Invalidation issue: VanillaJavaBuilder keeps stale files. #2941

damienmg · 2017-05-04T11:19:59Z

For JDK7 linux:
http://ci.bazel.io/job/Bazel/JAVA_VERSION=1.7,PLATFORM_NAME=ubuntu_16.04-x86_64/1467/display/redirect?page=changes
and windows:
http://ci.bazel.io/job/Bazel/JAVA_VERSION=1.8,PLATFORM_NAME=windows-x86_64/1467/console

damienmg · 2017-05-04T11:21:42Z

The windows failure seems like a permission error, infra flake?

JDK7 failure is relevant to autovalue apparently:

ERROR: /home/ci/workspace/Bazel/JAVA_VERSION/1.7/PLATFORM_NAME/linux-x86_64/src/main/java/com/google/devtools/build/lib/BUILD:543:1: Java compilation in rule '//src/main/java/com/google/devtools/build/lib:build-base' failed: java failed: error executing command external/local_jdk/bin/java -Xbootclasspath/p:external/bazel_tools/third_party/java/jdk/langtools/javac7.jar -XX:+TieredCompilation '-XX:TieredStopAtLevel=1' -jar ... (remaining 2 argument(s) skipped): com.google.devtools.build.lib.shell.BadExitStatusException: Process exited with status 1.
src/main/java/com/google/devtools/build/lib/skyframe/ArtifactSkyKey.java:105: error: interface expected here
src/main/java/com/google/devtools/build/lib/analysis/RuleDefinition.java:83: error: package AutoValue_RuleDefinition_Metadata does not exist
src/main/java/com/google/devtools/build/lib/analysis/actions/ExecutionRequirements.java:77: error: cannot find symbol
  symbol:   class AutoValue_ExecutionRequirements_ParseableRequirement
  location: class com.google.devtools.build.lib.analysis.actions.ExecutionRequirements.ParseableRequirement
src/main/java/com/google/devtools/build/lib/analysis/constraints/EnvironmentCollection.java:44: error: cannot find symbol
  symbol:   class AutoValue_EnvironmentCollection_EnvironmentWithGroup
  location: class com.google.devtools.build.lib.analysis.constraints.EnvironmentCollection.EnvironmentWithGroup
src/main/java/com/google/devtools/build/lib/skyframe/ArtifactSkyKey.java:48: error: incompatible types: com.google.devtools.build.lib.skyframe.ArtifactSkyKey.OwnedArtifact cannot be converted to com.google.devtools.build.skyframe.SkyKey
src/main/java/com/google/devtools/build/lib/skyframe/ArtifactSkyKey.java:136: error: method does not override or implement a method from a supertype
src/main/java/com/google/devtools/build/lib/skyframe/ArtifactSkyKey.java:141: error: method does not override or implement a method from a supertype
src/main/java/com/google/devtools/build/lib/skyframe/TargetCompletionValue.java:68: error: cannot find symbol
  symbol:   class AutoValue_TargetCompletionValue_TargetCompletionKey
  location: class com.google.devtools.build.lib.skyframe.TargetCompletionValue.TargetCompletionKey
src/main/java/com/google/devtools/build/lib/skyframe/TestCompletionValue.java:68: error: cannot find symbol
  symbol:   class AutoValue_TestCompletionValue_TestCompletionKey
  location: class com.google.devtools.build.lib.skyframe.TestCompletionValue.TestCompletionKey
src/main/java/com/google/devtools/build/lib/skyframe/AspectCompletionValue.java:58: error: cannot find symbol
  symbol:   class AutoValue_AspectCompletionValue_AspectCompletionKey
  location: class com.google.devtools.build.lib.skyframe.AspectCompletionValue.AspectCompletionKey
src/main/java/com/google/devtools/build/lib/skyframe/CollectTargetsInPackageValue.java:44: error: cannot find symbol
  symbol:   class AutoValue_CollectTargetsInPackageValue_CollectTargetsInPackageKey
  location: class com.google.devtools.build.lib.skyframe.CollectTargetsInPackageValue.CollectTargetsInPackageKey
src/main/java/com/google/devtools/build/lib/skyframe/TransitiveTraversalValue.java:146: error: incompatible types: com.google.devtools.build.lib.cmdline.Label cannot be converted to com.google.devtools.build.skyframe.SkyKey
src/main/java/com/google/devtools/build/lib/analysis/config/BuildConfiguration.java: note: Some input files use or override a deprecated API.
src/main/java/com/google/devtools/build/lib/analysis/config/BuildConfiguration.java: note: Recompile with -Xlint:deprecation for details.
src/main/java/com/google/devtools/build/lib/analysis/config/BuildConfiguration.java: note: Some input files use unchecked or unsafe operations.
src/main/java/com/google/devtools/build/lib/analysis/config/BuildConfiguration.java: note: Recompile with -Xlint:unchecked for details.

damienmg · 2017-05-04T11:44:38Z

JDK7 is due to bfdad90

laszlocsomor · 2017-05-04T13:31:45Z

@janakdr is taking a look

damienmg · 2017-05-04T13:48:43Z

Still keeping it hot because I haven't got time to look at the windows failures.

janakdr · 2017-05-04T13:54:48Z

Looks like this is a spurious breakage. Running "bazel clean" before the build makes it go away. I think we're consuming stale jars.

janakdr · 2017-05-04T13:55:10Z

This is a moral dup of #2936

dslomov · 2017-05-04T13:58:34Z

I am looking at Windows issue

…

On Thu, May 4, 2017 at 1:20 PM, Damien Martin-Guillerez < ***@***.***> wrote: For JDK7 linux: http://ci.bazel.io/job/Bazel/JAVA_VERSION=1.7,PLATFORM_ NAME=ubuntu_16.04-x86_64/1467/display/redirect?page=changes and windows: http://ci.bazel.io/job/Bazel/JAVA_VERSION=1.8,PLATFORM_ NAME=windows-x86_64/1467/console — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#2941>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AExrX2XxPAzaFDLw_OXZFmhTUCfgrHSJks5r2bRjgaJpZM4NQiRD> .

-- Google Germany GmbH *Erika-Mann-Straße 33, 80636 München, Germany*

damienmg · 2017-05-04T14:08:03Z

Reopening, we should figure out what is wrong both for windows and the correctness issue (figure out if this is an actual dup of #2936 ). I have a consistent repro.

dslomov · 2017-05-04T14:10:29Z

I can repro this locally on any incremental Java build with bazel-0.4.5-windows-x86_64.exe (that's out bootstrap Bazel, right?)

…

On Thu, May 4, 2017 at 4:08 PM, Damien Martin-Guillerez < ***@***.***> wrote: Reopening, we should figure out what is wrong both for windows and the correctness issue (figure out if this is an actual dup of #2936 <#2936> ). I have a consistent repro. — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#2941 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AExrX5pRyB5G8xv7ZIMZJvWLKIcX630jks5r2dvJgaJpZM4NQiRD> .

-- Google Germany GmbH *Erika-Mann-Straße 33, 80636 München, Germany*

damienmg · 2017-05-04T14:12:26Z

We bootstrap with bazel so yes On Thu, May 4, 2017 at 4:10 PM Dmitry Lomov <notifications@github.com> wrote:

…

I can repro this locally on any incremental Java build with bazel-0.4.5-windows-x86_64.exe (that's out bootstrap Bazel, right?) On Thu, May 4, 2017 at 4:08 PM, Damien Martin-Guillerez < ***@***.***> wrote: > Reopening, we should figure out what is wrong both for windows and the > correctness issue (figure out if this is an actual dup of #2936 > <#2936> ). I have a consistent > repro. > > — > You are receiving this because you commented. > Reply to this email directly, view it on GitHub > <#2941 (comment) >, > or mute the thread > < https://github.com/notifications/unsubscribe-auth/AExrX5pRyB5G8xv7ZIMZJvWLKIcX630jks5r2dvJgaJpZM4NQiRD > > . > -- Google Germany GmbH *Erika-Mann-Straße 33, 80636 München, Germany* — You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub <#2941 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ADjHf522UVWBnFQFMLsYw8FGK944tgkIks5r2dxcgaJpZM4NQiRD> .

damienmg · 2017-05-04T14:15:06Z

Also the linux bug is probably not a dupe of #2936 because it happens only with jdk7 bazel

dslomov · 2017-05-04T14:15:18Z

--strategy=Javac=standalone fixes this for 0.4.5. Now trying the latest Bazel On Thu, May 4, 2017 at 4:12 PM, Damien Martin-Guillerez < notifications@github.com> wrote:

…

We bootstrap with bazel so yes On Thu, May 4, 2017 at 4:10 PM Dmitry Lomov ***@***.***> wrote: > I can repro this locally on any incremental Java build > with bazel-0.4.5-windows-x86_64.exe (that's out bootstrap Bazel, right?) > > On Thu, May 4, 2017 at 4:08 PM, Damien Martin-Guillerez < > ***@***.***> wrote: > > > Reopening, we should figure out what is wrong both for windows and the > > correctness issue (figure out if this is an actual dup of #2936 > > <#2936> ). I have a consistent > > repro. > > > > — > > You are receiving this because you commented. > > Reply to this email directly, view it on GitHub > > <#2941# issuecomment-299195594 > >, > > or mute the thread > > < > https://github.com/notifications/unsubscribe-auth/ AExrX5pRyB5G8xv7ZIMZJvWLKIcX630jks5r2dvJgaJpZM4NQiRD > > > > . > > > > > > -- > Google Germany GmbH > *Erika-Mann-Straße 33, 80636 München, Germany* > > — > You are receiving this because you modified the open/close state. > > > Reply to this email directly, view it on GitHub > <#2941 (comment) >, > or mute the thread > <https://github.com/notifications/unsubscribe-auth/ ADjHf522UVWBnFQFMLsYw8FGK944tgkIks5r2dxcgaJpZM4NQiRD> > . > — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#2941 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AExrXwup7jD8yPaWI-EfBfV2VLk7GKz0ks5r2dzWgaJpZM4NQiRD> .

-- Google Germany GmbH *Erika-Mann-Straße 33, 80636 München, Germany*

janakdr · 2017-05-04T14:23:35Z

Simplest repro I have so far:

for br in c5672f8e bfdad902c ; do
git checkout $br
bazel clean
bazel build  --define JAVA_VERSION=1.7 //src/main/java/com/google/devtools/build/skyframe:skyframe
echo "For branch $br:"
jar tf bazel-bin/src/main/java/com/google/devtools/build/skyframe/libskyframe.jar | grep SkyKey
done

Output:
For branch c5672f8
com/google/devtools/build/skyframe/SkyKey$1.class
com/google/devtools/build/skyframe/SkyKey.class
For branch bfdad90

However, if you don't clean in between, you get:
For branch c5672f8
com/google/devtools/build/skyframe/SkyKey$1.class
com/google/devtools/build/skyframe/SkyKey.class
For branch bfdad90
com/google/devtools/build/skyframe/SkyKey$1.class
com/google/devtools/build/skyframe/SkyKey.class

The skyframe jar is getting stale files in it.

damienmg · 2017-05-04T14:26:06Z

Ok that sounds similar to the other bug so far.

…

On Thu, May 4, 2017 at 4:23 PM Janak Ramakrishnan ***@***.***> wrote: Simplest repro I have so far: for br in c5672f8 bfdad90 ; do git checkout $br bazel clean bazel build --define JAVA_VERSION=1.7 //src/main/java/com/google/devtools/build/skyframe:skyframe echo "For branch $br:" jar tf bazel-bin/src/main/java/com/google/devtools/build/skyframe/libskyframe.jar | grep SkyKey done Output: For branch c5672f8 com/google/devtools/build/skyframe/SkyKey$1.class com/google/devtools/build/skyframe/SkyKey.class For branch bfdad90 However, if you don't clean in between, you get: For branch c5672f8 com/google/devtools/build/skyframe/SkyKey$1.class com/google/devtools/build/skyframe/SkyKey.class For branch bfdad90 com/google/devtools/build/skyframe/SkyKey$1.class com/google/devtools/build/skyframe/SkyKey.class The skyframe jar is getting stale files in it. — You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub <#2941 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ADjHf3NqRDz6XOR6hJ8OpJdJrIhmppfvks5r2d9ugaJpZM4NQiRD> .

janakdr · 2017-05-04T14:38:01Z

Looks like bazel-out/local-fastbuild/bin/src/main/java/com/google/devtools/build/skyframe/_javac/skyframe/libskyframe_classes has the stale SkyKey.class file.

janakdr · 2017-05-04T14:43:21Z

Is this just as silly as, javac generates .class files into the specified --classdir, and then blindly bundles everything in --classdir into a jar? So if there are any extra files in --classdir, it'll read them too?

janakdr · 2017-05-04T14:46:33Z

Yup:

echo "hello" > bazel-out/local-fastbuild/bin/src/main/java/com/google/devtools/build/skyframe/_javac/skyframe/libskyframe_classes/com/google/devtools/build/skyframe/JanakImaginaryClass.class

(cd /private/var/tmp/_bazel_janakr/449f66540f7a827a317d06e34f37f945/execroot/bazel &&   exec env -     LC_CTYPE=en_US.UTF-8   external/local_jdk/bin/java -Xbootclasspath/p:external/bazel_tools/third_party/java/jdk/langtools/javac7.jar -XX:+TieredCompilation '-XX:TieredStopAtLevel=1' -jar external/bazel_tools/tools/jdk/JavaBuilder_deploy.jar @bazel-out/local-fastbuild/bin/src/main/java/com/google/devtools/build/skyframe/libskyframe.jar-2.params)

jar tf bazel-bin/src/main/java/com/google/devtools/build/skyframe/libskyframe.jar | grep Janak
com/google/devtools/build/skyframe/JanakImaginaryClass.class

janakdr · 2017-05-04T14:51:53Z

My guess is that https://github.com/bazelbuild/bazel/blob/master/src/java_tools/buildjar/java/com/google/devtools/build/buildjar/VanillaJavaBuilder.java#L257 should not just create the directory, it should delete any existing directory.

cushon · 2017-05-04T14:56:25Z

Is the action not sandboxed?

damienmg · 2017-05-04T14:59:02Z

It should be but it is also using workers no?

…

On Thu, May 4, 2017 at 4:56 PM Liam Miller-Cushon ***@***.***> wrote: Is the action not sandboxed? — You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub <#2941 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ADjHfx6L-uLp1JwjrmCq4oUMgMx9SFEwks5r2ecfgaJpZM4NQiRD> .

damienmg · 2017-05-04T15:00:07Z

The ijar for libskyframe is the same even though a SkyKey change in between the two changes. On Thu, May 4, 2017 at 4:58 PM Damien Martin-Guillerez <dmarting@google.com> wrote:

…

It should be but it is also using workers no? On Thu, May 4, 2017 at 4:56 PM Liam Miller-Cushon < ***@***.***> wrote: > Is the action not sandboxed? > > — > You are receiving this because you modified the open/close state. > Reply to this email directly, view it on GitHub > <#2941 (comment)>, > or mute the thread > <https://github.com/notifications/unsubscribe-auth/ADjHfx6L-uLp1JwjrmCq4oUMgMx9SFEwks5r2ecfgaJpZM4NQiRD> > . >

dslomov · 2017-05-04T15:01:43Z

Ok this does not reproduce on latest Bazel

…

On Thu, May 4, 2017 at 4:15 PM, Dmitry Lomov ***@***.***> wrote: --strategy=Javac=standalone fixes this for 0.4.5. Now trying the latest Bazel On Thu, May 4, 2017 at 4:12 PM, Damien Martin-Guillerez < ***@***.***> wrote: > We bootstrap with bazel so yes > > On Thu, May 4, 2017 at 4:10 PM Dmitry Lomov ***@***.***> > wrote: > > > I can repro this locally on any incremental Java build > > with bazel-0.4.5-windows-x86_64.exe (that's out bootstrap Bazel, right?) > > > > On Thu, May 4, 2017 at 4:08 PM, Damien Martin-Guillerez < > > ***@***.***> wrote: > > > > > Reopening, we should figure out what is wrong both for windows and the > > > correctness issue (figure out if this is an actual dup of #2936 > > > <#2936> ). I have a > consistent > > > repro. > > > > > > — > > > You are receiving this because you commented. > > > Reply to this email directly, view it on GitHub > > > <#2941 (comment) > t-299195594 > > >, > > > or mute the thread > > > < > > https://github.com/notifications/unsubscribe-auth/AExrX5pRyB > 5G8xv7ZIMZJvWLKIcX630jks5r2dvJgaJpZM4NQiRD > > > > > > . > > > > > > > > > > > -- > > Google Germany GmbH > > *Erika-Mann-Straße 33, 80636 München, Germany* > > > > — > > You are receiving this because you modified the open/close state. > > > > > > Reply to this email directly, view it on GitHub > > <#2941 (comment) > >, > > or mute the thread > > <https://github.com/notifications/unsubscribe-auth/ADjHf522U > VWBnFQFMLsYw8FGK944tgkIks5r2dxcgaJpZM4NQiRD> > > . > > > > — > You are receiving this because you commented. > Reply to this email directly, view it on GitHub > <#2941 (comment)>, > or mute the thread > <https://github.com/notifications/unsubscribe-auth/AExrXwup7jD8yPaWI-EfBfV2VLk7GKz0ks5r2dzWgaJpZM4NQiRD> > . > -- Google Germany GmbH *Erika-Mann-Straße 33, 80636 München, Germany*

-- Google Germany GmbH *Erika-Mann-Straße 33, 80636 München, Germany*

janakdr · 2017-05-04T15:01:43Z

@damienmg I think the _classes directory is the root of the issue.

janakdr · 2017-05-04T15:04:50Z

I don't see any warning about not sandboxing the action, even when I run explicitly with --spawn_strategy=sandboxed. The bug is still there with --spawn_strategy=standalone, of course.

cushon · 2017-05-04T15:07:25Z

Got it. I'm on it, the fix is going to look a lot like

bazel/src/java_tools/buildjar/java/com/google/devtools/build/buildjar/SimpleJavaLibraryBuilder.java

Line 116 in 9547eeb

// Necessary for local builds in order to discard previous outputs

.

damienmg · 2017-05-04T15:07:42Z

@dslomov \o/

@janakdr ok but that won't explain why the ijar hasn't changed.

janakdr · 2017-05-04T15:09:10Z

@damienmg I think the ijar is blindly reading from the _classes directory.

…

On Thu, May 4, 2017 at 11:07 AM, Damien Martin-Guillerez < ***@***.***> wrote: @dslomov <https://github.com/dslomov> \o/ @janakdr <https://github.com/janakdr> ok but that won't explain why the ijar hasn't changed. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#2941 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AJqlcEJNhw8mfMk6uIA4nSX7KbFFBS7Uks5r2enEgaJpZM4NQiRD> .

damienmg · 2017-05-04T15:10:40Z

The ijar reads the jar but the SkyKey.class should be different, no? On Thu, May 4, 2017 at 5:09 PM Janak Ramakrishnan <notifications@github.com> wrote:

…

@damienmg I think the ijar is blindly reading from the _classes directory. On Thu, May 4, 2017 at 11:07 AM, Damien Martin-Guillerez < ***@***.***> wrote: > @dslomov <https://github.com/dslomov> \o/ > > @janakdr <https://github.com/janakdr> ok but that won't explain why the > ijar hasn't changed. > > — > You are receiving this because you were mentioned. > Reply to this email directly, view it on GitHub > <#2941 (comment) >, > or mute the thread > < https://github.com/notifications/unsubscribe-auth/AJqlcEJNhw8mfMk6uIA4nSX7KbFFBS7Uks5r2enEgaJpZM4NQiRD > > . > — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#2941 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ADjHf2WXRjrl2uFlzuOKSfD5xuw4Btokks5r2eocgaJpZM4NQiRD> .

janakdr · 2017-05-04T15:17:05Z

No: the SkyKey class was moved to the skyframe-objects target, so the SkyKey.class file left behind here is stale. On Thu, May 4, 2017 at 11:10 AM, Damien Martin-Guillerez < notifications@github.com> wrote:

…

The ijar reads the jar but the SkyKey.class should be different, no? On Thu, May 4, 2017 at 5:09 PM Janak Ramakrishnan < ***@***.***> wrote: > @damienmg I think the ijar is blindly reading from the _classes directory. > > On Thu, May 4, 2017 at 11:07 AM, Damien Martin-Guillerez < > ***@***.***> wrote: > > > @dslomov <https://github.com/dslomov> \o/ > > > > @janakdr <https://github.com/janakdr> ok but that won't explain why the > > ijar hasn't changed. > > > > — > > You are receiving this because you were mentioned. > > Reply to this email directly, view it on GitHub > > <#2941# issuecomment-299213773 > >, > > or mute the thread > > < > https://github.com/notifications/unsubscribe-auth/ AJqlcEJNhw8mfMk6uIA4nSX7KbFFBS7Uks5r2enEgaJpZM4NQiRD > > > > . > > > > — > You are receiving this because you were mentioned. > Reply to this email directly, view it on GitHub > <#2941 (comment) >, > or mute the thread > <https://github.com/notifications/unsubscribe-auth/ ADjHf2WXRjrl2uFlzuOKSfD5xuw4Btokks5r2eocgaJpZM4NQiRD> > . > — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#2941 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AJqlcBwjJtn9KJyUaRTWkvGnf_ndMYDwks5r2ep6gaJpZM4NQiRD> .

janakdr · 2017-05-04T15:27:43Z

@dslomov are you sure that this works at HEAD with JDK7? There appears to be a discrepancy between VanillaJavaBuilder and BazelJavaBuilder, and I don't think it's gone at HEAD.

damienmg · 2017-05-04T15:35:04Z

Oh ok, @dslomov was speaking about the windows breakage.

…

On Thu, May 4, 2017 at 5:27 PM Janak Ramakrishnan ***@***.***> wrote: @dslomov <https://github.com/dslomov> are you sure that this works at HEAD with JDK7? There appears to be a discrepancy between VanillaJavaBuilder and BazelJavaBuilder, and I don't think it's gone at HEAD. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#2941 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/ADjHf3phQl1p1oEOf1SgLltoQV0A6FRVks5r2e50gaJpZM4NQiRD> .

damienmg · 2017-05-04T15:35:43Z

@cushon IIUC you are working on a fix so assigning to you

cushon · 2017-05-04T15:37:38Z

Yep, internal review cl/155081626.

to ensure outputs from any previous local builds are discarded. To cherry-pick for #2692. Fixes #2941 PiperOrigin-RevId: 155089391

This is a temporary workaround to avoid random failures of JDK7 builds until 0.5.0 is released. Change-Id: I1323c128978190274f65018f32fd6b9afe8e681f

janakdr · 2017-05-05T21:34:45Z

Should we make an announcement about this? People should know that they have to run clean before any JDK7 build if they want guaranteed correctness, no? In the case of the commit that triggered this issue, I changed a class to an interface, so the bug resulted in a compile-time error, but I could easily see it not being caught in other cases, and leading to incorrect runtime behavior.

to ensure outputs from any previous local builds are discarded. To cherry-pick for #2692. Fixes #2941 PiperOrigin-RevId: 155089391

This is a temporary workaround to avoid random failures of JDK7 builds until 0.5.0 is released. Change-Id: I1323c128978190274f65018f32fd6b9afe8e681f

damienmg added breakage P0 This is an emergency and more important than other current work. (Assignee required) labels May 4, 2017

damienmg self-assigned this May 4, 2017

damienmg assigned ulfjack and unassigned ulfjack May 4, 2017

damienmg assigned janakdr May 4, 2017

janakdr closed this as completed May 4, 2017

damienmg reopened this May 4, 2017

damienmg added the Release blocker label May 4, 2017

janakdr assigned cushon May 4, 2017

damienmg unassigned damienmg and janakdr May 4, 2017

damienmg changed the title ~~Bootstrap broken at HEAD~~ Invalidation issue: VanillaJavaBuilder keeps stale files. May 4, 2017

bazel-io closed this as completed in d953ca8 May 4, 2017

hlopko pushed a commit that referenced this issue May 5, 2017

Clean VanillaJavaBuilder output directories

a083dfa

to ensure outputs from any previous local builds are discarded. To cherry-pick for #2692. Fixes #2941 PiperOrigin-RevId: 155089391

hlopko pushed a commit that referenced this issue May 10, 2017

Clean VanillaJavaBuilder output directories

d5908dd

to ensure outputs from any previous local builds are discarded. To cherry-pick for #2692. Fixes #2941 PiperOrigin-RevId: 155089391

hlopko pushed a commit that referenced this issue May 10, 2017

Clean VanillaJavaBuilder output directories

e6f7a7a

to ensure outputs from any previous local builds are discarded. To cherry-pick for #2692. Fixes #2941 PiperOrigin-RevId: 155089391

hlopko pushed a commit that referenced this issue May 10, 2017

Clean VanillaJavaBuilder output directories

d720623

to ensure outputs from any previous local builds are discarded. To cherry-pick for #2692. Fixes #2941 PiperOrigin-RevId: 155089391

hlopko pushed a commit that referenced this issue May 11, 2017

Clean VanillaJavaBuilder output directories

1154fbb

to ensure outputs from any previous local builds are discarded. To cherry-pick for #2692. Fixes #2941 PiperOrigin-RevId: 155089391

hlopko pushed a commit that referenced this issue May 17, 2017

Clean VanillaJavaBuilder output directories

177079e

to ensure outputs from any previous local builds are discarded. To cherry-pick for #2692. Fixes #2941 PiperOrigin-RevId: 155089391

hlopko pushed a commit that referenced this issue May 17, 2017

Clean VanillaJavaBuilder output directories

c9d1d6c

to ensure outputs from any previous local builds are discarded. To cherry-pick for #2692. Fixes #2941 PiperOrigin-RevId: 155089391

hlopko pushed a commit that referenced this issue May 23, 2017

Clean VanillaJavaBuilder output directories

0417c1a

to ensure outputs from any previous local builds are discarded. To cherry-pick for #2692. Fixes #2941 PiperOrigin-RevId: 155089391

damienmg pushed a commit that referenced this issue May 26, 2017

Clean VanillaJavaBuilder output directories

8e0906c

to ensure outputs from any previous local builds are discarded. To cherry-pick for #2692. Fixes #2941 PiperOrigin-RevId: 155089391

hlopko pushed a commit that referenced this issue May 31, 2017

Clean VanillaJavaBuilder output directories

c708592

to ensure outputs from any previous local builds are discarded. To cherry-pick for #2692. Fixes #2941 PiperOrigin-RevId: 155089391

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Invalidation issue: VanillaJavaBuilder keeps stale files. #2941

Invalidation issue: VanillaJavaBuilder keeps stale files. #2941

damienmg commented May 4, 2017

damienmg commented May 4, 2017

damienmg commented May 4, 2017

laszlocsomor commented May 4, 2017

damienmg commented May 4, 2017

janakdr commented May 4, 2017

janakdr commented May 4, 2017

dslomov commented May 4, 2017 via email

damienmg commented May 4, 2017

dslomov commented May 4, 2017 via email

damienmg commented May 4, 2017 via email

damienmg commented May 4, 2017

dslomov commented May 4, 2017 via email

janakdr commented May 4, 2017 •

edited

Loading

damienmg commented May 4, 2017 via email

janakdr commented May 4, 2017

janakdr commented May 4, 2017

janakdr commented May 4, 2017

janakdr commented May 4, 2017

cushon commented May 4, 2017

damienmg commented May 4, 2017 via email

damienmg commented May 4, 2017 via email

dslomov commented May 4, 2017 via email

janakdr commented May 4, 2017

janakdr commented May 4, 2017

cushon commented May 4, 2017

damienmg commented May 4, 2017

janakdr commented May 4, 2017 via email

damienmg commented May 4, 2017 via email

janakdr commented May 4, 2017 via email

janakdr commented May 4, 2017

damienmg commented May 4, 2017 via email

damienmg commented May 4, 2017

cushon commented May 4, 2017

janakdr commented May 5, 2017

Invalidation issue: VanillaJavaBuilder keeps stale files. #2941

Invalidation issue: VanillaJavaBuilder keeps stale files. #2941

Comments

damienmg commented May 4, 2017

damienmg commented May 4, 2017

damienmg commented May 4, 2017

laszlocsomor commented May 4, 2017

damienmg commented May 4, 2017

janakdr commented May 4, 2017

janakdr commented May 4, 2017

dslomov commented May 4, 2017 via email

damienmg commented May 4, 2017

dslomov commented May 4, 2017 via email

damienmg commented May 4, 2017 via email

damienmg commented May 4, 2017

dslomov commented May 4, 2017 via email

janakdr commented May 4, 2017 • edited Loading

damienmg commented May 4, 2017 via email

janakdr commented May 4, 2017

janakdr commented May 4, 2017

janakdr commented May 4, 2017

janakdr commented May 4, 2017

cushon commented May 4, 2017

damienmg commented May 4, 2017 via email

damienmg commented May 4, 2017 via email

dslomov commented May 4, 2017 via email

janakdr commented May 4, 2017

janakdr commented May 4, 2017

cushon commented May 4, 2017

damienmg commented May 4, 2017

janakdr commented May 4, 2017 via email

damienmg commented May 4, 2017 via email

janakdr commented May 4, 2017 via email

janakdr commented May 4, 2017

damienmg commented May 4, 2017 via email

damienmg commented May 4, 2017

cushon commented May 4, 2017

janakdr commented May 5, 2017

janakdr commented May 4, 2017 •

edited

Loading