I've talked with Michael Meeks, one of libole2's maintainers. He
seemed to be very receptive of our wanting to use and develop libole2.
According to him, as soon as we port that small subset of glib,
everything should work XP after we wrap a few small functions, like the
get_size() function. We apparently do *not* have to come up with a
replacement for mmap(). Libole2 apparently has a feature (coded in for
Caolan's uses initially) to either use the mmap()'d version or the
*non-mmap()'d* one based on a runtime switch. This clears up the only
real non-glib related portability problem.
Michael Meeks<[EMAIL PROTECTED]> and Arturo Tena
<[EMAIL PROTECTED]> are its two official maintainers and ask that
all patches be sent to them. Michael also posted a tarball today of a
preliminary libole2 release version 0.1.0 which can be found here:
http://www.gnome.org/~michael/libole2-0.1.0.tar.gz Of course, the
newest snapshots can be found on cvs, info on which can be found in a
previous post of mine or by digging through the gnome website.
There is a libole2 mailing list to post questions or comments to:
[EMAIL PROTECTED]
Lastly, Michael suggests using the existing code across platforms,
instead of rewriting parts of libole2 to wrap existing MS API calls, as
had been previously mentioned on the list. I think that this makes
sense, and is a good idea (tm). A breakage on one platform is likely to
affect all platforms and be fixed quickly. Also, this reduces code size
and improves readability as well as reduces total development time.
Time for me to get coding.... I should have a GLib subset in later today
<probably>
Dom
