Hi James,
At 17.51 13/06/2005 -0700, James Berry wrote:
Here I am again with my weekly check-in and query about the state of
the 3.0 unstable branch.
I've got the rest of this week in town before I'm out through mid- July.
Do to that schedule, I want to see if we can get the 3.0
unstable branch merged to trunk during the next couple of days, in
order that it doesn't diverge too much from trunk, and so that I can
help with any issues that come up before I leave.
Thanks to all of you, the 3.0 unstable branch is now more stable,
though I'm sure there's still work to do. In particular Dave,
Alberto, Axel, and Greg have all contributed bug reports and platform-
specific fixes. Here's an update to the platform support (your
feedback and corrections welcome):
Configure/Build success:
- Mac OS X 10.4 jberry (command line build, not XCode)
- FreeBSD 4.11 jberry (minotaur.apache.org)
- Mingw ver? amassari
I tried both MinGW 3.1 and 4.1; currently it builds fine but has issues
with multithreading and automatic generation of the DLL.
- cygwin Axel
- SuSE Linux 9.3 Axel (gcc version 3.3.5 20050117
(prerelease) (SUSE Linux), GNU Make 3.80)
- Solaris (ver?) Greg Franks (Sunpro compiler)
- Solaris dbertoni (GCC -- with notes?)
- Windows VS.net amassari
Also Visual C++ 6.0 is building; as I don't have Visual .NET 2002 (aka VS
7.0) I cannot update those projects. Unless someone else owns it, I think
we could safely delete them from the source tree (there should be almost no
users of that version, as the 7.1 upgrade was sold at a nominal fee for
existing users).
The following are unknown, by me anyway:
- AIX (issues, according to dave)
- Unix Services for Windows (is that what it's called?)
The official name is Windows Services for Unix 3.5 (or Interix, from the
company that really provides the software); on that platform I can run
configure, but it doesn't write a valid config.h (it writes one with all
the #define commented out). Maybe because it writes a few warnings about
invalid tests inside "configure"...
- NetBSD
- OpenBSD
- True64
I still have to do the Borland migration, but it shouldn't take long (and
being not a strategic platform it can be done on the main trunk)
So, I am voting +1 to do the merge at the end of this week.
Alberto
Since we're building on many more platforms now, I propose that we
now merge to trunk. My intent was not that we do all development on
my temporary 3.0-unstable branch, but that we use it to get the thing
buildable for many of our developers.
Furthermore, I'd like to hold off making any major non-build-related
changes on the unstable branch, so that it doesn't diverge too much
from trunk.
So my proposal in short form:
- Merge 3.0 unstable to trunk, with your consensus, over the
next couple of days.
- Concentrate effort to fix any remaining non-building platform
that is critical to developer progress.
- Perform remainder of work for Xerces 3.0 on the main trunk.
Feedback?
James.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
