Re: Gtk2-1.021 make test failures



On Sun, 2004-01-04 at 18:41, muppet wrote:
On Sunday, January 4, 2004, at 04:35 PM, Ross McFarland wrote:
and this is the problem.  it will pick up TreeView.pm from the old  
inline stuff, which will pick up _config.pm and attempt to do the  
Inline booting, which will look for symbols that do not exist in the  
Gtk2.so that was already loaded by "use Gtk2".  all makes perfect sense  
now.

but i find it odd that this require does not fail at all on any of my  
systems, which do not have TreeView.pm anywhere.  i've been using that  
construct for quite some time.  8-(

i have a feeling that it requires it in a fail safe way, if it's not
there just go about its business. 


i don't know what to tell you.

well, basically, it means that we need to set up SimpleList differently  
to avoid this problem.

i'll get this here in a few. really all it should need is 

push @ISA, 'Gtk2::TreeView';

but i'll make sure of that.


Thanks again for your patience, am I making this 500 times harder than
it needs to be?? :-)

it's not your fault the problems are with mandrake having (for
understandable reasons) distributed a non-trivially modified version.

whoa, hang on --- Kevin's talking about mdk 9.1, which contains the old  
Inline-based gtk2-perl stuff from 4Q02/1Q03, not mdk 9.2, which was  
released 4Q03 with modified versions of Glib and Gtk2 0.95.

by the way:  
http://rpmfind.net/linux/RPM/mandrake/9.2/SRPMS/perl-Gtk2-0.95-
6mdk.src.html and  
http://rpmfind.net/linux/RPM/mandrake/9.2/SRPMS/perl-Glib-0.95-
3mdk.src.html --- IIRC, the many patches add support for XSetInputFocus  
(for the installer), several deprecated functions (for porting their  
old code), a few patches we later applied, and a hack for exception  
handling (which hadn't been fixed in our cvs by the time they hit their  
code freeze).

i'd forgotten about that. i was thinking 9.1 was the 0.95 hacked. 


-rm




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