Bug#534943: reportbug: python / GTK error about being unable to open display when used from console

[Sandro Tosi]

Jakub,

On Sun, Jun 28, 2009 at 16:40, Jakub Wilk<ubanus at users.sf.net> wrote:
> Sandro,
>
> I think I'd be better to keep this bugreport (or perhaps its clone) assigned
> to reportbug, at least to avoid more duplicate reports.

I think the solution is fix it in python-gtk2

>>> /usr/lib/pymodules/python2.5/gtk-2.0/gtk/__init__.py:72: GtkWarning:
>>> could not open display
>>>  warnings.warn(str(e), _gtk.Warning)
>>
>> this is a bug (if it's a bug) in python-gtk2, hence reassignig it to
>> that pacakge.
>
> I'd disagree.
>
> I find it perfectly reasonable that python-gtk2 issues a warning; in *most
> cases* the module has limited usefulness when $DISPLAY is not set to a valid
> value.

ok, so this is informative for a user, but how can this be useful for a program?

> Moreover, the issue can be easily fixed in reportbug: by suppressing the
> warning or by checking if $DISPLAY is set before trying to load the module.
> (I could provide a patch if you wish.)

No, this is to be done in python-gtk2. IMNSHO.

>>> This is mainly cosmetic, but reportbug should not rely on there being
>>> more
>>> than a console mode terminal available.
>>
>> We have to know what UIs are available: if there is no DISPLAY, gtk should
>> not issue a warning but raising an exception.
>
> A Python module should not raise exceptions at import time without
> a good reason. The solution you propose would break pydoc, for instance.

If it can't work without DISPLAY, then raise an exception; if it can
work without DISPLAY then remote the worning.

> BTW, I am currenly working around this bug by using the attached wrapper.

It doesn't seem an elegant solution.

Regards,
-- 
Sandro Tosi (aka morph, morpheus, matrixhasu)
My website: http://matrixhasu.altervista.org/
Me at Debian: http://wiki.debian.org/SandroTosi