[PATCH]: Allowing embedded newlines in strings.

[Jan Nieuwenhuizen]

Hi,

Below is a patch that optionally allows xgettext to grok strings that
include raw (unescaped) newlines.  Although this is strictly not
allowed for C/C++, gcc/g++ groks these strings, and we find them
especially useful (in LilyPond) for mimicking in-line literate
programming.

If there's anything else I can do to get this merged, please tell me.
Greetings,

Jan.

diff -purN --exclude=*~ --exclude=*.info* --exclude=*.html 
../gettext-0.11.2/doc/ChangeLog ./doc/ChangeLog
--- ../gettext-0.11.2/doc/ChangeLog     Wed Apr 24 19:50:31 2002
+++ ./doc/ChangeLog     Thu Jun 13 18:06:44 2002
@@ -1,3 +1,8 @@
+2002-06-13  Jan Nieuwenhuizen  <address@hidden>
+
+       * xgettext.texi: Document --embedded-newlines and
+       --no-embedded-newlines options.
+
 2002-04-24  Bruno Haible  <address@hidden>
 
        * gettext-0.11.2 released.
diff -purN --exclude=*~ --exclude=*.info* --exclude=*.html 
../gettext-0.11.2/doc/xgettext.texi ./doc/xgettext.texi
--- ../gettext-0.11.2/doc/xgettext.texi Tue Apr 23 13:33:34 2002
+++ ./doc/xgettext.texi Thu Jun 13 18:04:37 2002
@@ -107,6 +107,18 @@ POT file.
 Place comment block with @var{tag} (or those preceding keyword lines)
 in output file.
 
address@hidden -z
address@hidden --no-embedded-newlines
address@hidden address@hidden, @code{xgettext} option}
address@hidden address@hidden, @code{xgettext} option}
+Do not allow raw newlines embedded in strings.
+
address@hidden -Z
address@hidden --embedded-newlines
address@hidden address@hidden, @code{xgettext} option}
address@hidden address@hidden, @code{xgettext} option}
+Allow raw newlines embedded in strings.
+
 @end table
 
 @subsection Language=C/C++ specific options
diff -purN --exclude=*~ --exclude=*.info* --exclude=*.html 
../gettext-0.11.2/src/ChangeLog ./src/ChangeLog
--- ../gettext-0.11.2/src/ChangeLog     Wed Apr 24 19:50:31 2002
+++ ./src/ChangeLog     Thu Jun 13 17:53:21 2002
@@ -1,3 +1,16 @@
+2002-06-13  Jan Nieuwenhuizen  <address@hidden>
+
+       * x-python.c (phase7_getuc): Handle --embedded-newlines option.
+
+       * x-c.c (phase5_get): Handle --embedded-newlines option.
+
+       * xgettext.c (embedded_newlines): New variable.
+       (long_options): Add --embedded-newlines, --no-embedded-newlines
+       options.
+       (main): Parse --embedded-newlines, --no-embedded-newlines options.
+
+       * xgettext.h (embedded_newlines): New declaration.
+
 2002-04-24  Bruno Haible  <address@hidden>
 
        * gettext-0.11.2 released.
diff -purN --exclude=*~ --exclude=*.info* --exclude=*.html 
../gettext-0.11.2/src/x-c.c ./src/x-c.c
--- ../gettext-0.11.2/src/x-c.c Tue Feb 12 13:30:34 2002
+++ ./src/x-c.c Thu Jun 13 17:33:42 2002
@@ -971,12 +971,17 @@ phase5_get (tp)
          c = phase7_getc ();
          if (c == P7_NEWLINE)
            {
-             error_with_progname = false;
-             error (0, 0, _("%s:%d: warning: unterminated string literal"),
-                    logical_file_name, line_number - 1);
-             error_with_progname = true;
-             phase7_ungetc ('\n');
-             break;
+             if (embedded_newlines)
+               c = '\n';
+             else
+               {
+                 error_with_progname = false;
+                 error (0, 0, _("%s:%d: warning: unterminated string literal"),
+                        logical_file_name, line_number - 1);
+                 error_with_progname = true;
+                 phase7_ungetc ('\n');
+                 break;
+               }
            }
          if (c == EOF || c == P7_QUOTES)
            break;
diff -purN --exclude=*~ --exclude=*.info* --exclude=*.html 
../gettext-0.11.2/src/x-python.c ./src/x-python.c
--- ../gettext-0.11.2/src/x-python.c    Mon Mar 11 13:17:55 2002
+++ ./src/x-python.c    Thu Jun 13 18:20:31 2002
@@ -394,6 +394,8 @@ phase7_getuc (quote_char, triple, interp
              *backslash_counter = 0;
              return '\n';
            }
+         if (embedded_newlines)
+             return '\n';
          phase1_ungetc (c);
          error_with_progname = false;
          error (0, 0, _("%s:%d: warning: unterminated string"),
diff -purN --exclude=*~ --exclude=*.info* --exclude=*.html 
../gettext-0.11.2/src/xgettext.c ./src/xgettext.c
--- ../gettext-0.11.2/src/xgettext.c    Tue Apr 23 13:33:35 2002
+++ ./src/xgettext.c    Thu Jun 13 17:56:00 2002
@@ -81,6 +81,9 @@ static char *comment_tag;
    equality.  */
 bool substring_match;
 
+/* If true, grok newline characters embedded in a string. */
+bool embedded_newlines;
+
 /* Name of default domain file.  If not set defaults to messages.po.  */
 static const char *default_domain;
 
@@ -119,6 +122,7 @@ static const struct option long_options[
   { "debug", no_argument, &do_debug, 1 },
   { "default-domain", required_argument, NULL, 'd' },
   { "directory", required_argument, NULL, 'D' },
+  { "embedded-newlines", no_argument, NULL, 'Z' },
   { "escape", no_argument, NULL, 'E' },
   { "exclude-file", required_argument, NULL, 'x' },
   { "extract-all", no_argument, NULL, 'a' },
@@ -133,6 +137,7 @@ static const struct option long_options[
   { "language", required_argument, NULL, 'L' },
   { "msgstr-prefix", optional_argument, NULL, 'm' },
   { "msgstr-suffix", optional_argument, NULL, 'M' },
+  { "no-embedded-newlines", no_argument, NULL, 'z' },
   { "no-escape", no_argument, NULL, 'e' },
   { "no-location", no_argument, &line_comment, 0 },
   { "no-wrap", no_argument, NULL, CHAR_MAX + 3 },
@@ -220,7 +225,7 @@ main (argc, argv)
   default_domain = MESSAGE_DOMAIN_DEFAULT;
 
   while ((optchar = getopt_long (argc, argv,
-                                "ac::Cd:D:eEf:Fhijk::l:L:m::M::no:p:sTVw:x:",
+                                "ac::Cd:D:eEf:Fhijk::l:L:m::M::no:p:sTVw:x:zZ",
                                 long_options, NULL)) != EOF)
     switch (optchar)
       {
@@ -356,6 +361,12 @@ main (argc, argv)
       case 'x':
        read_exclusion_file (optarg);
        break;
+      case 'Z':
+       embedded_newlines = true;
+       break;
+      case 'z':
+       embedded_newlines = false;
+       break;
       case CHAR_MAX + 1:       /* --copyright-holder */
        copyright_holder = optarg;
        break;
@@ -582,6 +593,8 @@ Operation mode:\n\
   -x, --exclude-file=FILE.po     entries from FILE.po are not extracted\n\
   -c, --add-comments[=TAG]       place comment block with TAG (or those\n\
                                  preceding keyword lines) in output file\n\
+  -z, --no-embedded-newlines     forbid newlines embedded in strings 
(default)\n\
+  -Z, --embedded-newlines        allow newlines embedded in strings\n\
 "));
       printf ("\n");
       /* xgettext: no-wrap */
diff -purN --exclude=*~ --exclude=*.info* --exclude=*.html 
../gettext-0.11.2/src/xgettext.h ./src/xgettext.h
--- ../gettext-0.11.2/src/xgettext.h    Wed Jan 30 14:46:48 2002
+++ ./src/xgettext.h    Thu Jun 13 17:25:46 2002
@@ -33,6 +33,8 @@ extern int xgettext_omit_header;
 
 extern bool substring_match;
 
+extern bool embedded_newlines;
+
 /* Split keyword spec into keyword, argnum1, argnum2.  */
 extern void split_keywordspec PARAMS ((const char *spec, const char **endp,
                                       int *argnum1p, int *argnum2p));


-- 
Jan Nieuwenhuizen <address@hidden> | GNU LilyPond - The music typesetter
http://www.xs4all.nl/~jantien       | http://www.lilypond.org