tag 549758 + confirmed retitle 549758 gtk+extra2: FTBFS: GtkEntry has no member named n_bytes quit
On Mon, 5 Oct 2009 20:23:34 +0200 Lucas Nussbaum <lu...@lucas-nussbaum.net> wrote: > During a rebuild of all packages in sid, your package failed to build on > amd64. I've been waiting for this. http://lists.debian.org/debian-devel/2009/05/msg00627.html The bug title blames this section: > configure.in:33: error: possibly undefined macro: LT_CURRENT > If this token and others are legitimate, please use > m4_pattern_allow. See the Autoconf documentation. > configure.in:34: error: possibly undefined macro: LT_REVISION > configure.in:35: error: possibly undefined macro: LT_AGE However, without running autoreconf (and therefore without updating configure or generating this warning), the package FTBFS on amd64 with the same error: > Relevant part: > > /bin/bash ../libtool --tag=CC --mode=compile gcc -DHAVE_CONFIG_H -I. -I.. > > -I -I.. -D_REENTRANT -I/usr/include/gtk-2.0 -I/usr/lib/gtk-2.0/include > > -I/usr/include/atk-1.0 -I/usr/include/cairo -I/usr/include/pango-1.0 > > -I/usr/include/pixman-1 -I/usr/include/freetype2 -I/usr/include/directfb > > -I/usr/include/libpng12 -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include > > -g -O2 -Wall -MT gtkitementry.lo -MD -MP -MF .deps/gtkitementry.Tpo > > -c -o gtkitementry.lo gtkitementry.c > > libtool: compile: gcc -DHAVE_CONFIG_H -I. -I.. -I -I.. -D_REENTRANT > > -I/usr/include/gtk-2.0 -I/usr/lib/gtk-2.0/include -I/usr/include/atk-1.0 > > -I/usr/include/cairo -I/usr/include/pango-1.0 -I/usr/include/pixman-1 > > -I/usr/include/freetype2 -I/usr/include/directfb -I/usr/include/libpng12 > > -I/usr/include/glib-2.0 -I/usr/lib/glib-2.0/include -g -O2 -Wall -MT > > gtkitementry.lo -MD -MP -MF .deps/gtkitementry.Tpo -c gtkitementry.c -fPIC > > -DPIC -o .libs/gtkitementry.o > > gtkitementry.c: In function 'gtk_entry_real_insert_text': > > gtkitementry.c:696: error: 'GtkEntry' has no member named 'n_bytes' This breakage was inevitable and, AFAICT, is not worth fixing. (Indeed, I don't think it *is* fixable without completely breaking/removing GtkSheet and requiring a SONAME bump in gtk+extra2 which is an utterly pointless exercise or copying large parts of Gtk+ << 2.17 into the gtk +extra2 sources which is even more pointless.) gtk+extra2 should be removed as soon as something is sorted out for libgtkada or, if that cannot be done soon, despite libgtkada. If it wasn't for libgtkada (and only libgtkada), gtk+extra2 would already have been removed from Debian unstable and testing. http://lists.debian.org/debian-devel/2009/06/msg00977.html Ludovic: now is the time to adopt gtk+extra2 and/or subsume it into Ada. I don't think we can wait for upstream to do something in 8 months time. Honestly, I don't think adoption is the right approach. Either libgtkada subsumes the parts of gtk+extra2 that it needs into it's own sources or libgtkada just has to get by without gtk+extra2 completely *and quickly*. I really don't see any other option - there is no good reason to keep gtk+extra2 around until Ada upstream can do something about it in June 2010. gtk+extra2 needs to be gone before the end of 2009, certainly long before any freezes begin for Squeeze. gtk+extra2 is already orphaned, quicklist has already been removed from testing and unstable and the inevitable FTBFS bug is now present in gtk +extra2. I cannot help feeling that this bug is like the head of the Hydra - remove one and two more will follow. I'm certainly not willing to tackle it. Ubuntu has an FTBFS issue too: https://bugs.launchpad.net/ubuntu/+source/gtk+extra2/+bug/422184 There's no build log linked from that bug though, so I can't be certain it isn't the same bug. This Debian bug reports a final fail in gtkitementry.c which is part of the GtkSheet codebase which also gets the blame for the Ubuntu bug as a result of changes in GTK+ 2.17. $ dpkg -l libgtk2.0-dev libgtk2.0-dev 2.18.1-1 I think my own views are clear - gtk+extra2 is not maintainable and has no place in Squeeze IMHO. I've already orphaned the package, I have absolutely no intention of fixing this FTBFS by any other means other than reassigning/cloning it to ftp.debian.org as an RM RoQA - something needs to be done with respect of libgtkada urgently. I always suspected that GtkSheet was using Gtk+ in ways that were not strictly part of the GTK+ API - it was never just about deprecated symbols that could be handled smoothly through transitions. I had deep misgivings (but not enough time to find proof) that gtk+extra2 was being underhand in how it utilised parts of Gtk+. I do not think that gtkitementry.c is the only instance of this approach. Upstream were being too clever by half and then decided to move on to other things. Can't help thinking that someone upstream knew this would happen. I'm regretting my decision to maintain gtk+extra2 but, at the time, I wanted to see quicklist survive. Live and learn - let's dump gtk+extra2 over the side ASAP. Please. I'm sorry for libgtkada but I think we have to get rid of gtk+extra2 despite such issues - and quickly. -- Neil Williams ============= http://www.data-freedom.org/ http://www.nosoftwarepatents.com/ http://www.linux.codehelp.co.uk/
pgpzqQvhNPIHM.pgp
Description: PGP signature
