GNU bug report logs - #36402
Installer: null pointer exception during partitioning

Previous Next

Full log

Message #38 received at 36402 <at> debbugs.gnu.org (full text, mbox):

From: Ludovic Courtès <ludo <at> gnu.org>
To: Mathieu Othacehe <m.othacehe <at> gmail.com>
Cc: 36402 <at> debbugs.gnu.org, Juan <r5jm <at> protonmail.com>
Subject: Re: bug#36402: installation error
Date: Thu, 05 Sep 2019 10:32:18 +0200

Hello!

Mathieu Othacehe <m.othacehe <at> gmail.com> skribis:

> And here is what was going wrong:
>
> ped_device_get and ped_device_get_next can return pointers to already
> existing device object. So set-pointer-finalizer! was possibly called
> multiple times on the same device pointer, resulting in calling
> ped_device_destroy multiple times on the same device pointer.
>
> To prevent that, I created a weak value hash table to make sure that one
> <device> object maps to exactly one device pointer, and that the pointer
> finalizer is set only once. See commit b35839b.

Good catch!

I confirm that:

  guix build guile-parted --with-branch=guile-parted=master --check

passed several times in a row.  :-)

b35839b LGTM!

(‘define-wrapped-pointer-type’ takes care of this, but we can’t use it
while we use bytestructures (info "(guile) Void Pointers and Byte
Access").)

It seems to me that the fix should be not just for ‘pointer->device!’
but for all the ‘pointer->RECORD!’ procedures, where we potentially have
similar scenarios, and where we’d rather have:

  (eq? (pointer->X ptr) (pointer->X ptr))

So perhaps you should define your own ‘define-wrapped-type’ macro that
does ‘define-record-type’ + the weak hash table thing, and replace all
‘define-record-type’ instances in structs.scm with
‘define-wrapped-type’.  How does that sound?

Thank you!

Ludo’.

Previous Next

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