| 1 |
On Sun, Apr 21, 2019 at 8:03 PM <andrey.vul@×××××.com> wrote: |
| 2 |
> |
| 3 |
> |
| 4 |
> On Sun, Apr 21, 2019, 05:23 Adam Carter, <adamcarter3@×××××.com> wrote: |
| 5 |
>> |
| 6 |
>> On Sun, Apr 21, 2019 at 12:24 PM <andrey.vul@×××××.com> wrote: |
| 7 |
>>> |
| 8 |
>>> Hi, folks. |
| 9 |
>>> |
| 10 |
>>> When texinfo tries to run to process the texinfo files as part of the |
| 11 |
>>> compile step, perl appears to die with free(): invalid pointer. |
| 12 |
>>> |
| 13 |
>>> build.log: http://dpaste.com/2Z0K93S |
| 14 |
>>> emerge --info: http://dpaste.com/1KD9J3R |
| 15 |
>>> |
| 16 |
>> |
| 17 |
>> If you've just updated perl (i just got 5.28.2 today, ~amd64) you'll need to run perl-cleaner to get your perl stuff working again |
| 18 |
> |
| 19 |
> No change before or after, and texinfo doesn't use Perl modules not part of the standard distribution. |
| 20 |
> Is there a way to get the texinfo build to run Perl in debug mode so something like a backtrace can be emitted? |
| 21 |
|
| 22 |
I have a backtrace for sys-apps/texinfo-6.6-r1: |
| 23 |
(gdb) bt |
| 24 |
#0 __GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:50 |
| 25 |
#1 0x00007f9edf3397a5 in __GI_abort () at abort.c:79 |
| 26 |
#2 0x00007f9edf395450 in __libc_message |
| 27 |
(action=action@entry=do_abort, fmt=fmt@entry=0x7f9edf4a7707 "%s\n") at |
| 28 |
../sysdeps/posix/libc_fatal.c:181 |
| 29 |
#3 0x00007f9edf39cfd8 in malloc_printerr |
| 30 |
(str=str@entry=0x7f9edf4a589b "free(): invalid pointer") at |
| 31 |
malloc.c:5341 |
| 32 |
#4 0x00007f9edf39e964 in _int_free (av=<optimized out>, p=<optimized |
| 33 |
out>, have_lock=<optimized out>) at malloc.c:4165 |
| 34 |
#5 0x00007f9edfbd3b8a in xs_abort_empty_line |
| 35 |
(self=self@entry=0x55b7cc609128, current=current@entry=0x55b7cc901bc8, |
| 36 |
additional_spaces_in=additional_spaces_in@entry=0x55b7cc901e68) at misc.c:74 |
| 37 |
#6 0x00007f9edfbd48ea in xs_merge_text |
| 38 |
(self=self@entry=0x55b7cc609128, current=0x55b7cc901bc8, |
| 39 |
text_in=text_in@entry=0x55b7cbf15408) |
| 40 |
at misc.c:303 |
| 41 |
#7 0x00007f9edfbd2da4 in XS_Texinfo__MiscXSXS_merge_text |
| 42 |
(my_perl=0x55b7cb1f9290, cv=<optimized out>) at MiscXS.c:266 |
| 43 |
#8 0x00007f9edf8391f4 in Perl_pp_entersub (my_perl=0x55b7cb1f9290) at |
| 44 |
pp_hot.c:5232 |
| 45 |
#9 0x00007f9edf7f326a in Perl_runops_debug (my_perl=0x55b7cb1f9290) |
| 46 |
at dump.c:2535 |
| 47 |
#10 0x00007f9edf75a58d in S_run_body (oldscope=1, |
| 48 |
my_perl=0x55b7cb1f9290) at perl.c:2694 |
| 49 |
#11 perl_run (my_perl=0x55b7cb1f9290) at perl.c:2617 |
| 50 |
#12 0x000055b7cabf428d in main (argc=<optimized out>, argv=<optimized |
| 51 |
out>, env=<optimized out>) at perlmain.c:122 |
| 52 |
(gdb) q |
| 53 |
|
| 54 |
By the looks of it, something suspicious is going on in the compile of |
| 55 |
MiscXS.so. This is similar to https://bugs.gentoo.org/622576 but |
| 56 |
because the issue occurred with sys-apps/texinfo-6.6-r1, it can't be |
| 57 |
exactly that. |
Archives