Matlab disabled

[vhda]

Just to open this for discussion: why is matlab parser being disabled to overcome the mapping conflict with objective c?
Why not use --langmap or --language-force?

[masatake]

Good question!

You are right. I disabled it temporary to pass tests; from the name of file(*.s) matlab parser is choose always in tests.
Even if there is no test related issue, it is nice that a proper parser is choose automatically.
I think ctags should have preprocessing stage for choosing proper parser; we can choose a parser from scores returned from the stage. The most of preprocessing stage may be language specific.
However, some can be generic; we can use modeline of editors as hit. See http://www.wireshark.org/tools/modelines.html.

determineInterpreter is good place to start. determineInterpreter is used only if a language parser cannot be choose from file suffix. It also should be used when conflicting is occured from the suffix rule.

[masatake]

I started working on this area. See #39.

[masatake]

Sorry, look at #41.
#41 introduces change incompatible with exuberant ctags.
Please, review it.

My colleague tells me 2-gram can be used for scoring.

[masatake]

h of C++ conflicts that of ObjectiveC
cmd of DosBatch conflicts that of REXX

[vhda]

In the meanwhile, might I suggest that we reenable the Matlab parser and adjust the failing test cases such that they use the correct parser?
I don't it makes sense to disable functionality for testability purposes.

[masatake]

With the modeline patches we can adjust the failing test cases such that they use the correct parser.
However, I would like to fix this issue completely with the 2gram aproach; we don't have serious deadline now.

Do YOU need matlab parser now?

[vhda]

I thought we were trying to coordinate our efforts with the SF maintainers.
I don't think they will love to see parsers disabled just to make test
units pass and may even take this effort not seriously because of that.

On Mon, Jun 23, 2014 at 1:23 PM, Masatake YAMATO notifications@github.com
wrote:

With the modeline patches we can adjust the failing test cases such that
they use the correct parser.
However, I would like to fix this issue completely with the 2gram aproach;
we don't have serious deadline now.

Do YOU need matlab parser now?

—
Reply to this email directly or view it on GitHub
#38 (comment).

[masatake]

I think they will love to see that ctags chooses a proper parser automatically from input.
However, this is just my opinion.

Do you have any good patch?
What I have is very ugly and doesn't work with out modeline patches.

[vhda]

Please check pull request #46.

[masatake]

2-gram engine shows good detections:)
After cleaning up the table generator, I will send pull request.

% ./ctags -o - Test/.m > /dev/null
./ctags -o - Test/.m > /dev/null
about Test/matlab_backtracking.m winner: MatLab
about Test/matlab_backtracking.m winner: MatLab
about Test/matlab_test.m winner: MatLab
about Test/matlab_test.m winner: MatLab
about Test/objc.m winner: ObjectiveC
about Test/objc.m winner: ObjectiveC
about Test/objectivec_implementation.m winner: ObjectiveC
about Test/objectivec_implementation.m winner: ObjectiveC

[vhda]

That's great news :)
Keep it going!

[masatake]

Could you try https://github.com/masatake/ctags/tree/tg?
Matlab parser is enabled.

[vhda]

Your work looks great, but I still did not have the time to run some tests. Please give me a day or two to do so.

[vhda]

Great work!
I don't have anything to add.

[masatake]

Thank you for testing.
I will start merging modeline branch first.

[vhda]

Excellent work Masatake!
I have much to learn from you :)

On Mon, Jul 7, 2014 at 11:42 AM, Masatake YAMATO notifications@github.com
wrote:

Thank you for testing.
I will start merging modeline branch first.

—
Reply to this email directly or view it on GitHub
#38 (comment).

[masatake]

tg branch is merged.

I have to add tg tables for

h of C++ conflicts that of ObjectiveC
cmd of DosBatch conflicts that of REXX

But these are rather unimportant task.