[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: fixing debconf stdin bug



Blars Blarson <blarson@blars.org> wrote:
> In article <E1AfNAX-0000CF-S3@mid.downhill.at.eu.org> 
> ametzler@downhill.at.eu.org writes:
>>Blars Blarson <blarson@blars.org> wrote:
>>> I submitted a patch to the bug (142944) in debconf where it does not
>>> restore stdin when doing db_stop.

>> You are not trying to get this into debconf before sarge are you?

> If it doesn't break anything, why shouldn't it?

I was referencing
| Joey Hess believes that this patch will cause problems...

As JH is debconf's maintainer I took this warning quite seriously.

> If it does cause problems for more than one or two packages, I would
> agree that it should wait until the packages have been fixed.  (So
> most likely after sarge.)

No disagreement here.

> So far, I havn't found any evidence it breaks any package, and I have
> been looking.  (Added all packages that depend on debconf to my local
> mirror, had a program search through all postinst scripts, and
> manually examined 78 postinst scripts.)

> The open fd 3 to the debconf frontend is more likely to cause
> problems, and none of the 78 packages are doing anything special about
> that.  (If the fd isn't closed, the debconf process will be left
> waiting for input from the deamon the postinst started.)

If I understand you correctly this kind of problem depends on the
_daemon's_ sourcecode and not only on the maintainerscript and cannot
reliably be found just by checing the latter, but by installing the
package, can it?
               cu andreas
-- 
Hey, da ist ein Ballonautomat auf der Toilette!
Unofficial _Debian-packages_ of latest unstable _tin_
http://www.logic.univie.ac.at/~ametzler/debian/tin-snapshot/



Reply to: