Re: Word wrapping in PO

[Dwayne Bailey]

That should solve things, but from my limited reading on the cod ethis
will only solce it for formats understood by Gettext.  This is OK, but
some of our PO files that are holding OpenOffice.org data e.g. 
%PROGRAMNAME% will still break.

On Wed, 2007-12-26 at 16:45 +0100, Bruno Haible wrote:
> Dwayne Bailey wrote in [1]:
> > Two more cases I've now seen that make things rather confusing.
> > 
> > Variables: %s can break over two lines
> > Tags: </b> it is perfectly permissible to break these over lines.
> > 
> > Unfortunately the Unicode algorithms are very focused on text and will
> > break anywhere not always in ideal places.
> 
> I'm applying this; it fixes the problem with breaking format directives like
> %s.
> 
> 2007-12-22  Bruno Haible  <address@hidden>
> 
>       * write-po.c (wrap): Avoid breaking line in the middle of a format
>       directive.
>       Reported by Dwayne Bailey <address@hidden>.
> 
> diff -r -c3 --exclude='*.po*' --exclude='*.info*' --exclude='*_*.html' 
> --exclude='*.*.html' --exclude='*.[13]' --exclude='*.1.in' 
> --exclude=Makefile.in --exclude=aclocal.m4 --exclude=configure 
> --exclude=config.h.in --exclude=version.texi --exclude=stamp-vti 
> --exclude='po-*-gen*.[ch]' --exclude='*.o' --exclude='*.lo' --exclude='*.gmo' 
> --exclude=ABOUT-NLS --exclude=gnulib-lib --exclude=gnulib-m4 
> --exclude=Makefile.gnulib --exclude='javadoc[12]' --exclude=CVS 
> gettext-cvs/gettext-tools/src/write-po.c 
> gettext-work/gettext-tools/src/write-po.c
> *** gettext-cvs/gettext-tools/src/write-po.c  Mon Oct 22 04:01:35 2007
> --- gettext-work/gettext-tools/src/write-po.c Sat Dec 22 12:39:40 2007
> ***************
> *** 522,527 ****
> --- 522,528 ----
>   {
>     const char *canon_charset;
>     char *fmtdir;
> +   char *fmtdirattr;
>     const char *s;
>     bool first_line;
>   #if HAVE_ICONV
> ***************
> *** 583,589 ****
>   
>     /* Determine the extent of format string directives.  */
>     fmtdir = NULL;
> !   if (is_stylable (stream) && value[0] != '\0')
>       {
>         bool is_msgstr =
>       (strlen (name) >= 6 && memcmp (name, "msgstr", 6) == 0);
> --- 584,591 ----
>   
>     /* Determine the extent of format string directives.  */
>     fmtdir = NULL;
> !   fmtdirattr = NULL;
> !   if (value[0] != '\0')
>       {
>         bool is_msgstr =
>       (strlen (name) >= 6 && memcmp (name, "msgstr", 6) == 0);
> ***************
> *** 597,604 ****
>           struct formatstring_parser *parser = formatstring_parsers[i];
>           char *invalid_reason = NULL;
>           void *descr;
> !         char *fdp;
> !         char *fd_end;
>   
>           fmtdir = XCALLOC (len, char);
>           descr = parser->parse (value, is_msgstr, fmtdir, &invalid_reason);
> --- 599,607 ----
>           struct formatstring_parser *parser = formatstring_parsers[i];
>           char *invalid_reason = NULL;
>           void *descr;
> !         const char *fdp;
> !         const char *fd_end;
> !         char *fdap;
>   
>           fmtdir = XCALLOC (len, char);
>           descr = parser->parse (value, is_msgstr, fmtdir, &invalid_reason);
> ***************
> *** 607,617 ****
>   
>           /* Locate the FMTDIR_* bits and transform the array to an array
>              of attributes.  */
>           fd_end = fmtdir + len;
> !         for (fdp = fmtdir; fdp < fd_end; fdp++)
>             if (*fdp & FMTDIR_START)
>               {
> !               char *fdq;
>                 for (fdq = fdp; fdq < fd_end; fdq++)
>                   if (*fdq & (FMTDIR_END | FMTDIR_ERROR))
>                     break;
> --- 610,622 ----
>   
>           /* Locate the FMTDIR_* bits and transform the array to an array
>              of attributes.  */
> +         fmtdirattr = XCALLOC (len, char);
> +         fdap = fmtdirattr;
>           fd_end = fmtdir + len;
> !         for (fdp = fmtdir, fdap = fmtdirattr; fdp < fd_end; fdp++, fdap++)
>             if (*fdp & FMTDIR_START)
>               {
> !               const char *fdq;
>                 for (fdq = fdp; fdq < fd_end; fdq++)
>                   if (*fdq & (FMTDIR_END | FMTDIR_ERROR))
>                     break;
> ***************
> *** 621,633 ****
>                      its end. It is a bug in the ->parse method.  */
>                   abort ();
>                 if (*fdq & FMTDIR_ERROR)
> !                 memset (fdp, ATTR_INVALID_FORMAT_DIRECTIVE, fdq - fdp + 1);
>                 else
> !                 memset (fdp, ATTR_FORMAT_DIRECTIVE, fdq - fdp + 1);
>                 fdp = fdq;
>               }
>             else
> !             *fdp = 0;
>   
>           break;
>         }
> --- 626,639 ----
>                      its end. It is a bug in the ->parse method.  */
>                   abort ();
>                 if (*fdq & FMTDIR_ERROR)
> !                 memset (fdap, ATTR_INVALID_FORMAT_DIRECTIVE, fdq - fdp + 1);
>                 else
> !                 memset (fdap, ATTR_FORMAT_DIRECTIVE, fdq - fdp + 1);
> !               fdap += fdq - fdp;
>                 fdp = fdq;
>               }
>             else
> !             *fdap = 0;
>   
>           break;
>         }
> ***************
> *** 733,744 ****
>       }
>         portion = XNMALLOC (portion_len, char);
>         overrides = XNMALLOC (portion_len, char);
> -       memset (overrides, UC_BREAK_UNDEFINED, portion_len);
>         attributes = XNMALLOC (portion_len, char);
>         for (ep = s, pp = portion, op = overrides, ap = attributes; ep < es; 
> ep++)
>       {
>         char c = *ep;
> !       char attr = (fmtdir != NULL ? fmtdir[ep - value] : 0);
>         if (is_escape (c))
>           {
>             switch (c)
> --- 739,754 ----
>       }
>         portion = XNMALLOC (portion_len, char);
>         overrides = XNMALLOC (portion_len, char);
>         attributes = XNMALLOC (portion_len, char);
>         for (ep = s, pp = portion, op = overrides, ap = attributes; ep < es; 
> ep++)
>       {
>         char c = *ep;
> !       char attr = (fmtdirattr != NULL ? fmtdirattr[ep - value] : 0);
> !       char brk = UC_BREAK_UNDEFINED;
> !       /* Don't break inside format directives.  */
> !       if (attr == ATTR_FORMAT_DIRECTIVE
> !           && (fmtdir[ep - value] & FMTDIR_START) == 0)
> !         brk = UC_BREAK_PROHIBITED;
>         if (is_escape (c))
>           {
>             switch (c)
> ***************
> *** 754,760 ****
>               }
>             *pp++ = '\\';
>             *pp++ = c;
> !           op++;
>             *op++ = UC_BREAK_PROHIBITED;
>             *ap++ = attr | ATTR_ESCAPE_SEQUENCE;
>             *ap++ = attr | ATTR_ESCAPE_SEQUENCE;
> --- 764,770 ----
>               }
>             *pp++ = '\\';
>             *pp++ = c;
> !           *op++ = brk;
>             *op++ = UC_BREAK_PROHIBITED;
>             *ap++ = attr | ATTR_ESCAPE_SEQUENCE;
>             *ap++ = attr | ATTR_ESCAPE_SEQUENCE;
> ***************
> *** 777,783 ****
>             *pp++ = '0' + (((unsigned char) c >> 6) & 7);
>             *pp++ = '0' + (((unsigned char) c >> 3) & 7);
>             *pp++ = '0' + ((unsigned char) c & 7);
> !           op++;
>             *op++ = UC_BREAK_PROHIBITED;
>             *op++ = UC_BREAK_PROHIBITED;
>             *op++ = UC_BREAK_PROHIBITED;
> --- 787,793 ----
>             *pp++ = '0' + (((unsigned char) c >> 6) & 7);
>             *pp++ = '0' + (((unsigned char) c >> 3) & 7);
>             *pp++ = '0' + ((unsigned char) c & 7);
> !           *op++ = brk;
>             *op++ = UC_BREAK_PROHIBITED;
>             *op++ = UC_BREAK_PROHIBITED;
>             *op++ = UC_BREAK_PROHIBITED;
> ***************
> *** 790,796 ****
>           {
>             *pp++ = '\\';
>             *pp++ = c;
> !           op++;
>             *op++ = UC_BREAK_PROHIBITED;
>             *ap++ = attr | ATTR_ESCAPE_SEQUENCE;
>             *ap++ = attr | ATTR_ESCAPE_SEQUENCE;
> --- 800,806 ----
>           {
>             *pp++ = '\\';
>             *pp++ = c;
> !           *op++ = brk;
>             *op++ = UC_BREAK_PROHIBITED;
>             *ap++ = attr | ATTR_ESCAPE_SEQUENCE;
>             *ap++ = attr | ATTR_ESCAPE_SEQUENCE;
> ***************
> *** 838,843 ****
> --- 848,855 ----
>                 insize = inptr - ep;
>                 memcpy_small (pp, ep, insize);
>                 pp += insize;
> +               *op = brk;
> +               memset_small (op + 1, UC_BREAK_PROHIBITED, insize - 1);
>                 op += insize;
>                 memset_small (ap, attr, insize);
>                 ap += insize;
> ***************
> *** 855,868 ****
>                     *pp++ = c;
>                     ep += 1;
>                     *pp++ = *ep;
> !                   op += 2;
>                     *ap++ = attr;
>                     *ap++ = attr;
>                   }
>                 else
>                   {
>                     *pp++ = c;
> !                   op++;
>                     *ap++ = attr;
>                   }
>               }
> --- 867,881 ----
>                     *pp++ = c;
>                     ep += 1;
>                     *pp++ = *ep;
> !                   *op++ = brk;
> !                   *op++ = UC_BREAK_PROHIBITED;
>                     *ap++ = attr;
>                     *ap++ = attr;
>                   }
>                 else
>                   {
>                     *pp++ = c;
> !                   *op++ = brk;
>                     *ap++ = attr;
>                   }
>               }
> ***************
> *** 1118,1123 ****
> --- 1131,1138 ----
>       }
>     while (*s);
>   
> +   if (fmtdirattr != NULL)
> +     free (fmtdirattr);
>     if (fmtdir != NULL)
>       free (fmtdir);
>   
> 
> [1] http://lists.gnu.org/archive/html/bug-gnu-utils/2007-12/msg00054.html
-- 
Dwayne Bailey
Translate.org.za

+27-12-460-1095 (w)
+27-83-443-7114 (cell)