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

Re: NEW: hydrogen

On Mon, Oct 03, 2005 at 07:56:18AM +0200, Marc Espie wrote:
> On Sun, Oct 02, 2005 at 07:03:48PM -0700, Jacob Meuser wrote:
> > but that's probably a dangerous patch.  some qt applications do
> > use ~/.qt for storing user configuration files.
> Of course it's a dangerous patch! that's not what I meant. No, it would mean
> really tweaking the code, so that the .qt creation doesn't happen unless
> there's really some file you need to store there... which is a tad bit more
> interesting.

I tested the patch, and AFAICS, it has no ill effects.  it just removes
~/.qt creation from the QSettings constructor.  the directory will be
created (and a config file written) in the destructor if any
settings have changed.  IMHO, this is correct behaviour: only make
the directory if it needs to be written to, not just because something
_might_ be done with the settings.

> > might it not be better to just set $HOME to somewhere that is OK to
> > write to?
> Well, to me, it's a bug in qt.  I don't see how a toolkit has a business
> mucking with prefs files if you just run uic3/moc3...

uic is calling QSettings::writeEntry() (thus modifying the settings)
when it creates .cpp files, but it doesn't call the function when
creating .h files.  but I cannot figure out where in uic writeEntry()
is getting called.  maybe this will give someone some clues though.