Dear Paul,
I wrote to mono-fedora a couple of weeks ago as below. (and before that, I followed your postings, etc about mono 2.8 quite closely). Anyhow, I grabbed 2.8-8 off koji, made two changes (I made the first one, then it complained about unchecked files, so I added the 2nd):
---------------------------------------- @ -337,7 +337,7 @@
%configure --with-ikvm-native=yes --with-jit=yes --with-xen_opt=yes \ --with-moonlight=yes --with-profile2=yes \ - --with-libgdiplus=installed --with-sgen=no \ + --with-libgdiplus=installed --with-sgen=yes \ %if ! %{with_mono4} --with-profile4=no \ %endif @@ -398,6 +398,7 @@ %defattr(-,root,root,-) %doc AUTHORS COPYING.LIB ChangeLog NEWS README %{_bindir}/mono +%{_bindir}/mono-sgen %{_bindir}/monodir %{_bindir}/mono-test-install %{_bindir}/mono-gdb.py -----------------------
and it built to completion with f14 x86_64 (my system is basically up-to-date f14 - just with also libgdiplus grabbed from koji). So could you have another look and/or see what's wrong with on koji's build farm?
Thanks. Hin-Tak
--- On Thu, 2/12/10, Hin-Tak Leung htl10@users.sourceforge.net wrote:
Hi - I have been following the progress of integration of mono 2.8 into fedora. e.g. http://koji.fedoraproject.org/koji/buildinfo?buildID=206128
I am a little disappointed with the sgen being disabled on 64-bit build - I assume that means the simple generational gabbage collector, rather than 'sgen - Mono XML Serializer GENerator'. That's the change I have been waiting for almost two years for some applications I run which does strange things memory-wise and needs a large heap otherwise(?).
Can somebody post and summarise the problem? I am on f14 and have a mono git (and regularly cross-compile mono with some local changes) and I'd like to get this fixed, if something needs fixing.