Bug#963983: closed by Gianfranco Costamagna (Re: python3-pyside2.qtsvg: segfault on import)

2020-07-01 Thread Dmitry Shachnev
Control: severity -1 important

On Tue, Jun 30, 2020 at 12:50:36AM +0200, ydir...@free.fr wrote:
> The other pyside2 packages (code, gui etc) were still at 5.15.0-1~exp1.
>
> If I just update qtsvg to 5.15.0-2 the segfault still happens.
>
> Update the rest of pyside2 triggers a huge cascading of upgrades to Qt 5.14
> which had not been done when I had installed the exp packages, and when I
> updated qtsvg initially.
>
> Among them:
>
>  Unpacking libqt5svg5:amd64 (5.14.2-2) over (5.12.5-2) ...
>
> With all this upgraded, I confirm there is no segfault any more.
>
> Thus, the problem is likely caused by an undeclared versioned dependency on an
> underlying lib, and libqt5svg5 looks like a good candidate.  

I have now bumped Qt build-dependencies in Git, that should help.

In the mean time, downgrading the severity of this bug as it should not
block Qt migration to testing (especially when we require new Qt).

--
Dmitry Shachnev


signature.asc
Description: PGP signature


Bug#963983: closed by Gianfranco Costamagna (Re: python3-pyside2.qtsvg: segfault on import)

2020-06-29 Thread ydirson
reopen 963983
thanks

The other pyside2 packages (code, gui etc) were still at 5.15.0-1~exp1.

If I just update qtsvg to 5.15.0-2 the segfault still happens.

Update the rest of pyside2 triggers a huge cascading of upgrades to Qt 5.14
which had not been done when I had installed the exp packages, and when I
updated qtsvg initially.

Among them:

 Unpacking libqt5svg5:amd64 (5.14.2-2) over (5.12.5-2) ...

With all this upgraded, I confirm there is no segfault any more.

Thus, the problem is likely caused by an undeclared versioned dependency on an
underlying lib, and libqt5svg5 looks like a good candidate.  

Not dealing with it is likely to lead to other insufficient partial updates.