Re: [Gtk-osx-devs] Third and Forth problem with gnucash on 10.6

[Vincent Lucarelli <vincent lucarelli gmail com>]

Hi John,

You do know that I maintain OSX App bundles for Gnucash, right? The download link for the 2.2.9 stable release is on the website (http:www.gnucash.org). Very highly recommended for production use. You can help the project by testing the unstable release on a copy of your data, and of course you will need to get svn going if you're going to write patches.

I've committed myself to using gnucash (moved 12 years worth of data and not running Quicken in parallel).  So I'm a little nervous that

there isn't a critical mass of people involved with gnucash on OSX and one day I'll be stuck.  So, I'm trying to get involved and not just

download a precompiled binary.  I'll starting test out 2.3.10 on a copy of my data.

The project wiki (for which http://gtk-osx.sourceforge.net provides an easy entry point) has information about some of the issues one encounters building gtk-osx projects in general, including a page on 64-bit.

Do in particular be careful to keep your MacPorts and Gtk-OSX stuff separate (paying particular attention to the various paths set in environment variables). Cross-pollenating between the two has been known to make a mess.

There are some significant issues with both Pango and Cairo over the transition from ATSUI to CoreText, particularly with Pango.

We (meaning Paul Davis of Ardour and I along with a few less-consistent kvetchers) have badgered one of the Gtk+ core devs, 

Kristian Reitveld, into doing some work on the quartz implementations and getting patches incorporated. So, yes, the goal is to move everything that makes sense into the appropriate library. In the meantime, Gtk-OSX can provide add-on libraries like ige-mac-integration and patchfiles which are automatically applied by jhbuild to tarballs. Unfortunately, jhbuild has no provision for automatically patching VCS checkouts (which makes sense, because the risk is quite high that a patch won't apply cleanly against a

checkout unless the patch is in some dusty corner of the code).

Thanks for this info.

Vince