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

Re: QT on alpha potato does not compile



On Thu, 7 Sep 2000, Ullrich Martini wrote:

> I am trying to compile qt 2.2 on a alpha with potato, g++ 2.95.2-13
> using  rkrustys patches from the
> intel qt 2.2 diffs. I get lots of internal compiler errors on the files
> generated by moc (moc_*.cpp), and
> uic segfaults when compiling  tools/designer/designer/listboxeditor.ui
> 
> It looks like a alpha-gcc problem. The internal compiler errors go away
> if the files
> are compiled without optimization. I will file a bug against gcc because
> of that.

Good idea...make sure that the bug submission has complete info for
forwarding to the gcc list (see their submission guidelines).  I only ask
because I'll probably be the one reproducing it when someone asks :-)

> The uic problem occurred when building the designer.  I think the Qt
> designer should go into
> a separate package anyway.

Agreed, but I'm not the maintainer, so my opinions matter little :-P

> The kde people had similiar bugs against kde2 betas in their bugtracking
> system, but they closed them because they had a modified version of qt
> which appearently fixed the problems. It looks like those patches didn't
> find their way back to the trolls.
> Anyway, did someone succeed in building qt-2.2 on a alpha/potato?

I removed optimisation for the woody builds of the beta Qt
versions.  We run into quite a few optimiser bugs on Alpha, so we're used
to working around them if they're easily defeatable.  FYI, though, most of
the ICEs that I ran into were because of problems in the code that I'm
compiling.  Either way, they shouldn't have come up (since the code was
technically correct, albeit loosely), but it may be worth looking at the
offending code to see what it might be choking on.  I've also found that
it helps to try compiling the offending files with '-save-temps -v' so
that I can see what's going on in the generated assembly (if any) and from
a command point of view.  I then play with the optimisation switches (the
individual ones, not just -O2/-O1/etc) to see which is causing the
problem and hopefully why.  Don't think I got around to it yet on the Qt
code, though.

C


-- 
To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org


Reply to: