GNU bug report logs - #30626
26.0.91; Crash when traversing a `stream-of-directory-files'

Previous Next

Full log

View this message in rfc822 format

From: Noam Postavsky <npostavs <at> gmail.com>
To: John Mastro <john.b.mastro <at> gmail.com>
Cc: Michael Heerdegen <michael_heerdegen <at> web.de>, Nicolas Petton <nicolas <at> petton.fr>, 30626 <at> debbugs.gnu.org
Subject: bug#30626: 26.0.91; Crash when traversing a `stream-of-directory-files'
Date: Sun, 04 Mar 2018 10:56:36 -0500

Noam Postavsky <npostavs <at> gmail.com> writes:

> John Mastro <john.b.mastro <at> gmail.com> writes:
>
>> (let* ((stream0 (stream-range 1 1000000))
>>        (stream stream0))
>>   (setq stream0 nil) ;; <<< Inserted by compiler
>>   (while (not (stream-empty-p stream))
>>     (cl-callf stream-rest stream)))
>>
>> If the code does reference stream0 later, locals clearing can't help
>> you, but that's considered a "if it hurts, don't do it" situation.
>>
>> This probably isn't practical for Emacs, especially since it could only
>> work for byte-compiled code, but thought the prior art may be of
>> interest.
>
> Not sure how doable this solution is, but the fact that it works only
> for byte-compiled code seems fine to me.  The interpreted case is doomed
> to fail anyway since the interpreter doesn't prune redundant variables
> from closures.

Hmm, I think it won't work by itself though, just doing

    (stream-flush (stream-range 1 1000000))

also crashes, due to the head of the stream being referenced from the C
stack somewhere (I can get the address from gdb, but I can't figure out
how to get to the corresponding C variable from there).

Previous Next

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