Joost and Nakata,
Is there any documentation within the qa project describing these
processes a little, that might help?
:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.:.
Best Regards
Per Eriksson
Co-Lead Swedish Language Project
Marketing Contact Sweden
OpenOffice.org Community
Phone: +46 70 560 10 33
Email: [EMAIL PROTECTED]
Web: http://sv.openoffice.org/
tora - Takamichi Akiyama skrev:
Hi,
Andre Schnabel wrote:
Well .. a valid answer for OpenSource develoopers to the question
"why do you this and that" is "Because I can!" ;)
Good answer!
for the current issue, thare are some positive points, that we have
double builds for Mac (these are the same as having double builds for
Windows / Linux):
- Sun is not providing PPC builds - we have to use Maho's builds for
PPC. In case, we find a defect, it is easier to proofe, if the defect
is in PPC-port or in Maho's build environment, as we now can compare
two different processor architectures from one build provider.
Good.
- if we find problems in Sun's builds, we can easily test if these
are source issues or if these were introduced by the build
environment (we had this sometimes for Linux / windows - comparing
Sun's and Pavel's builds was sometimes very helpfull)
Not actually.
Normally, when we encounter or are reported an unexpected behavior
around OpenOffice.org, doing the same reproduction instructions among
different versions of stable builds and developer's build on various
platforms. e.g. 2.4.1 on Windows, 2.4.1 on Linux, DEV300_m30 on Windows,
DEV300_m30 on Linux.
I think comparison between builds based on identical source codes,
built by different persons and different build environments does not
help for investigation of a bug.
Build process must be reliable. If we can not trust the build process,
we can not work on the bug. Efforts of debugging his/her build
instructions and/or build environment should be avoided.
If the person who built and widely announced his/her builds in the
public ML, releases@ and he/she does not know what is the difference
from Sun's build. Why can we trust the builds?
- we do not only release binaries, we release the source also. Mahos
builds verify, that the source works correctly on a vanilla build
environment.
Not only Maho but also lots of people on the globe has built OOo
from the source codes and reported many problems in the ML dev@
[EMAIL PROTECTED], other relevant MLs, and/or issue tracker.
Maho's efforts are highly helpful. But it is not the only one.
- Sun does not provide builds for all localizations - so some native
lang teams need to use Maho's buiilds.
It can be applied if the native lang teams officially release them.
You know, only 10s order of native lang teams release OpenOffice.org.
Enough reasons? ;)
I am afraid, no.
The original point in this thread is to avoid confusion regarding
the duplication of the Max OS X Intel builds.
Pavel does not announce his builds in the ML releases@ while Maho does.
Why?
One more point is that what is the official build. It, however, seems
that the decision is left to each local community.
Best,
Tora
---------------------------------------------------------------------
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]
