Add support for Travis-CI integration

[florentx]

It gives basic support for continuous integration.

This is tested, and the build passes: https://travis-ci.org/florentx/asciidoc/builds/24401205

[michel-kraemer]

+1 looks good to me.

Thanks a lot! This PR will be very valuable, in particular it will allow us to check future pull requests automatically. We just need to enable the Travis hook in the projects settings after the PR has been merged.

Cheers,
Michel

[michel-kraemer]

Just one question. The log file of the 2.6 build contains several error messages (related to syntax errors) but it does not fail. Is this supposed to happen?

[florentx]

Hello Michel, I noticed these errors and other limitations as well.
But I thought it would be easier to review the PR if I keep it small and focused.

I did further research about these warnings, and I found that the test suite is a bit too indulgent:

• if some dependency is missing, errors are printed but test is marked PASSED
• if a filter spawns a subprocess and it crashes (like it is the case for 2.6 here), test is marked PASSED
• if a test source or a backend is missing, the test is SKIPPED but the exit code is 0
• if a macro or anything else returns a warning, the warning is silenced, and result is PASSED

I've started to address all these points in a separate branch compat26 of my repo, and I will propose a PR soon.

Then I discovered a last limitation: doing tests/testasciidoc.py update then tests/testasciidoc.py run with the same revision is slow and dumb: it controls that the output is identical when running asciidoc twice.
This is the intended behavior as explained here: http://asciidoc.org/testasciidoc.html#_rationale
It is acceptable when running the tests manually while developing new features but it's not compatible with the continuous integration approach.
We need to think more about this.

I see 2 solutions so far:

• running tests/testasciidoc.py update and committing the result
• or computing an MD5SUMS file of the expected output, storing it in the tests/data directory, and implementing a new flag tests/testasciidoc.py run --compare-md5 which could be used for Travis-CI.

I will open a separate issue for it.

[florentx]

This is the current state for the compat26 branch: florentx/asciidoc@with-travis...compat26

[michel-kraemer]

You are calling doctest.testmod() twice here. Is this intended? If so, why?

Michel

[michel-kraemer]

I like this PR and we can merge it. I just have a very minor comment (see my previous inline comment).

Cheers,
Michel

[florentx]

I've updated the patch, thank you for your review

[elextr]

@michel-kraemer have you set up a travis account and the asciidoc github repository so that travis will work? http://docs.travis-ci.com/user/getting-started/

[michel-kraemer]

I'm on it...

[elextr]

Ok, I was just gonna say it would be appropriate to add a status icon to the readme (although only if its green ;-)

[florentx]

Great 👍

[michel-kraemer]

OK. Seems to work:
https://travis-ci.org/asciidoc/asciidoc

I just added the status icon to the README file.

Cheers,
Michel