[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
Re: [platform-releng-dev] Build I-200411091436 is unusable
|
I don't think this is about process. Breaking a build not only forces the
releng team to start a new one, but this also makes a lot of people waste
their time. I'm waiting for new API from a component, and I cannot use it
because an unrelated component has made the build unusable.
Self-hosting on an integration build is also important as this helps us
detect problems early. The later an integration build is available, the
later the problems will be detected.
For these reasons, I think integration builds should be taken more
seriously.
Jerome
Dorian Birsan
<birsan@xxxxxxxxx
m> To
Sent by: platform-releng-dev@xxxxxxxxxxx
platform-releng-d cc
ev-admin@eclipse.
org Subject
Re: [platform-releng-dev] Build
I-200411091436 is unusable
11/10/2004 04:46
PM
Please respond to
platform-releng-d
ev
-1
I think the process makes sense for the milestone builds, but I find that
putting too much process around integration builds distracts us from doing
real work.
The releng project page shows two (slightly inconsistent) schedules for
integration builds, but none of them mentions the Tuesday morning as being
THE build.
|---------------+-----+-----+---------------------------------------------|
| Build Type | Date| Date| Failure Policy |
| | (Loc| (UTC| |
| | al) | ) | |
|---------------+-----+-----+---------------------------------------------|
| 3.1 | Wed | Wed | Same day rebuild if compile errors, missing |
| Integration | 8:00| 13:0| plugins, the build is unusable or if the |
| builds from | EDT | 0 | drops do not get posted.Otherwise, a rebuild|
| *.map files if| | UTC | will occur Thursday 8:00 EDT. |
| requested | | | |
|---------------+-----+-----+---------------------------------------------|
then:
Build Notes
We have decided to make some minor changes in our build story in order
to better focus the responsibility for getting good integration builds
on the developers:
For the 3.1 integration builds:
- build time is 8:00 am (Eastern time) on Tuesday mornings.
- in the event of an integration build failure due to compile errors,
missing plugins or if the drops are not posted,
there will be a same day rebuild upon submission of a fix.
- Otherwise, there will be a rebuild at 12:00 pm Wednesday morning.
- in the event of a double integration build failure, all other work
will stop until
guaranteed-to-be-successful build contributions are prepared for a
rebuild at 8:00 am Thursday morning.
- if the Thursday 8:00 am (Eastern time) integration build fails we will
rebuild until success
- once a good integration build is achieved, it will be marked as such
on the website.
-the releng team will often request a go/no go status from all teams to
determine if the build is suitable for the week's testing. If you
provide status and then recontribute to a subsequent build, you are
required to revalidate the build and provide an updated status.
-Dorian
Olivier
Thomann/Ottawa/IBM@IBMCA
Sent by: To
platform-releng-dev-admin@ecli platform-releng-dev@eclipse.o
pse.org rg
cc
11/10/2004 10:38 AM Subject
Re: [platform-releng-dev]
Build I-200411091436 is
Please respond to unusable
platform-releng-dev
+1.
Each submission for a rebuild must be explained on this mailing list before
the corresponding map is updated. If no major issue is raised against a
component, that component cannot submit new code on a rebuild. It has to
wait for next integration build.
Olivier
Pascal
Rapicault/Ottawa/IBM@IBMCA
Sent by:
platform-releng-dev-admin@ecli To
pse.org platform-releng-dev@eclipse.o
rg
cc
2004-11-10 10:31
Subject
Re: [platform-releng-dev]
Please respond to Build I-200411091436 is
platform-releng-dev unusable
I'm a little puzzled by this rebuild.
Initially yesterday's rebuild of 14:36 should have contained ONLY the fix
for the SWT bug that made the Tuesday's 08:00 build unusable.
However a diff of the map files shows that in complement of SWT, the doc,
the platform and the udpate components submitted new content increasing the
likelyhood of breakage whereas we are seeking stability.
I think that Sonia and Kim should have the power to veto unexpected
submissions (things that have not been raised as being major issues on the
mailing list) in case of an integration rebuild.
PaScaL
Sonia
Dimitrov/Ottawa/IBM@IBMCA
Sent by:
platform-releng-dev-admin@ecli To
pse.org platform-releng-dev@eclipse.o
rg
cc
11/10/2004 09:59 AM
Subject
Re: [platform-releng-dev]
Please respond to Build I-200411091436 is
platform-releng-dev unusable
The build will begin at 10 am ET.
Sonia
Dorian
Birsan/Toronto/IBM@IBMCA
Sent by:
platform-releng-dev-admin@ecli To
pse.org platform-releng-dev@eclipse.o
rg
cc
11/10/2004 09:54 AM
Subject
Re: [platform-releng-dev]
Please respond to Build I-200411091436 is
platform-releng-dev unusable
I have recently released a patch for cdcFoundation that serializes
platform.xml "manually", as opposed to using the standard Transformer class
from the J2SE. The comment written at the top of the file caused the xml
parser to choke.
The fix is released and ready for a rebuild.
-Dorian
Sonia
Dimitrov/Ottawa/IBM@IBMCA
Sent by:
platform-releng-dev-admin@ecli To
pse.org platform-releng-dev@eclipse.o
rg
cc
11/10/2004 09:01 AM
Subject
Re: [platform-releng-dev]
Please respond to Build I-200411091436 is
platform-releng-dev unusable
Please let us know when the fix for this is released, we will run a
rebuild.
Sonia
Wassim Melhem/Toronto/IBM@IBMCA
Sent by:
platform-releng-dev-admin@eclip To
se.org platform-releng-dev@eclipse.o
rg
cc
11/10/2004 02:38 AM
Subject
[platform-releng-dev] Build
Please respond to I-200411091436 is unusable
platform-releng-dev
Said build is unusable, as self-hosting is impossible.
I logged a defect against the Platform/Update component.
https://bugs.eclipse.org/bugs/show_bug.cgi?id=78262
Wassim.
_______________________________________________
platform-releng-dev mailing list
platform-releng-dev@xxxxxxxxxxx
http://dev.eclipse.org/mailman/listinfo/platform-releng-dev