Bazel: Release - November 2017 - Target RC date: 2017-11-06 - name: 0.8.0

Created on 20 Sep 2017  Â·  40Comments  Â·  Source: bazelbuild/bazel

Release tracking bug for november 2017:

  • Target cut: November, 6th -- during the Bazel Conference

  • Name: 0.8.0

/cc @dslomov. I should be the RM for this one but I would gladly exchange with yours in December seeing that I will be at the repro build summit.

release

Most helpful comment

rc5 promoted to release

All 40 comments

I can be an RM for this one. How about we cut on Mon Nov. 6 (I can cut while at the summit)

I plan to cut at cff0dc94f6a8e16492adf54c88d0b26abe903d4c
Nightly looks good: https://ci.bazel.io/view/Bazel%20bootstrap%20and%20maintenance/job/bazel/job/nightly/109/

Sounds good I'll let you know if the one remaining issue is genuine

On Sat, Nov 4, 2017, 9:13 AM Dmitry Lomov notifications@github.com wrote:

I plan to cut at cff0dc9
https://github.com/bazelbuild/bazel/commit/cff0dc94f6a8e16492adf54c88d0b26abe903d4c
Nightly looks good:
https://ci.bazel.io/view/Bazel%20bootstrap%20and%20maintenance/job/bazel/job/nightly/109/

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
https://github.com/bazelbuild/bazel/issues/3772#issuecomment-341879883,
or mute the thread
https://github.com/notifications/unsubscribe-auth/ADjHf3Jptre4yYWBggcAb4_Fuf2it_YDks5szByQgaJpZM4Pdv-n
.

We do not have CI coverage for it, which means you will have to check with
the release candidate and report regression. Cc @meteorcloudy.

Technically since this was not working on 0.7, this would not be a
regression. Testing those is hard however. Anyway if you can test the
proposed baseline and tell us if that works before we cut the RC that would
probably be the best.

On Sat, Nov 4, 2017, 2:59 PM David Ostrovsky notifications@github.com
wrote:

What about this #3965 https://github.com/bazelbuild/bazel/issues/3965?

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
https://github.com/bazelbuild/bazel/issues/3772#issuecomment-341899006,
or mute the thread
https://github.com/notifications/unsubscribe-auth/ADjHf0Ij19Ma1brEwjA5rsqgIhu4gm5hks5szG3DgaJpZM4Pdv-n
.

Dmitry, platforms documentation was just submitted internally. Can we get
it into the release?

On Sat, Nov 4, 2017, 1:35 PM Damien Martin-Guillerez <
[email protected]> wrote:

We do not have CI coverage for it, which means you will have to check with
the release candidate and report regression. Cc @meteorcloudy.

Technically since this was not working on 0.7, this would not be a
regression. Testing those is hard however. Anyway if you can test the
proposed baseline and tell us if that works before we cut the RC that would
probably be the best.

On Sat, Nov 4, 2017, 2:59 PM David Ostrovsky notifications@github.com
wrote:

What about this #3965 https://github.com/bazelbuild/bazel/issues/3965?

—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
<https://github.com/bazelbuild/bazel/issues/3772#issuecomment-341899006
,
or mute the thread
<
https://github.com/notifications/unsubscribe-auth/ADjHf0Ij19Ma1brEwjA5rsqgIhu4gm5hks5szG3DgaJpZM4Pdv-n

.

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub
https://github.com/bazelbuild/bazel/issues/3772#issuecomment-341915472,
or mute the thread
https://github.com/notifications/unsubscribe-auth/AAT7f7jSBOjLHkQLM-kZgHdd8VemzdI9ks5szKBwgaJpZM4Pdv-n
.

@damienmg @meteorcloudy

Trying to build gerrit master with bazel master is failing for me on Windows now, with:

$ bazel build --workspace_status_command=tools/workspace-status.cmd headless
Loading:
Loading: 0 packages loaded
INFO: Analysed target //:headless (0 packages loaded).
INFO: Found 1 target...
[0 / 14] BazelWorkspaceStatusAction stable-status.txt
ERROR: C:/users/davido/projects/gerrit/java/gerrit/BUILD:1:1: Building java/gerrit/libprolog-predicates.jar (19 source files) failed (Exit 1)
java\gerrit\PRED_commit_edits_2.java:99: error: incompatible types: ImmutableList cannot be converted to List<Edit>
          List<Edit> edits = entry.getEdits();
                                           ^
Target //:headless failed to build
Use --verbose_failures to see the command lines of failed build steps.

This happens here on most recent Java 8 version available on Windows 64bit:

$ "$JAVA_HOME/bin/javac" -version
javac 1.8.0_152

This looks like this JDK bug and it seems that this diff in gerrit fixed it:

$ bazel build --workspace_status_command=tools/workspace-status.cmd headless
Loading:
Loading: 0 packages loaded
INFO: Analysed target //:headless (0 packages loaded).
INFO: Found 1 target...
[0 / 9] BazelWorkspaceStatusAction stable-status.txt
Target //:headless up-to-date:
  C:/msys64/tmp/_bazel_davido/_biltsqa/execroot/gerrit/bazel-out/x64_windows-fastbuild/bin/headless.war
INFO: Elapsed time: 1,868s, Critical Path: 0,76s
INFO: Build completed successfully, 3 total actions

So do you mean the problem is on gerrit side?

On Sun, Nov 5, 2017, 10:39 PM David Ostrovsky notifications@github.com
wrote:

@damienmg https://github.com/damienmg @meteorcloudy
https://github.com/meteorcloudy

Trying to build gerrit master with bazel master is failing for me on
Windows now, with:

$ bazel build --workspace_status_command=tools/workspace-status.cmd headless
Loading:
Loading: 0 packages loaded
INFO: Analysed target //:headless (0 packages loaded).
INFO: Found 1 target...
[0 / 14] BazelWorkspaceStatusAction stable-status.txt
ERROR: C:/users/davido/projects/gerrit/java/gerrit/BUILD:1:1: Building java/gerrit/libprolog-predicates.jar (19 source files) failed (Exit 1)
javagerritPRED_commit_edits_2.java:99: error: incompatible types: ImmutableList cannot be converted to List
List edits = entry.getEdits();
^
Target //:headless failed to build
Use --verbose_failures to see the command lines of failed build steps.

This happens here on most recent Java 8 version:

$ "$JAVA_HOME/bin/javac" -version
javac 1.8.0_152

It seems that ths diff fixed it http://paste.openstack.org/show/625563/:

$ bazel build --workspace_status_command=tools/workspace-status.cmd headless
Loading:
Loading: 0 packages loaded
INFO: Analysed target //:headless (0 packages loaded).
INFO: Found 1 target...
[0 / 9] BazelWorkspaceStatusAction stable-status.txt
Target //:headless up-to-date:
C:/msys64/tmp/_bazel_davido/_biltsqa/execroot/gerrit/bazel-out/x64_windows-fastbuild/bin/headless.war
INFO: Elapsed time: 1,868s, Critical Path: 0,76s
INFO: Build completed successfully, 3 total actions

—
You are receiving this because you were mentioned.

Reply to this email directly, view it on GitHub
https://github.com/bazelbuild/bazel/issues/3772#issuecomment-342007928,
or mute the thread
https://github.com/notifications/unsubscribe-auth/ADjHf85EW_9nYL2JV06K2_MKqKePToFuks5szir9gaJpZM4Pdv-n
.

I believe, this is a JDK bug. A workaround can be provided in gerrit. I also verified, that the same problem exists in Bazel 0.7 already (so this is definitely not a regression in upcoming 0.8 release). It's not clear to me why this problem does not exist on sane platforms (Linux and Mac Os X).

I cut with devtools/blaze/bazel/admin/release.sh create 0.8.0 cff0dc9

please cherry pick internal change 174503874 which has been submitted but not exported yet, its a release blocker that would break Android AAR support (internal bug 68858732) /cc @jin

rc2: devtools/blaze/bazel/admin/release.sh create 0.8.0 cff0dc9 8a49b15

We need to cherry-pick e28d3af92227cd60287d27d9efa7593ae3e0509f for fixing https://github.com/bazelbuild/bazel/issues/3927

e28d3af is already in the release, no need to cherrypick

Good to know, thanks!

Please cherry-pick 275ae45b1228bdd0f912c4fbd634b29ba4180383 to fix #3956

275ae45 fails to apply cleanly, and after I fix merge conflicts, scripts/release.sh fails to detect that it needs to create rc3 (it creates rc1 instead). I will not be able to investigate until Friday.

Got it, scripts/release/release.sh relies on local state

rc3: scripts/release/release.sh create --force_rc=3 0.8.0 cff0dc9 8a49b15 275ae45

recreated rc3 with the same command since I botched conflict resolution

3rd time's the charm

Great, thanks! I don't see #3956 happen again in 0.8.0rc3.

Will need one more cherry-pick: fix for https://github.com/bazelbuild/bazel/issues/4035

We also have a cc_proto_library regression in 0.7.0 on Windows https://github.com/bazelbuild/bazel/issues/3985
Sent https://bazel-review.googlesource.com/#/c/bazel/+/21790/ to fix it.

ok, waiting for both of those to land before cutting rc4

One (future) cherry-pick has landed (fix for #4035) https://github.com/bazelbuild/bazel/commit/d0bf589f2716b3d139c210930371a684c6e158eb
Waiting on a fix for #3985

rc4: scripts/release/release.sh create 0.8.0 cff0dc9 8a49b15 275ae45 d0bf589 9738f35

We need to cherry-pick cl/176654245

rc5: scripts/release/release.sh create --force_rc=5 0.8.0 cff0dc9 8a49b15 275ae45 d0bf589 9738f35 0d6ff47

I think that this release should be blocked, untill #4108 is fixed.

Not a regression so no.

On Thu, Nov 23, 2017, 7:55 AM David Ostrovsky notifications@github.com
wrote:

Should this release be blocked, untill #4108
https://github.com/bazelbuild/bazel/issues/4108 is fixed?

—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
https://github.com/bazelbuild/bazel/issues/3772#issuecomment-346542210,
or mute the thread
https://github.com/notifications/unsubscribe-auth/ADjHf5Bv1bTA8dcraLh0Hq4EoR52RhkTks5s5RbJgaJpZM4Pdv-n
.

rc5 promoted to release

Unfortunately, I think https://github.com/bazelbuild/bazel/issues/4173 is a regression.

Bazel 0.8.0 native package for Alpine Linux is here: https://github.com/davido/bazel-alpine-package/releases/tag/0.8.0.

Was this page helpful?
0 / 5 - 0 ratings