Граф коммитов

369 Коммитов

Автор SHA1 Сообщение Дата
Takashi Kokubun a1be11eb12
Revert "Remove arm64 from allow_failures"
This reverts commit 212f4d49ba.

It worked on PR, but master branch builds have another build issue.
https://travis-ci.org/ruby/ruby/jobs/608303393
2019-11-06 09:53:02 -08:00
Takashi Kokubun 212f4d49ba
Remove arm64 from allow_failures 2019-11-06 09:10:27 -08:00
Nobuyoshi Nakada 84b69c520e
Edit vm_opts.h for VM options
Revert 46f175ed5c partially.

Also tool/ruby_vm scripts refer the options in vm_opts.h.
2019-10-26 02:54:29 +09:00
Nobuyoshi Nakada 46f175ed5c
Use cppflags to pass preprocessor options 2019-10-23 02:53:36 +09:00
Alan Wu 97a03a0356 Add a cron job to keep CPDEBUG working 2019-10-22 12:21:02 +09:00
Takashi Kokubun 5c80dffd05
Revert "Disable ccache on arm64 build"
This reverts commit e8124729fb.

It worked once, but soon after that it stopped working again
https://travis-ci.org/ruby/ruby/jobs/598491972
2019-10-16 08:28:53 -07:00
Takashi Kokubun eef25f95b7
Maintain the intention of comment in Travis matrix 2019-10-15 23:04:34 -07:00
Takashi Kokubun e8124729fb
Disable ccache on arm64 build
Follow up https://github.com/ruby/ruby/pull/2559

According to experiments in https://github.com/ruby/ruby/pull/2560,
disabling cache seems to make arm64 build succeed.
2019-10-15 22:54:43 -07:00
Takashi Kokubun 4ccb49bad3
Enforce --jit-debug test by another way 2019-10-15 22:47:20 -07:00
Takashi Kokubun 759576a03f
Fix test path on travis build 2019-10-15 22:01:25 -07:00
Takashi Kokubun 2679e10942
Test without --jit-debug by default
and let RUBY_DEBUG=1 job test it.

We usually don't use --jit-debug and we should test no --jit-debug by
default.
2019-10-15 21:37:01 -07:00
Jun Aruga d97229de5c Add arm64 case on Travis. (#2559)
misc-16234
2019-10-15 20:59:40 -07:00
Takashi Kokubun 38069a3a55
Try using Ruby 1.9.3 on Travis xenial
Isn't it there? http://rubies.travis-ci.org/
2019-09-18 18:54:09 +09:00
卜部昌平 1364217bcd add missing dependency for .travis.yml 2019-09-09 21:27:40 +09:00
Takashi Kokubun 1a9cc3b27c Avoid defining unused instructions 2019-09-03 14:22:44 +09:00
Takashi Kokubun 6c3ab99eae
Now it's always os: linux 2019-09-03 01:11:18 +09:00
Takashi Kokubun 52f0c31558
Fix accidental leftover for Linux
In 91aa8bfff8, my understanding of the branch was inverted.
2019-09-03 01:06:29 +09:00
Takashi Kokubun 91aa8bfff8
Drop unstable Travis osx in favor of GitHub Actions
For some reason, the Travis osx environment has been really unstable.
It failed on today's cron too:
https://travis-ci.org/ruby/ruby/builds/579843163

As we have almost the same test environment (including OpenSSL version)
in GitHub Actions and it seems to be more stable and faster, I think
there's no motivation to maintain Travis osx CI environment.

By removing this, we'd be able to simplify .travis.yml as well.
2019-09-03 01:01:48 +09:00
Takashi Kokubun 1d68d29c14
Drop obsoleted IRC channel reference
travis-ci user does not live in #ruby-ja. Therefore the notification
isn't working anymore.
2019-09-03 00:36:19 +09:00
Takashi Kokubun f1d8d0f47b
Drop obsoleted branch references
We don't have topic branches on ruby.git anymore.
2019-09-03 00:35:24 +09:00
Takashi Kokubun 60a49a6ff9
Fix SUPPORT_JOKE CI on Travis
It looks like a flag which cannot be enabled on configure.
2019-09-03 00:07:37 +09:00
Takashi Kokubun 834e771860
Run daily cron against SUPPORT_JOKE 2019-09-02 23:48:38 +09:00
Takashi Kokubun cda5745c1b
Automatically detect missing symbols
which are usually optimized away by -O3.

This CI can detect missing exports like
ea84a68075 which was needed for
761346a960.
2019-09-01 16:31:15 +09:00
Takashi Kokubun 602fbd4338
Limit the apt addon enablement to trusty
8897fa60fd worked.
But we want to retry that part and this commit is trying to fix the
problem.
2019-08-31 17:34:25 +09:00
Takashi Kokubun 8897fa60fd
Try using built-in apt source for gcc-8
58a373e9ed and
3868df8333
didn't work.
2019-08-31 17:18:59 +09:00
Takashi Kokubun 3868df8333
Try to fix apt-get update error for trusty (take 2)
https://travis-ci.org/ruby/ruby/jobs/579116510
2019-08-31 17:05:53 +09:00
Takashi Kokubun 58a373e9ed
Try to suppress errors in BASERUBY Travis
https://travis-ci.org/ruby/ruby/jobs/579108692
2019-08-31 16:44:44 +09:00
Kazuhiro NISHIYAMA 682aaf6af5
Replace numbered parameters with named parameters [ci skip] 2019-08-29 14:35:09 +09:00
Takashi Kokubun 235d810c2e
Deprecate alerting multiple channels
Some CIs report to two channels, and some others report to only one.
This makes it consistent. Only alert channel should be alerted.
2019-08-20 22:33:32 +09:00
Takashi Kokubun 97c377e2b1
Mark Travis osx cron-only
because it has often hanged like
https://travis-ci.org/ruby/ruby/jobs/573691637,
and we also have almost the same test suite on GitHub Actions now, which
seems to be stable in `make check` so far.
2019-08-19 18:13:09 +09:00
Takashi Kokubun 829b800c10
Reoder Travis matrix.include
to clarify what's running for every commit.
2019-08-18 19:10:24 +09:00
Takashi Kokubun a5f9306065
Simplify sudo specification 2019-08-12 20:20:59 +09:00
Takashi Kokubun 1bec27f05b
Add exec to apt-get update
to propagate exit status to travis_retry properly.
2019-08-12 19:52:38 +09:00
Takashi Kokubun b1ef14176b
We did not have tool/ before checkout
anyway we don't need authorization here.
Also retry does not seem to work in the original version, so let's
extend this with retries as a separate github action later.
2019-08-12 14:37:26 +09:00
Takashi Kokubun 4f10a61eaa
Stop relying on actions/checkout
because it randomly fails on authorization like:
https://github.com/ruby/ruby/runs/190887455
Also the backoff seems too short. Maybe we need tool/travis_retry.sh for
this too.

Cloning ruby/ruby does not need authorization. We don't need to use
actions/checkout.
2019-08-12 14:30:43 +09:00
Takashi Kokubun b1ad628cdf
Resurrect travis_wait for test-all
as we dropped -v.
2019-08-12 13:45:57 +09:00
Takashi Kokubun 0609087a6b
Parallelize osx test-all too 2019-08-10 23:57:20 +09:00
Nobuyoshi Nakada c21c001733
Re-enable parallel build/test on OSX 2019-08-10 23:30:50 +09:00
Nobuyoshi Nakada 00e7ef7e8d
Suppress flags messages [ci skip] 2019-08-10 19:50:46 +09:00
Nobuyoshi Nakada 5e3fb0b357
Touch the checked out source to fix clock skew on all platform 2019-08-10 16:33:32 +09:00
Nobuyoshi Nakada c5fbe2a19f
Moved options to $travis_apt_get_options 2019-08-10 15:48:39 +09:00
Takashi Kokubun 330e3f19c3
Refactor .travis.yml by introducing travis_retry.sh
Not using official travis_retry.bash, because it's not supporting to
modify backoff seconds.
0ad8f1886b/lib/travis/build/bash/travis_retry.bash

Not using official `travis_apt_get_update` function because it does not
propagate exit status to be used by retries.
0ad8f1886b/lib/travis/build/bash/travis_apt_get_update.bash

Co-Authored-By: Nobuyoshi Nakada <nobu@ruby-lang.org>
2019-08-07 21:47:03 +09:00
Takashi Kokubun a0fe396555
Simplify break logic of Travis retries
Co-Authored-By: Nobuyoshi Nakada <nobu@ruby-lang.org>
2019-08-07 09:40:11 +09:00
Kazuhiro NISHIYAMA b839721ae7
.travis.yml: Use `-eq` instead `=` to compare integers 2019-08-07 09:03:08 +09:00
Takashi Kokubun 2d6b92c5f6
Retry ubuntu-toolchain-r-test apt source setup
In 614c90fe21, I assumed
apt-add-repository has been stable recently, but I saw PR randomly
failed for it today.

This commit only deals with "ubuntu-toolchain-r-test" and does NOT deal
with `llvm-toolchain-xenial-8` intentionally, because what it does
(mainly curl from build.travis-ci.org) seems to be more reliable than
"ubuntu-toolchain-r-test". Also `&clang-8` jobs are basically
allow_failures and it's less important to be stabilized.
2019-08-07 01:26:33 +09:00
Takashi Kokubun 255e22a827
Use gcc-8 for BASERUBY test
as it failed https://travis-ci.org/ruby/ruby/jobs/567942969, unlike my
fork CI.

This just rollbacks a minor change in
798474deaf10dbea70c02724b5acf4487bbed660.
2019-08-06 00:50:48 +09:00
Takashi Kokubun 614c90fe21
Control Travis apt retries on our own
because Travis team does not do it for us:
https://github.com/travis-ci/travis-build/pull/1712

The retried part has failed often even in one day:
https://travis-ci.org/ruby/ruby/jobs/567802384
https://travis-ci.org/ruby/ruby/jobs/567802388
https://travis-ci.org/ruby/ruby/jobs/567695879
https://travis-ci.org/ruby/ruby/jobs/567666931

For doing it easily, this also changes major aliases to
compiler-specific ones, because partially updating `before_install`
logic using `env` key was too hard and we needed to directly write
`before_install` for each package set. As a bonus of it, it may also
skip installing unnecessary packages when just `os: linux` is needed.

I'll revert this if this patch does not contribute to stabilize CI.
2019-08-06 00:33:33 +09:00
Takashi Kokubun 44ddfec0bd
Add missing backslash 2019-08-03 01:07:48 +09:00
Takashi Kokubun 4eb54c49f7
Stop using homebrew addon
to control erros on our own.
2019-08-03 00:55:37 +09:00
Takashi Kokubun 247b06d400
Revert "Give up `brew update` on Travis"
This reverts commit e05f397f08.
Failed... https://travis-ci.org/ruby/ruby/jobs/566961111

We should talk to Travis people instead...
2019-08-02 23:01:54 +09:00
Hiroshi SHIBATA 3e4e8dbe20
Uninstall ntlm-http and plist for failure of osx test.
It contains too old configuration that is autorequire. It will be
  removed at the RubyGems 3.1.0.
2019-07-31 07:47:03 +08:00
Takashi Kokubun a5ea55feb4
Actually remove travis_wait
I forgot to amend again... Details are explained in
f6a6b21f09
2019-07-29 20:57:57 +09:00
Takashi Kokubun f6a6b21f09
`travis_wait` silences any output
and thus it does not work with `-v` for investigating hangs well.
This seems to be by design: https://github.com/travis-ci/travis-ci/issues/4190

Also I simplified a comment about `homebrew.update`.
I had this in-flight change while editing e05f397f08
but forgot to ammend this.
2019-07-29 20:55:19 +09:00
Takashi Kokubun e05f397f08
Give up `brew update` on Travis
No `brew update` causes "Error: Your Homebrew is outdated" like https://travis-ci.org/ruby/ruby/jobs/547485832,
but doing `brew update` is also problematic like https://travis-ci.org/ruby/ruby/jobs/564916879.

Hoping that the former case is more rare, let's try no `brew update`
again.
2019-07-29 20:25:23 +09:00
Takashi Kokubun 5af28fec2f
Resurrect -v to debug osx problems
Formerly we did f432fd6ea5, but it did not
eliminate our problems: https://travis-ci.org/ruby/ruby/jobs/564804923
2019-07-29 19:49:20 +09:00
Takashi Kokubun 4b4d8a662b
Stop isolating test_gc_compact on CI
I think it's been stable these days.
2019-07-29 19:45:32 +09:00
Nobuyoshi Nakada 61f0f71e5b
Moved osx_image
Set osx_image under each configuration, as it decides the OS (and
kernel) version not only Xcode version, and the configuration name
contains the kernel version.
2019-07-29 11:44:56 +09:00
Takashi Kokubun adf13625ae
Extend travis_wait to 50min for osx
but shorten --timeout-scale for now to avoid finishing with timeout
if possible.

timeout: https://travis-ci.org/ruby/ruby/jobs/564370175
2019-07-27 23:46:40 +09:00
Takashi Kokubun e3b613a669
Include travis osx timeout
for failures like https://travis-ci.org/ruby/ruby/jobs/564351066
2019-07-27 21:34:35 +09:00
Nobuyoshi Nakada 957a29fc6e
Bump osx_image on Travis CI to xcode11
Also Homebrew is up-to-date and "Updating Homebrew" takes less
than one minute.
2019-07-26 09:21:04 +09:00
Nobuyoshi Nakada 538ba984c8
Split ls line 2019-07-24 21:07:12 +09:00
Nobuyoshi Nakada 96b0d7cd6f
GNU ls -T has different meaning 2019-07-24 21:03:01 +09:00
Takashi Kokubun 65a9d4b0f7
Fix typo [ci skip] 2019-07-24 20:51:29 +09:00
Takashi Kokubun 7b1893c771
Resurrect timestamp debug log
because it failed again https://travis-ci.org/ruby/ruby/jobs/563026412
even after 676df311d9.
2019-07-24 20:50:51 +09:00
Nobuyoshi Nakada 676df311d9
Reset mtime of all files on osx
Often checked out files are in the future on OSX image.
2019-07-23 22:46:03 +09:00
Nobuyoshi Nakada e8d4f0cbc7
Show seconds and in the modified time order
[ci skip]
2019-07-23 17:59:43 +09:00
Nobuyoshi Nakada 0338c44bde
Retry to update Unicode timestamp 2019-07-23 17:39:03 +09:00
Samuel Williams c584dd8460
Move travis coroutine check to cron only. 2019-07-20 19:44:49 +12:00
Takashi Kokubun e7b5b9144a
--debug was not functional either
Even after 19d592dc82, the retry seems not
functional. Let's just add --debug from the beginning because an output
of each step is collapsed and we don't see `make up` output so often.
2019-07-20 15:45:34 +09:00
Samuel Williams b8242bce23
Add `ucontext` and `copy` coroutine implementations to test matrix. 2019-07-18 20:54:54 +12:00
Nobuyoshi Nakada 19d592dc82
Somehow `if` didn't work 2019-07-16 18:41:11 +09:00
Nobuyoshi Nakada 76c6cf2cbc
Print debugging information when updating failed 2019-07-16 17:23:43 +09:00
Takashi Kokubun d1e2650aca
Force-fetch unicode update only when it's needed
nobu said that we could be banned if we aggressively downloaded unicode
file from Travis.
2019-07-15 13:49:15 +09:00
Takashi Kokubun 73346823e1
Try to prevent random build failure on Travis osx 2019-07-15 11:04:16 +09:00
Nobuyoshi Nakada ac2866005b
Add empty depend file if extconf.rb only exists 2019-07-14 03:04:23 +09:00
Nobuyoshi Nakada d0ef5c21fb
Check dependencies on Travis-CI 2019-07-09 23:13:43 +09:00
Kazuhiro NISHIYAMA 07e9a1d998
Debugging failures on x86_64-darwin17 of travis-ci
https://travis-ci.org/ruby/ruby/jobs/553346285
2019-07-03 10:23:01 +09:00
Takashi Kokubun 35cf3e8904
The default branch is changed to master
and now Travis should run on master, instead of trunk.

(this is testing git push to master branch)
2019-07-01 00:14:12 +09:00
Takashi Kokubun 310a42e89c
Explain rubyspec Travis job in job name
Eregon did not want 7b9dd379bf
because people only see Travis job name and don't see .travis.yml.

I believe the new job name is simple and informative enough to explain
why it exists.
2019-06-27 22:32:28 +09:00
Benoit Daloze 2c02d5714e Add a comment explaining what to do if specs fail on Ruby 2.4 2019-06-27 11:57:23 +02:00
Takashi Kokubun 148f50fc78
Update homebrew on Travis
because Travis is crashing like "Error: Your Homebrew is outdated.
Please run `brew update`."
https://travis-ci.org/ruby/ruby/jobs/547485832
2019-06-19 10:09:37 +09:00
Kazuhiro NISHIYAMA 42ac8890ef
Fail test if load ~/.irbrc 2019-06-11 12:33:34 +09:00
Nobuyoshi Nakada c7477c3e1f
Moved Makefile.in under template 2019-06-07 09:09:58 +09:00
Nobuyoshi Nakada 200c840b95
Revert "Moved Makefile.in under template"
This reverts commits:
* 6f9d5fafe0
* bb3c89b643

And remove the dependency of Makefile on Makefile.in
transitionally.
2019-06-07 09:03:25 +09:00
Nobuyoshi Nakada bb3c89b643
Makefile.in moved under template 2019-06-07 08:58:53 +09:00
Takashi Kokubun 468b475e61
.travis.yml: Minor reorder for shortening 2019-05-29 16:09:07 +09:00
Takashi Kokubun 3e54ff67e6
Test BASERUBY: Ruby 1.9.3 on Travis
We have no clear assertion or check of BASERUBY requirement.
I want to make the current situation more explicit.

I'm NOT saying we should support Ruby 1.9.3 here,
but I'm just checking the situation as per 05bc14d81a.

FYI, at this moment Ruby 1.8.7 did not work with this Travis config, like:
https://travis-ci.org/k0kubun/ruby/builds/538459100
2019-05-29 06:45:33 +09:00
Takashi Kokubun d070523e7b
Drop unused Travis config: universal-darwin17
This has been unused since b7f5c573ef.
2019-05-29 05:27:55 +09:00
Takashi Kokubun cf904d9f9f
Avoid doubly building Travis and AppVeyor
but on "master" branch.

For Pull Request, I changed the approach from
d9b338a53f and
277e68825a.
2019-05-28 10:02:31 +09:00
Takashi Kokubun be851692ce
Build trunk branch too to trigger Travis on PR 2019-05-25 07:24:59 -07:00
Kazuhiro NISHIYAMA 2541578096
Add leaked-globals to .travis.yml 2019-05-24 14:32:29 +09:00
Urabe, Shyouhei 8fce83339b disable _FORTIFY_SOURCE
Sanitizers and fortifications do not interface, and there is currently
no plan for them to work together.

See also https://github.com/google/sanitizers/issues/247
2019-05-23 17:24:53 +09:00
Nobuyoshi Nakada 1ccc2eeba0
Pretend to update all Unicode files 2019-05-23 14:18:26 +09:00
Takashi Kokubun 277e68825a
.travis.yml: Test master branch instead of trunk 2019-05-22 08:33:44 -07:00
Nobuyoshi Nakada af1502175b
Touch the timestamp file for Unicode files 2019-05-19 05:45:30 +09:00
Kazuhiro NISHIYAMA e8e415b534
Add workaround for `Permission denied` of `cp`
see r67347
recent log: https://travis-ci.org/ruby/ruby/jobs/529640417
2019-05-08 16:46:18 +09:00
Takashi Kokubun 229e5053a1
Drop -fs from rubyspec to see test results easily
Recently `ruby/spec on Ruby 2.4` seems stable.
2019-05-08 14:59:21 +09:00
Urabe, Shyouhei a116f04cca suppress some UBSAN sanitizers
They are not "undefined".  UBSAN reports them because it thinks
they are "often unintentional".  We see the report rather annoying.
2019-04-29 21:52:44 +09:00
Kazuhiro NISHIYAMA 69cad44fac
Reduce debug output
because I found machine clock problem
see r67347
https://travis-ci.org/ruby/ruby/jobs/525784924
https://travis-ci.community/t/mtime-of-source-codes-are-sometimes-newer-than-build-time-clock-skew/3215
2019-04-29 12:21:50 +09:00
Kazuhiro NISHIYAMA f0776e3203
Add more debug print for random CI failure on osx Travis
and remove `git status` with noisy rvm trace log

see r67347
2019-04-29 01:23:20 +09:00