Table inline code

[MustafaHaddara]

There's probably a better way to do this -- I feel like tables.py should import from inlinepatterns or something like that.

Tests passed, and I added 4 tests for tables. On the master branch, the first 3 pass. These commits get the 4th one working.

Resolves #436

[mitya57]

I think this should be named _row_has_unpaired_backticks, as this really has little to do with code.

[mitya57]

In general the idea looks good to me 👍, though I have left some comments on the code style issues.

Travis fails on Coverage check on Python 3.2 because Coverage doesn't seem to support that version, should we maybe exclude that check from the Travis config?

[waylan]

Interesting. What about when multiple backticks are used to delimit a codespan? Might want to add to the test some more cells:

`` some | code ``
``` some | code ```
```` some | code ````
`` some ` | ` code ``
``` some ` | ` code ```
```` some ` | ` code ````

Note that each of the first three should result in the following HTML:

<code>some | code</code>

And the last three:

 <code>some ` | ` code</code>

[waylan]

According to the coverage report, this is never tested. Might want to add a test for it.

[waylan]

Travis fails on Coverage check on Python 3.2 because Coverage doesn't seem to support that version, should we maybe exclude that check from the Travis config?

Yes, Coverage 4.0 does not support Python 3.2. We could require Coverage 3.x, but I don't really think it matters much at this point.

[waylan]

@MustafaHaddara, @mitya57 already pointed out the specific issues to you, but the Flake8 test shows all those style/formatting related issues for you as well.

[MustafaHaddara]

@waylan @mitya57 I've made the style fixes and updates to the test...but attempting to use multiple backticks to denote a code block is still broken.

[waylan]

@MustafaHaddara this looks like a good start. When the multiple backticks issue gets worked out, either @mitya57 or I will merge it. Thanks for the report and getting the work started on this.

[waylan]

The Flake8 test is indicating this has trailing whitespace.

[MustafaHaddara]

@waylan @mitya57 updated to deal with the issue of multiple backticks. I changed the implementation strategy from the manual string parsing I initially started with and instead, I use inlinepatterns.BacktickPattern to find inline code blocks inside of a table cell.

[waylan]

Please don't do this. Lets keep all inline parsing in the inline parsing step of the parser. If you want to use the regex for codespans to identify them, I understand, but when the table blockparser is done, the cell should still contain Markdown formatted inline text. If you don't do it that way, then other (inline) extensions could fail in unpredictable ways on table cells with codespans.

[MustafaHaddara]

@waylan @mitya57 fixed as requested. All tests pass for me. Speaking of tests, I didn't notice there was already some tables tests in /tests/extensions/extra/ so I had created a new tables test in /tests/extensions/. Do you want me to remove the test file I added and move those test cases to the older file?

And would you like me to rebase and squash all of these commits into one commit to encompass the entire patch?

[waylan]

The tables extension is part of extra, so yes I would prefer the tests be under tests/extensions/extra/. And yes, it would make sense to squash all of the commits.

[waylan]

Any reason why you're not using match.end(group) here?

Every thing else looks good to me.

[MustafaHaddara]

@waylan In response to your point about match.end(group), that gives us the end of the group -- and we only have groups for the first code block delimiter, the text inside the code block, and the text after the code block. Asking for the end of a group won't help.

What we really want is the beginning of the text outside the code block, ie. match.start(group).

Please review, and if it is good with you I'll rebase into one commit.

[waylan]

Looks good. Clean up the Flake8 errors and rebase and I'll merge.

[MustafaHaddara]

The Flake8 test didn't run, and I'm not sure why. It threw ERROR: InvocationError: '/home/travis/build/waylan/Python-Markdown/.tox/flake8/bin/flake8 /home/travis/build/waylan/Python-Markdown/markdown /home/travis/build/waylan/Python-Markdown/tests /home/travis/build/waylan/Python-Markdown/setup.py /home/travis/build/waylan/Python-Markdown/run-tests.py'

Is there a more detailed log I can look at somewhere?

[waylan]

The Travis Report (linked below) shows the details for each build. Just select the failing build and it will give you a full log. As per our earlier discussion, you can ignore the Python 3.2 build due to a version incompatibility with Coverage.

[MustafaHaddara]

Rebased, all ready to go.