[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Chicken-users] Travis CI
From: |
Mario Domenech Goulart |
Subject: |
Re: [Chicken-users] Travis CI |
Date: |
Thu, 01 May 2014 01:42:25 +0000 |
User-agent: |
Gnus/5.13 (Gnus v5.13) Emacs/24.3.50 (gnu/linux) |
Hi Hugo,
On Wed, 30 Apr 2014 20:56:30 -0300 Hugo Arregui <address@hidden> wrote:
> Sorry for the late response!.
No problem at all!
>>Instead of just running "chicken-install -test", you can install
>>salmonella and run it. It'll perform some extra checks that
>>chicken-install doesn't. In case you run salmonella, you'll probably
>>also want to run "salmonella-log-viewer salmonella.log" to show details
>>of the build/test steps.
>
> Excellent advice!
Actually, now that I have thought more about it, it doesn't sound so
simple. The only way to determine whether installation or tests of your
egg failed/succeeded is by inspecting the salmonella log file. It's
certainly possible, but not as straightforward as checking salmonella's
exit code, as you probably do with chicken-install. Salmonella will
only exit a code different from zero if it crashes, so, it's not
meaninful for your purposes.
Some time ago I thought about making a command line tool to query the
log file, but since I didn't have an immediate use for it, I ended up
not implementing it. :-) For example, you could run something like:
$ salmonella-query --action=install --status awful-blog salmonella.log
That would return the installation status (chicken-install's exit code)
for awful-blog.
Maybe I should add a tool like that to salmonella.
>>If you want to use that infrastructure to test your project, it's just a
>>matter of adding it to the official coop. It'll be automatically added
>>to the daily test procedure. New eggs are aways welcome!
>
> Thanks!, I think I will. Nevertheless, I was thinking on Travis as
> another instance of testing, as is build every time you push, you can
> catch some mistakes before the coop tests run (that is why I don't
> care too much about the old chicken version). Or maybe also prior to
> release it as an egg.
That's very good, indeed. From some perspective, having an old chicken
is also good, because we don't usually test eggs with older chickens.
>>P.S.: sorry for the spiffy-request-vars bug. :-) Fixing it is in my
>>TODO list.
>
> Haha, you never miss a thing!. If its for me, please don't bother
> to fix it, is an old version and I don't need it.
For you and for and for the health of the coop. :-)
chicken-install chokes if you run, for example, "chicken-install -test
awful" when you don't have spiffy-request-vars installed. Since
spiffy-request-vars' tests depend on awful, and awful's installation
depend on spiffy-request-vars, it generates a circular dependency. It's
actually only circular because chicken-install mixes up installation and
test phases (I wouldn't say it is an issue, just the way chicken-install
works).
Best wishes.
Mario
--
http://parenteses.org/mario