Re: Compile Error on FC5 (was "Re: gtkdatabox- released")

Hi Eric,

sounds weird. I would assume a compiler bug, too.

Currently I am thinking about getting rid of this part, anyway. It is
used in two signals only and they could also work with one pointer
instead of two. I have to give it a few more thoughts and expect to do
something here within the next two weeks.

It is the closely related to the top_left / bottom_right stuff mentioned
by Laurent. Maybe I am going to introduce a new struct which contains
all the data in one.



Eric Work wrote:
> Roland,
> I executed the two commands you mentioned below on both my 64-bit
> Ubuntu system and a 32-bit Fedora 8 VM and both produce the same
> output, identical to what's included in (except minor
> formatting).  I'm not sure why 64-bit Fedora 7/8/9 does not compile
> without that patch.  The strange thing is that 32-bit Fedora 7/8/9 and
> 64-bit Ubuntu 7.10 does compile fine.  Sounds like some kind
> of tool problem (gcc?).  From what I can tell glib-genmarshal seems to
> do the right thing tagging the function prototype.  It sounds like
> you're going to be making a few changes pretty soon so I'll wait until
> you get the next version out before I update the fedora repos.
> -Eric
> On Nov 27, 2007 1:56 PM, Roland Bock <box2OO6 eudoxos de> wrote:
>> Hi Eric,
>> thanks for the patch!
>> The patch worked fine for me and a new version is uploaded. I must admit
>> that I am a bit confused that that was necessary, because I created the
>> file using a glib tool:
>> glib-genmarshal --header --prefix=gtk_databox_marshal --internal
>> gtkdatabox_marshal.list
>> glib-genmarshal --body --prefix=gtk_databox_marshal  --internal
>> gtkdatabox_marshal.list
>> The header seems to be just fine, containing the "G_GNUC_INTERNAL" from
>> your patch, but the C file does not contain it.
>> Can you try to run the glib-genmarshal on your machine and see if it
>> produces code that works for you? In that case I think of removing the
>> gtkdatabox_marshal.[ch]. The files would be created by the make process.
>> Regards,
>> Roland

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