Re: Support print0 with completions

[Shigio YAMAGUCHI]

Hello,
> My request is simply to consistently add the same
> behavior for print0 with completions.

OK. Below is a patch that fixes the bug.
It will be included in the next release.
Thank you!

Regards,
Shigio

Index: global/global.c
===================================================================
RCS file: /sources/global/global/global/global.c,v
retrieving revision 1.319
diff -c -r1.319 global.c
*** global/global.c 1 Mar 2021 22:23:56 -0000 1.319
--- global/global.c 8 Jul 2021 01:44:17 -0000
***************
*** 148,153 ****
--- 148,154 ----
  char *single_update;
  char *path_style;
  char *print_target;
+ int newline = '\n';

  /*
   * Path filter
***************
*** 791,798 ****
  xflag = 0;
  if (nflag > 1)
  nosource = 1; /* to keep compatibility */
! if (print0)
  set_print0();
  if (cflag && match_part == 0)
  match_part = MATCH_PART_ALL;
  /*
--- 792,801 ----
  xflag = 0;
  if (nflag > 1)
  nosource = 1; /* to keep compatibility */
! if (print0) {
! newline = '\0';
  set_print0();
+ }
  if (cflag && match_part == 0)
  match_part = MATCH_PART_ALL;
  /*
***************
*** 1025,1031 ****
  flags |= GTOP_IGNORECASE;
  for (gtp = gtags_first(gtop, prefix, flags); gtp; gtp = gtags_next(gtop)) {
  fputs(gtp->tag, stdout);
! fputc('\n', stdout);
  count++;
  }
  if (debug)
--- 1028,1034 ----
  flags |= GTOP_IGNORECASE;
  for (gtp = gtags_first(gtop, prefix, flags); gtp; gtp = gtags_next(gtop)) {
  fputs(gtp->tag, stdout);
! fputc(newline, stdout);
  count++;
  }
  if (debug)
***************
*** 1171,1177 ****
  continue;
  }
  *p = '\0';
! puts(line);
  }
  #if (defined(_WIN32) && !defined(__CYGWIN__)) || defined(__DJGPP__)
  if (pclose(ip) != 0)
--- 1174,1181 ----
  continue;
  }
  *p = '\0';
! fputs(line, stdout);
! fputc(newline, stdout);
  }
  #if (defined(_WIN32) && !defined(__CYGWIN__)) || defined(__DJGPP__)
  if (pclose(ip) != 0)
***************
*** 1236,1242 ****
  gfind_close(gp);
  for (path = dbop_first(dbop, NULL, NULL, DBOP_KEY); path != NULL;
path = dbop_next(dbop)) {
  fputs(path, stdout);
! fputc('\n', stdout);
  }
  dbop_close(dbop);
  }
--- 1240,1246 ----
  gfind_close(gp);
  for (path = dbop_first(dbop, NULL, NULL, DBOP_KEY); path != NULL;
path = dbop_next(dbop)) {
  fputs(path, stdout);
! fputc(newline, stdout);
  }
  dbop_close(dbop);
  }

  }

On Wed, Jul 7, 2021 at 10:25 PM Ergus <spacibba@aol.com> wrote:
>
> Hi:
>
> Recently we are using global from emacs and it works pretty fine in most
> of the case. We always use in the package consistently --print0 to
> simplify the parsing code, BUT we found that --completion option does
> not support --print0. It just ignores it and prints one
> candidate/line. My request is simply to consistently add the same
> behavior for print0 with completions.
>
> Thanks in advance,
> Ergus.
>
> PD: BTW, I don't know if it is too complex, but if you have some shared
> object where you expose some C bindings it may be very useful to provide
> an emacs module, and inter-language support for global throw python for
> example.
>


--
Shigio YAMAGUCHI <shigio@gnu.org>
PGP fingerprint:
26F6 31B4 3D62 4A92 7E6F  1C33 969C 3BE3 89DD A6EB