
 511 +$(srcdir)/stamp-cat-id: $(srcdir)/$(PACKAGE).pot
 512         rm -f cat-id-tbl.tmp
 513         sed -f ../intl/po2tbl.sed $(srcdir)/$(PACKAGE).pot \
 514                 | sed -e "s/@PACKAGE NAME@/$(PACKAGE)/" > cat-id-tbl.tmp
 515 @@ -180,7 +180,8 @@
 516
 517  check: all
 518
 519 -cat-id-tbl.o: ../intl/libgettext.h
 520 +cat-id-tbl.o: $(srcdir)/cat-id-tbl.c $(top_srcdir)/intl/libgettext.h
 521 +       $(COMPILE) $(srcdir)/cat-id-tbl.c
 522
 523  dvi info tags TAGS ID:
 524
 525 @@ -196,7 +197,7 @@
 526  maintainer-clean: distclean
 527         @echo "This command is intended for maintainers to use;"
 528         @echo "it deletes files that may require special tools to rebuild."
 529 -       rm -f $(GMOFILES)
 530 +       rm -f $(GMOFILES) cat-id-tbl.c stamp-cat-id
 531
 532  distdir = ../$(PACKAGE)-$(VERSION)/$(subdir)
 533  dist distdir: update-po $(DISTFILES)
 534 @@ -207,7 +208,7 @@
 535         done
 536
 537  update-po: Makefile
 538 -       $(MAKE) $(PACKAGE).pot
 539 +       $(MAKE) $(srcdir)/$(PACKAGE).pot
 540         PATH=`pwd`/../src:$$PATH; \
 541         cd $(srcdir); \
 542         catalogs='$(CATALOGS)'; \