[PyQt] PyQt5 5.7.1 Bug: debug build on Windows broken

[Hans-Peter Jansen]

On Dienstag, 17. Januar 2017 18:15:42 Dmitry Shachnev wrote:
> Hi all,
> 
> On Tue, Jan 17, 2017 at 02:00:09PM +0100, Hans-Peter Jansen wrote:
> > Phil, please keep in mind, that *many* Linux distributions just build
> > debug
> > versions of *all* packages in order to maintain the symbol tables.
> > 
> > The build process strips the symbol tables from binaries later on, and
> > packages them separately in debug info (and debug source) packages.
> > 
> > If you debug a crash with gdb, it lists all debug info packages, that
> > needs be be installed for a full backtrace with symbols.
> > 
> > At openSUSE, we already patch configure.py to remove the *implicit*
> > qml_debug flag, since enabling this option has security implications (in
> > that it opens a local network port on the host):
> > 
> > Index: b/configure.py
> > ===================================================================
> > --- a/configure.py
> > +++ b/configure.py
> > @@ -2243,7 +2243,7 @@ def pro_add_qt_dependencies(target_confi
> > 
> >          pro_lines.append('QT += %s' % ' '.join(add))
> >      
> >      pro_lines.append(
> > 
> > -            'CONFIG += %s' % ('debug qml_debug' if debug else 'release'))
> > +            'CONFIG += %s' % ('debug' if debug else 'release'))
> > 
> >      if metadata.cpp11:
> >          pro_lines.append('CONFIG += c++11')
> 
> We now have this patch in Debian too (thanks Pete!).

Hehe. Funnily, the fix_qreal_check.diff mentioned below in "openSUSE patches of PyQt" is 
yours. Thank you, Dmitry. The world "feels" small sometimes.

You might want to comment on this, why Phil should consider your approach.

Cheers,
Pete
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.riverbankcomputing.com/pipermail/pyqt/attachments/20170117/ae890c78/attachment.html>