GNU bug report logs - #73204
Tree-sitter in clojure-ts-mode causes memory leak during parsing

Previous Next

Full log

Message #14 received at 73204 <at> debbugs.gnu.org (full text, mbox):

From: vinurs <hi <at> vinurs.me>
To: Yuan Fu <casouri <at> gmail.com>
Cc: 73204 <at> debbugs.gnu.org
Subject: Re: bug#73204: Tree-sitter in clojure-ts-mode causes memory
 leak during parsing
Date: Sat, 14 Sep 2024 17:14:59 +0800

[Message part 1 (text/plain, inline)]

Hi,
Thank you for your fix! I just compiled the latest Emacs-30 code and tested several files that had problems before. It’s working very well so far.

------------

Sincerely,



Haiyuan Zhang, Vinurs

BinaryBrain LLC


开 2024年9月14日 在 15:59:05, Yuan Fu (casouri <at> gmail.com) 写到:



> On Sep 12, 2024, at 9:24 PM, Yuan Fu <casouri <at> gmail.com> wrote:  
>  
>  
>  
>> On Sep 12, 2024, at 9:21 AM, vinurs <hi <at> vinurs.me> wrote:  
>>  
>> Hi,  
>>  
>> I am using the precompiled version from https://emacsformacosx.com/builds  
>> https://emacsformacosx.com/emacs-builds/Emacs-2024-09-11_00-09-10-69e1aca041c57ba425425d31471e1c8f86d3bf04-universal.dmg  
>>  
>> I have chosen the latest Emacs 30 version. The min configuration is as follows:  
>>  
>> (add-to-list 'treesit-extra-load-path (expand-file-name "tree-sitter/" user-emacs-directory))  
>> (add-to-list 'load-path "~/.emacs.d/site-lisp/clojure-ts-mode”)  
>> (require 'clojure-ts-mode)  
>> ;; debug  
>> (setq debug-on-quit t)  
>>  
>> and I only use clojure-ts-mode package  
>>  
>> cd ~/.emacs.d/site-lisp  
>> git clone https://github.com/clojure-emacs/clojure-ts-mode.git  
>>  
>> Then, after opening Emacs,  
>> I opened the attached file b.clj, moved to line 57,  
>> and entered the following characters:  
>> (comment  
>>  
>> At this point, Emacs freezes and becomes unresponsive. I noticed that Emacs' memory usage keeps increasing:  
>> 2024-09-12 21:12:23: 806624 KB  
>> 2024-09-12 21:12:28: 810720 KB  
>> 2024-09-12 21:12:33: 829840 KB  
>> 2024-09-12 21:12:38: 1582256 KB  
>> 2024-09-12 21:12:43: 4678816 KB  
>> 2024-09-12 21:12:48: 7831776 KB  
>> 2024-09-12 21:12:53: 10982384 KB  
>> 2024-09-12 21:12:58: 14140048 KB  
>> 2024-09-12 21:13:03: 17296176 KB  
>> 2024-09-12 21:13:08: 20440960 KB  
>> 2024-09-12 21:13:13: 23583008 KB  
>> 2024-09-12 21:13:18: 26726272 KB  
>> 2024-09-12 21:13:24: 29857312 KB  
>> 2024-09-12 21:13:29: 33011456 KB  
>> 2024-09-12 21:13:34: 36156624 KB  
>> 2024-09-12 21:13:39: 38685440 KB  
>> 2024-09-12 21:13:44: 34858560 KB  
>> 2024-09-12 21:13:49: 30892768 KB  
>>  
>> Then, Emacs eventually gets killed by the system.  
>> The bt.log is the information I was able to retrieve  
>> ------------  
>> Sincerely,<b.clj><bt.log>  
>  
> Hi vinurs,  
>  
> Thank you for your detailed report! This looks very similar to bug#72863, which we determined to be a tree-sitter bug. In both that bug and this, ts_parser_parse falls into an inf loop.  
>  
> Your test file also uncovered another Emacs bug. I need fix this ASAP (emacs 30 pre-release is already out 😬). Once that’s fixed I’ll check if the hang is due to the same tree-sitter bug.  
>  
> Yuan  

I fixed the bug I discovered; and guess what, that totally fixed this bug! So this bug isn’t caused by the tree-sitter bug, despite having similar symptoms.  

When the next pretest comes out, please give it a try and see if you can still see the bug. Or you can clone the emacs-30 branch and check it out now.  

Thanks again for the report! Would’ve been really bad if we didn’t discover this bug before releasing Emacs 30.  

Yuan  

[Message part 2 (text/html, inline)]

Previous Next

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