Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Release 0.5.2 #3086

Closed
hlopko opened this issue May 31, 2017 · 39 comments
Closed

Release 0.5.2 #3086

hlopko opened this issue May 31, 2017 · 39 comments
Assignees
Labels
P1 I'll work on this now. (Assignee required) type: process

Comments

@hlopko
Copy link
Member

hlopko commented May 31, 2017

We need more releases :) 0.5.2 incoming.

a477805 has an error in the commit message. Instead of action_config 'generic' it should say 'cc-flags-make-variable'.

@hlopko hlopko added category: misc > release / binary P1 I'll work on this now. (Assignee required) type: process labels May 31, 2017
@hlopko hlopko assigned buchgr and hlopko and unassigned buchgr May 31, 2017
@hlopko
Copy link
Member Author

hlopko commented Jun 6, 2017

Make sure fix for #3064 is included. The SHA of the fix is 7d8943e .

@buchgr
Copy link
Contributor

buchgr commented Jun 6, 2017

Thanks so much @mhlopko

@petemounce
Copy link
Contributor

petemounce commented Jun 6, 2017

What is the overhead to doing releases that makes them less frequent? (I mean: it sounds like the process is largely automated end-to-end).

@hlopko
Copy link
Member Author

hlopko commented Jun 6, 2017

Right now, our release candidates get more thorough testing by teams using Bazel internally. It's in the process of being automated, but we're not there yet. But ideally yeah, we would release more often, and regularly.

@petemounce
Copy link
Contributor

@mhlopko say - what are the chances of getting an RC cut for 0.5.2, so that we can give early feedback, but also get past our blocker (#3064)?

@hlopko
Copy link
Member Author

hlopko commented Jun 8, 2017

@petemounce If we'll have a green CI, I think we can cut something tomorrow (or soonish). It's been a while since the last baseline anyway.

@ulfjack
Copy link
Contributor

ulfjack commented Jun 9, 2017

I just cut a canary on @mhlopko's behalf. Waiting for Jenkins to process it. (I couldn't wait for a green CI, maybe that was a mistake.)

@aehlig
Copy link
Contributor

aehlig commented Jun 9, 2017

6802831 fixes a deadlock in --experimental_ui; can we cherry-pick it to the next release candidate? Thanks.

@ulfjack
Copy link
Contributor

ulfjack commented Jun 9, 2017

I filed #3159 as a reminder to cherrypick the change.

@petemounce
Copy link
Contributor

@mhlopko CI looks quite gorgeously green (though I don't know which builds are which, particularly). Friendly ping?

@ulfjack
Copy link
Contributor

ulfjack commented Jun 13, 2017

We need a second cherrypick, which isn't merged yet.

@ittaiz
Copy link
Member

ittaiz commented Jun 13, 2017 via email

@ulfjack
Copy link
Contributor

ulfjack commented Jun 13, 2017

We also need to cherrypick 019935d.

@hlopko
Copy link
Member Author

hlopko commented Jun 13, 2017

Just baked rc2. If there are no blockers and no more release candidates, we can release on Friday at soonest (following the policy of at least 7 days from initial RC to a release). But since we needed 23 cherry-picks and 9 RCs for 0.5.0 I'm a little bit skeptical. I'll update this thread with all the news.

@petemounce
Copy link
Contributor

https://chocolatey.org/packages/bazel/0.5.2-rc2 is available. As with 0.5.1, this is the msysless flavour.

@hlopko
Copy link
Member Author

hlopko commented Jun 13, 2017

Thanks Pete for a quick packaging!

@ulfjack
Copy link
Contributor

ulfjack commented Jun 14, 2017

We'll need another cherrypick for #3189.

@ulfjack
Copy link
Contributor

ulfjack commented Jun 14, 2017

Btw. the release notes are missing a line break between the cherrypicks.

@buchgr
Copy link
Contributor

buchgr commented Jun 16, 2017

We need a cherry-pick of 1d82d19 ... as this fixes the JDK7 build.

@ulfjack
Copy link
Contributor

ulfjack commented Jun 20, 2017

I also marked #3055 as a release blocker.

@hlopko
Copy link
Member Author

hlopko commented Jun 20, 2017

Possibly we'll need 9eea05d too

@aj-michael
Copy link
Contributor

7dec005, which is in the current RC was just rolled back internally for breaking bazel query --output package, so I'd like to request a cherrypick of the rollback commit (which has not yet been exported -- internal id is 159718294).

@hlopko
Copy link
Member Author

hlopko commented Jun 22, 2017

I just created RC3 with all the mentioned cherry picks:

https://storage.googleapis.com/bazel/0.5.2/rc3/index.html

@hlopko
Copy link
Member Author

hlopko commented Jun 22, 2017

And since I forgot one important cherry pick, we have another rc:
https://storage.googleapis.com/bazel/0.5.2/rc4/index.html

@petemounce
Copy link
Contributor

@hlopko
Copy link
Member Author

hlopko commented Jun 23, 2017

Thank you Pete!

@meteorcloudy
Copy link
Member

@mhlopko Is there still a chance to cherry-pick 5b80458 and 9b3f76e ?
This will make TF GPU build work on Windows.

@ulfjack
Copy link
Contributor

ulfjack commented Jun 27, 2017

Is that a regression? Otherwise I'd prefer if we push 0.5.2 and then start on 0.5.3 right away.

@meteorcloudy
Copy link
Member

Compare to 0.5.0, it's not, because GPU build is broken with it already.
But compare to previous version that works for TF Windows GPU build, it is.
I was thinking leave them to 0.5.3, but TF now wants to use Bazel to build a C library with GPU support.
Similar to http://ci.tensorflow.org/view/All/job/release-libtensorflow-windows/
How soon will 0.5.3 come?

@meteorcloudy
Copy link
Member

If #3248 is no longer a release blocker, I guess we can release 0.5.2 today? If that's the case, I am fine leaving it to 0.5.3. :)

@hlopko
Copy link
Member Author

hlopko commented Jun 27, 2017

It's a release blocker for the next release, be it 0.6.0 or 0.5.3.

@meteorcloudy
Copy link
Member

I see, so let's release 0.5.2 and then start 0.5.3 ASAP

@damienmg
Copy link
Contributor

2 more cherry-picks on their way.

bazel-io pushed a commit that referenced this issue Jun 27, 2017
To cherry-pick for #3086.

Change-Id: I05db2aef04a7cd6cec22c918f00e4b47bae3a1f4
PiperOrigin-RevId: 160249482
bazel-io pushed a commit that referenced this issue Jun 27, 2017
Double quoting make bash interpret the backquote and bazel is not
on the path of the resulting process, showing error on the console.

To cherry-pick for #3086.

Change-Id: I15a4cdbd349300c87bd90d56ecb8df63ecf62dfb
PiperOrigin-RevId: 160250275
@hlopko
Copy link
Member Author

hlopko commented Jun 27, 2017

False alarm, still no more cherry picks for 0.5.2.

@hlopko
Copy link
Member Author

hlopko commented Jun 27, 2017

And here it comes: https://github.com/bazelbuild/bazel/releases/tag/0.5.2 :) Thank you all for all the help and testing!

@petemounce
Copy link
Contributor

Published https://chocolatey.org/packages/bazel/0.5.2 . I'm not a bot, promise :p

@hlopko
Copy link
Member Author

hlopko commented Jun 27, 2017

@hlopko
Copy link
Member Author

hlopko commented Jun 28, 2017

I think we're free to close this issue now. Lets try to get 0.6.0 next?

@hlopko hlopko closed this as completed Jun 28, 2017
@buchgr
Copy link
Contributor

buchgr commented Jun 29, 2017

Congrats on 0.5.2 @mhlopko !

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
P1 I'll work on this now. (Assignee required) type: process
Projects
None yet
Development

No branches or pull requests

9 participants