Caolan McNamara

a test set of linux SRC680_m157 x86_64 openoffice.org rpms available from http://people.redhat.com/caolanm/x86_64

64bit OOo
Microsoft Office import/export works
Native format xml import/export works
Graphic import/export works (i.e. fixes missing graphics in help)
and happily gcj java wizards also work, which was cool

Actually I don’t know if anything doesn’t work, though I’m sure there’s a few lurking ULONG containing structs being SvStream::Writen somewhere.

A dump of the 64 bit patches I built with, sliced horizontally. Mostly a mixture of kendy and pavel’s work with altered binary IO love. Hopefully slice them up vertically next week, see what’s already been submitted and what hasn’t, get their status fixed up and get it all glued together. Then there’s the tedium of fixing binfilter which is curently blithely ignored, but that’ll be a do-what-the-doppelganger-modules-do

Stuffing in the various 64bit workspaces and patches into a test build and with a few cleanups last night I got a running working OOB680_m1 x86_64 OOo running seemingly happily on x86_64 this morning.

Though it is afflicted by problems loading native format documents from 32bit output.

foundations:
vunet’s OOo story on an OOo foundation seems to have made a few leaps not supported by some of the named parties in the story. louis’s reponse, and an IBM response. Personally I’m really dubious about the benefits of foundations. Sounds suspicously like a talking shop to me, which we have more than enough of.

font work:
Whacked up a simple patch to honour fontconfig’s embeddedbitmap and antialiasing per-font overrides in OOo. Though meanwhile our layered glyph substitution mechanism seems a bit busted at the moment, and is baffling me a bit.