Yonggang Luo, I've been trying to figure out how to understand what you
are doing for over a year now without success. You've been doing some
amazing work, and I get the vague impression that you are working on a
Chinese fork of Thunderbird. Now you are proposing some specific changes
to Thunderbird infrastructure, but you are really not a Thunderbird
contributor. Could you give some sort of public statement of how your
technical work fits into the broader picture of products?
I've been trying to get people working on open-source email clients to
talk to teach other and figure out ways to work together better. In my
dreams, we all work together to make an amazing product rather than have
small teams each duplicating each other's work. We'd love to be able to
work with you, even if you are not specifically contributing to
Thunderbird. But I have no context now to understand how.
R. Kent James
Chair, Thunderbird Council
On 12/3/2015 8:52 PM, 罗勇刚(Yonggang Luo) wrote:
I think the relationship between Thunderbird/SeaMonkey and
gecko-dev should be like the relation ship between Atom editor vs Electron.
So there is no need that Thunderbird/SeaMonkey chasing the mozilla-central.
And considerate the hard for contributing code in mozilla infrastructure
for thunderbird,
I suggest to completely to leave the mozilla infrastructure.
To reach this target, we need to do the following things.
1、All the repos are git based and hosted on github
I've already writing scripts to mirror all branches/tags of comm &
gecko-dev,
They are now placed at
https://github.com/mail-apps/comm
https://github.com/mail-apps/gecko-dev
I have the scripts to do that.
After the migration, we only need to maintain the gecko-dev mirror
2、Thunderbird/SeaMonkey bugs moved to github.
I didn't write the scripts yet,
3、The leaving Makefile.in should be clear up,
4、Removing all the #if #endif in
xul/js/mozilla files. This is not a must be procedure
We may hack the mozbuild system that
generate install_manifests for those files.
5、Leaving the C/C++ building system from moz.build
This is very important, cause the current moz.build for building
are to tight with the lowering mozilla building infrastructure, and if
we want to leaving the mozilla source tree, this would be a
unavoidable step.
For building C/C++, we may choose gyp/cmake/qbs or other means
that's was a self-contained echo-system and that's doesn't depends on
the complicated mozbuild eco-system.
For example: LLVM are leaving autoconf to CMake.
We may choose, cause Thunderbird/SeaMonkey is much
smaller than gecko-dev, so that's won't be a big deal.
6、 Building the All Thunderbird C/C++ compoents as a single
thunderbird.dll or other things. I've already implement that, and
doesn't
need to modify much of the source code.
7、Getting all the locales install to be build with pure python.
This is necessary, and because of this, we may also need to convert all
the locales
repo from hg to git.
Or we may using a much simpler method. place all the locales files into
the comm source-tree.
8、Building xulrunner distribution for Thunderbird/Gecko like
https://github.com/atom/electron/releases, so we only need to downloading
those binaries.
9、Packaging the mozbuild building system into the xulrunner, so that we
won't need
the whole mozilla source-tree.
10、 The xulrunner only chasing for stable version of firefox, unless there
is much need of new functional, then the xulrunner only based on ESR
version of firefox.
_______________________________________________
dev-platform mailing list
dev-platform@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-platform