Add REPORT as valid HTTP verb #390
Conversation
Resolves #387
|
#387 issue and this PR actually brings us to the need of rethinking how we validate HTTP verbs. |
As of bundler 1.14.0 installing rainbow gem fails due sto bug in rubygems that was fixed. For details see rainbow issue discussion: ku1ik/rainbow#48 (comment)
|
@tarcieri you might wanna about some of the changes made with this pr ;D |
|
@tarcieri I tend to merge this PR as failures on jruby now are pretty specific to travis and bundler. Somhow it simply fails to run on jruby on travis - but runs locally. |
| - 2.0.0 | ||
| - 2.1 | ||
| - 2.2 | ||
| - 2.3.0 | ||
| - 2.3.1 | ||
| - 2.3.2 | ||
| - 2.3.3 |
There was a problem hiding this comment.
What's the point in testing every version of 2.3 under the sun?
There was a problem hiding this comment.
Honestly I have no answer on this question.
| - 2.4.0 | ||
| - ruby-head | ||
| - jruby-9.1.6.0 | ||
| - jruby-9.1.7.0 |
There was a problem hiding this comment.
If you put jruby at the top it will start running first... which is a good thing as it takes the longest
| - 2.4.0 | ||
| - ruby-head |
There was a problem hiding this comment.
I used to be a fan of putting "ruby-head" in but nowadays I think it's just pointless
There was a problem hiding this comment.
Well, I tend to agree. I mean I think it would make sense if build would be triggered at least nightly if not on each ruby-head commit.
| @@ -3,6 +3,7 @@ | |||
| .config | |||
| .rvmrc | |||
| .yardoc | |||
| .ruby-version | |||
There was a problem hiding this comment.
What? I can see an argument for not having one, but why put it in .gitignore?
There was a problem hiding this comment.
Adding .ruby-version to .gitignore so that nobody will commit it accidentally... :D I mean I'm fine when .ruby-version is committed in repo, when only that ruby version is supported (e.g. for some ruby app), but overwise it's as good as adding ruby 2.4 line into Gemfile.
There was a problem hiding this comment.
I've added a .ruby-version to most projects, with the latest version being supported. It automatically configures multiple development tools and is quite handy that way.
There was a problem hiding this comment.
Somehow it was giving me more of a headache than helping. But I will revert that commit.
|
Looks good... well aside from the JRuby failure |
- define FLAKY_ENV const instead of flaky_env? helper - use `:flaky` tag on examples to mark that it should be excluded on `FLAKY_ENV`, instead of `skip "..." if FLAKY_ENV` - do not run tests coverage analysis on FLAKY_ENV
ixti
force-pushed
the
enhancement/report-http-verb
branch
from
750b530 to
5f0935e
Compare
|
@tarcieri master is also failing on jruby ;)) btw |
|
@ixti I'm asking on the Bundler Slack about it. Maybe we should pin to a known-working Bundler. |
|
This appears to be the issue: rubygems/bundler#5340 |
|
This will be fixed in Bundler 1.14.2 |
|
Yeah. Thanks for digging into that. I have locked down bundler on travis for now to the one that worked last time for us ;)) |
ixti
force-pushed
the
enhancement/report-http-verb
branch
4 times, most recently
from
d163fa1 to
a5a001f
Compare
|
Okay. I have no clue how to lock bundler... |
ixti
force-pushed
the
enhancement/report-http-verb
branch
from
a5a001f to
acac075
Compare
|
Let me take a crack at just fixing that issue... |
ixti
force-pushed
the
enhancement/report-http-verb
branch
2 times, most recently
from
e33f338 to
79797ad
Compare
ixti
force-pushed
the
enhancement/report-http-verb
branch
from
79797ad to
84dcb1f
Compare
|
After some time and couple of rounds when I thought I finally got insane - I have managed travis to run with bundler / rubygems known to work :D |
|
Looks good! |
Upstream changes (from CHANGES.md): ## 2.2.0 (2017-02-03) * [#375](httprb/http#375) Add support for automatic Gzip/Inflate ([@Bonias]) * [#390](httprb/http#390) Add REPORT to the list of valid HTTP verbs ([@ixti]) ## 2.1.0 (2016-11-08) * [#370](httprb/http#370) Add Headers#include? ([@ixti]) * [#364](httprb/http#364) Add HTTP::Response#connection ([@janko-m]) * [#362](httprb/http#362) connect_ssl uses connect_timeout (Closes #359) ([@TiagoCardoso1983])
Resolves #387.