2.12 update #428

ittaiz · 2018-02-20T20:49:21Z

this is just master with cherry-picking 08f3ca5 thanks to @hmemcpy who grouped together the 2.12 relevant changes to ease this branch update

* jmh normalizes jar timestamps build_is_identical e2e test spits out differences in md5 * change jar creator to default to normalizing timestamps

build_is_identical stabilization

) * Output statsfile with line 0 = compilation time * Actually specify that the statsfile is an output :) * Use Java nio utilities to write statsfile; adjust to property file inspired syntax * Add a test for the statsfile output

…lbuild#409) When executing non-locally, the wrong version of Java would be used, referencing a path that didn't exist. This change adds the _java_runtime attribute to have to locate the java runtime on the machine where the Scala code is being executed, and selects the correct configuration for the host_javabase. As a minor cleanup, this also removes the now no-longer-necessary _java attribute. TESTED=Ran unit tests and also ran a custom Scala unit test through: previously, the Scala unit test failed with an incorrect java path error, now it passes. Change-Id: I9c2829bded04ec2d9d014094bd9e72cd984fe6e6

Resolves: bazelbuild#412

The IntelliJ plugin assumes that source_jar and source_jars are both present on the jars struct. This sets those to empty values, so the IntelliJ plugin does not fail.

Without it, builds fail on Windows as it uses semi-colon instead.

…with specs2 v4

* use transitive_proto_path_flags if exist * Update scala_proto.bzl

googlebot · 2018-02-20T20:49:28Z

So there's good news and bad news.

👍 The good news is that everyone that needs to sign a CLA (the pull request submitter and all commit authors) have done so. Everything is all good there.

😕 The bad news is that it appears that one or more commits were authored by someone other than the pull request submitter. We need to confirm that all authors are ok with their commits being contributed to this project. Please have them confirm that here in the pull request.

Note to project maintainer: This is a terminal state, meaning the cla/google commit status will not change from this State. It's up to you to confirm consent of the commit author(s) and merge this pull request when appropriate.

ittaiz · 2018-02-20T21:17:28Z

note re cla- I verified it's all approved authors since it's just master with manually taken changes by @hmemcpy

johnynek · 2018-02-21T03:15:31Z

👍

ittaiz · 2018-02-21T04:05:12Z

I want to keep the history close to master and not squash the commits but this means force pushing into 2.12.
@johnynek wdyt?

johnynek · 2018-02-21T06:15:37Z

I’m confused about this diff since it seems to have diffs that are already in 2.12. For instance I don’t know why the hash change on the jars since that was already updated.

Anyway, I view 2.12 branch as an alpha type branch.

That said, if you force push, I worry if that could break users using Shas now on the branch. I don’t actually know what happens to the shas. I guess they will still be accessible.

ittaiz · 2018-02-21T06:37:48Z

I'm indeed not sure why they are showing up in the diff.
I'm also not sure what will happen to the sha but from a small experimentation I did on a different repo it seems they are still accessible (although I thought they wouldn't be).
I think we can maintain this in two ways:

Merge master into this branch constantly- the pro is that this is the git way to do this but the con is that it creates unclear history (to me) and we already saw regressions sneak in.
Cherry-pick the 2.12 commit on top of master each time and force push here- less clean and unclear on why we're seeing that diff.
Cherry-pick the 2.12 commit on top of master each time and push to a new branch each time master_$date_2.12- not the git way but solves the above issues.

WDYT?
Internally we use option 3 for rules_scala HEAD because we also cherry-pick a few other things we have yet been able to iron out

johnynek · 2018-02-21T06:59:23Z

As long as we don’t break anyone, I’m good. I would say I would prefer the 1. You mention where we merge in diffs from master. That said I hope we can get to making the versions selectable in master soon. Sadly I have been very busy with a project and having a hard time finding time for anything not in the critical path of that.

On Tue, Feb 20, 2018 at 20:37 Ittai Zeidman ***@***.***> wrote: I'm indeed not sure why they are showing up in the diff. I'm also not sure what will happen to the sha but from a small experimentation I did on a different repo it seems they are still accessible (although I thought they wouldn't be). I think we can maintain this in two ways: 1. Merge master into this branch constantly- the pro is that this is the git way to do this but the con is that it creates unclear history (to me) and we already saw regressions sneak in. 2. Cherry-pick the 2.12 commit on top of master each time and force push here- less clean and unclear on why we're seeing that diff. 3. Cherry-pick the 2.12 commit on top of master each time and push to a new branch each time master_$date_2.12- not the git way but solves the above issues. WDYT? Internally we use option 3 for rules_scala HEAD because we also cherry-pick a few other things we have yet been able to iron out — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#428 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAEJdsGuj2R5Zc1v6GGEX-iqHFKDVfSHks5tW7nCgaJpZM4SMpG4> .

-- P. Oscar Boykin, Ph.D. | http://twitter.com/posco | http://pobox.com/~boykin

ittaiz · 2018-02-22T07:45:05Z

ok,
@anchlovi will start working on having it selectable on master next week using the plan we discussed in #399 and to lesser extent in #407
So I'm keeping this open but not merging it for the time being

ittaiz · 2018-03-09T08:21:54Z

Seems like superseded by #435

ittaiz and others added 12 commits January 18, 2018 09:47

jmh normalizes jar timestamps (bazelbuild#396)

375b34a

* jmh normalizes jar timestamps build_is_identical e2e test spits out differences in md5 * change jar creator to default to normalizing timestamps

fix e2e

77fd23d

Merge pull request bazelbuild#398 from ittaiz/fixup_tests

7d02925

build_is_identical stabilization

Update README scala version to latest HEAD (bazelbuild#414)

f257275

Resolves: bazelbuild#412

Add struct members that IntelliJ plugin expects to exist

7c5209d

The IntelliJ plugin assumes that source_jar and source_jars are both present on the jars struct. This sets those to empty values, so the IntelliJ plugin does not fail.

Put the statsfile output in the rule output struct

827124a

Use OS-specific classpath separator

24bc74b

Without it, builds fail on Windows as it uses semi-colon instead.

Adding an overload that is called dynamically, in case we’re running …

dbe55fb

…with specs2 v4

use transitive_proto_path_flags if exist (bazelbuild#425)

8c1f7d0

* use transitive_proto_path_flags if exist * Update scala_proto.bzl

Adding 2.12 commits

c856d38

googlebot added the cla: no label Feb 20, 2018

ittaiz closed this Mar 9, 2018

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

2.12 update #428

2.12 update #428

ittaiz commented Feb 20, 2018

googlebot commented Feb 20, 2018

ittaiz commented Feb 20, 2018

johnynek commented Feb 21, 2018

ittaiz commented Feb 21, 2018

johnynek commented Feb 21, 2018

ittaiz commented Feb 21, 2018

johnynek commented Feb 21, 2018 via email

ittaiz commented Feb 22, 2018

ittaiz commented Mar 9, 2018

2.12 update #428

2.12 update #428

Conversation

ittaiz commented Feb 20, 2018

googlebot commented Feb 20, 2018

ittaiz commented Feb 20, 2018

johnynek commented Feb 21, 2018

ittaiz commented Feb 21, 2018

johnynek commented Feb 21, 2018

ittaiz commented Feb 21, 2018

johnynek commented Feb 21, 2018 via email

ittaiz commented Feb 22, 2018

ittaiz commented Mar 9, 2018