GNU bug report logs - #72238
31.0.50; Segfault after treesit_delete_query

Previous Next

Full log

Message #26 received at 72238 <at> debbugs.gnu.org (full text, mbox):

From: Stefan Kangas <stefankangas <at> gmail.com>
To: Yuan Fu <casouri <at> gmail.com>
Cc: Noah Peart <noah.v.peart <at> gmail.com>, Eli Zaretskii <eliz <at> gnu.org>,
 john muhl <jm <at> pub.pink>, 72238 <at> debbugs.gnu.org
Subject: Re: bug#72238: 31.0.50; Segfault after treesit_delete_query
Date: Sun, 29 Sep 2024 18:42:12 -0700

Yuan Fu <casouri <at> gmail.com> writes:

>> On Jul 22, 2024, at 4:17 AM, Eli Zaretskii <eliz <at> gnu.org> wrote:
>>
>>> Cc: Noah Peart <noah.v.peart <at> gmail.com>, 72238 <at> debbugs.gnu.org
>>> From: Yuan Fu <casouri <at> gmail.com>
>>> Date: Sun, 21 Jul 2024 17:01:31 -0700
>>>
>>>
>>>
>>>> On Jul 21, 2024, at 3:24 PM, john muhl <jm <at> pub.pink> wrote:
>>>>
>>>> Noah Peart <noah.v.peart <at> gmail.com> writes:
>>>>
>>>>> After recent build, emacs is crashing frequently when using tree-sitter
>>>>> modes.
>>>>>
>>>>> The following backtrace was generated after opening a few C files,
>>>>> enabling c-ts-mode, and moving the cursor around a bit. I dont have
>>>>> precise steps for repeatability at this point (sorry!), as it seems to
>>>>> occur during garbage collection.
>>>>>
>>>>> Let me know if there is more useful output I could provide.
>>>>
>>>> Could you try reverting 1abf3bdd7ed and see if that helps?
>>>
>>> Hey guys, sorry for the segfault. It’s indeed caused by 1abf3bdd7ed, I just fixed it on master.
>>
>> Thanks, but I'm confused, since I see a fix both on emacs-30 and on
>> master.  What exactly did you fix and on which branch(es)?  Did the
>> problem exist on emacs-30 (it was reported for master)?
>
> Yeah, there’re two bugs: 1) we shouldn’t delete the query object without
> checking for NULL, and 2) we shouldn’t delete the cursor object without checking
> for NULL. In emacs-30, only query could be NULL, cursor is always initialized;
> in emacs-master, due to my recent change, both query and cursor could be NULL.
>
> So I added checks for both query and cursor in one commit and pushed to master, and cherry-picked the commit to emacs-30.
>
> Yuan

Should this bug be closed then?

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.