[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [STUMP] Travis CI broken on Github
From: |
David Bjergaard |
Subject: |
Re: [STUMP] Travis CI broken on Github |
Date: |
Tue, 11 Aug 2015 16:25:16 -0400 |
User-agent: |
Gnus/5.13 (Gnus v5.13) Emacs/24.5 (gnu/linux) |
Woohoo, thanks!
Evan <address@hidden> writes:
> Hey David,
>
> Please see issue [217] and PR [218] about this issue. I haven't attempted to
> build this pr
> on anything but ccl, so travis will tell us how it fairs shortly. If all goes
> well, this
> should solve ccl's build problems.
>
> [217]https://github.com/stumpwm/stumpwm/issues/217
> [218]https://github.com/stumpwm/stumpwm/pull/218
>
> -E
>
> On 08/11/2015 10:25 AM, David Bjergaard wrote:
>> Hi Evan,
>>
>> Responses inline:
>>
>> Evan <address@hidden> writes:
>>
>>> When you say "sbcl always passes" do you mean even when broken code is
>>> committed?
>> I mean when sbcl builds successfully if the code is OK. If you commit broken
>> code, then sbcl will fail. Conversely clisp and ccl always fail even if the
>> code
>> is clearly fine (ie you add a comment and commit it, or change a file that
>> doesn't depend on the build)
>>>
>>> Where is the code for the travis builds kept? Do we have a repo containing
>>> the directives
>>> that travis follows?
>> As far as I can tell the recipe is stumpwm/.travis.yml and "install.sh"
>> which
>> is the main driver of the code. The latest and greatest is here:
>> https://raw.githubusercontent.com/luismbo/cl-travis/master/install.sh
>>>
>>> With regards to ditching travis if we can't figure out how to get the
>>> ccl/clisp builds to
>>> work properly, it seems that we're still better off with one of the
>>> platforms being
>>> tested, even if it means you have to investigate every failed travis build.
>>> It's not
>>> ideal, and it'd probably be worth it to either cut out or move the
>>> ccl/clisp builds to
>>> their own branch while we investigate what's wrong, but ditching travis all
>>> together is
>>> probably unnecessary.
>> Ok agreed, but I would prefer that we be able to have a "build passing"
>> badge on
>> stumpwm's main page so users can at least judge whether or not the master
>> branch
>> is safe to use.
>>
>> David
>>
>>
>>
>>>
>>> -E
>>>
>>> On 08/07/2015 03:34 AM, David Bjergaard wrote:
>>>> sbcl always passes, while clisp and ccl fail:
>>>> ccl: https://travis-ci.org/stumpwm/stumpwm/jobs/74369075
>>>> clisp: https://travis-ci.org/stumpwm/stumpwm/jobs/74369077
>>>> sbcl: https://travis-ci.org/stumpwm/stumpwm/jobs/74369074
>>>>
>>>> David
>>>>
>>>> Evan <address@hidden> writes:
>>>>
>>>>> Can you give an example of how the travis builds are broken?
>>>>>
>>>>> -E
>>>>>
>>>>> On 08/06/2015 03:16 AM, David Bjergaard wrote:
>>>>>> Hi,
>>>>>>
>>>>>> Does anyone know how to fix the travis builds? Its making it really hard
>>>>>> to know
>>>>>> if its safe to merge PRs or if I'm not breaking stuff. If it can't be
>>>>>> fixed, I
>>>>>> would prefer that we remove it and go back to the old way (waiting for
>>>>>> bug
>>>>>> reports :$)
>>>>>>
>>>>>> Dave
>>>>>>
>>>>>> _______________________________________________
>>>>>> Stumpwm-devel mailing list
>>>>>> address@hidden
>>>>>> https://lists.nongnu.org/mailman/listinfo/stumpwm-devel
>>>>>>