[PyQt] Re: configure.py: -e semantic

[Phil Thompson]

On Tuesday 10 July 2007 5:47 pm, Giovanni Bajo wrote:
> On 7/10/2007 6:33 PM, Phil Thompson wrote:
> >> currently, -e MODULE means "activate MODULE, checking that it's built in
> >> Qt too".
> >
> > No, it means "enable checks for MODULE", the default being "enable checks
> > for all modules".
> >
> >> This is a little counter-intuitive to me: I would expect "-e
> >> MODULE" to mean "I really really want MODULE to be there", and thus
> >> configure.py to abort if the module(s) I specified are not built in Qt
> >> and thus cannot be compiled.
> >>
> >> Do you instead consider the current behaviour better, or at least just
> >> as potentially useful? If not, I will submit a patch to change it so
> >> that configure.py fails if the module can't be activated.
> >>
> >> Otherwise, would you mind if I add a way to tell configure.py to abort
> >> if the enabled modules can't be activated?
> >
> > I'd rather leave it as it is.
>
> To avoid wasting time, I want to add a check in my build scripts that
> the Qt build I'm building against is correct, that is it contains all
> the modules I need. I have far too many builds around and always manage
> to get something wrong.
>
> What would be the suggested way to check (after configure.py and before
> make, possibly) which modules have been enabled? I thought of checking
> the existince of the Qt* directories in the PyQt directory, but those
> directories could already be there before configure.py.

Import the generated pyqtconfig and check the value of pyqt_modules.

Phil