Re: abortion when doing i x twice

[Gavin Smith]

On Mon, May 15, 2017 at 08:38:50PM +0200, Benno Schulenberg wrote:
> 
> On Mon, May 15, 2017, at 18:39, Gavin Smith wrote:
> > On Mon, May 15, 2017 at 10:55:18AM +0200, Benno Schulenberg wrote:
> > > To reproduce:
> > > 
> > > MALLOC_CHECK_=2  info/ginfo doc/info-stnd.info
> > > 
> > > Then type:  i  x  <Enter>  i  x  <Enter>
> > > 
> > > Aborted.
> > 
> > I could find nothing wrong here at all.  Do you have valgrind installed?
> > Maybe try running
> > 
> > alias VAL='valgrind --log-file=val.log'
> > VAL  info/ginfo doc/info-stnd.info
> 
> ==5362== Invalid write of size 1
> ==5362==    at 0x4026107: strcpy (mc_replace_strmem.c:311)
> ==5362==    by 0x805751D: replace_in_documentation (infodoc.c:816)
> ==5362==    by 0x805068E: report_index_match (indices.c:437)
> ==5362==    by 0x8050793: info_next_index_match (indices.c:485)
> ==5362==    by 0x8050162: info_index_search (indices.c:264)
> ==5362==    by 0x805EAA5: info_read_and_dispatch (session.c:253)
> ==5362==    by 0x805EA01: info_session (session.c:221)
> ==5362==    by 0x8055E25: main (info.c:1079)
> ==5362==  Address 0x4b652cf is 0 bytes after a block of size 71 alloc'd
> ==5362==    at 0x4024F20: malloc (vg_replace_malloc.c:236)
> ==5362==    by 0x806CA84: xmalloc (xmalloc.c:41)
> ==5362==    by 0x8056F5B: replace_in_documentation (infodoc.c:663)
> ==5362==    by 0x805068E: report_index_match (indices.c:437)
> ==5362==    by 0x8050793: info_next_index_match (indices.c:485)
> ==5362==    by 0x8050162: info_index_search (indices.c:264)
> ==5362==    by 0x805EAA5: info_read_and_dispatch (session.c:253)
> ==5362==    by 0x805EA01: info_session (session.c:221)
> ==5362==    by 0x8055E25: main (info.c:1079)

The first thing I notice is that it says there were 71 bytes allocated,
but the call in 'report_index_match' has fewer characters, about 61. So
you are not using English error messages. What locale settings are you
using? (I tried using LANG=de_DE.utf8 but didn't reproduce the error.)

The other idea is that many message strings have been changed when 
removing full stops, and this was one of the few that haven't changed,
so this message will likely be one of those that has a translation 
available.