[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: midi2ly produces faulty output when using "-e"
From: |
David Kastrup |
Subject: |
Re: midi2ly produces faulty output when using "-e" |
Date: |
Wed, 27 Jul 2016 13:51:58 +0200 |
User-agent: |
Gnus/5.13 (Gnus v5.13) Emacs/25.1.50 (gnu/linux) |
Mojca Miklavec <address@hidden> writes:
> On 27 July 2016 at 12:32, David Kastrup wrote:
>> Mojca Miklavec writes:
>>
>>> If I run midi2ly with -e, this is what I get as lilypond output:
>>> f,4 <f4 a4 c'4 >4 <a'4 c''4 f4 a4 c'4 >4
>>>
>>> The problem is that durations are not allowed inside chords with
>>> single "brackets".
>>>
>>> This looks like a bug to me.
>>
>> Tracker issue: 4945 (https://sourceforge.net/p/testlilyissues/issues/4945/)
>> Rietveld issue: 308830043 (https://codereview.appspot.com/308830043)
>> Issue description:
>> Fix midi2ly chord durations. Contains commits: midi2ly should
>> omit redundant chord durations midi2ly -e should not print
>> durations in chords
>>
>> No idea how feasible it is for you to fetch/test the patch.
>
> Thank you. It's just a python script, so applying a patch is easy (no
> need to recompile :).
>
> Does this deserve a test case? That is: do you have/run a test suite?
> If so, I could try to come up with a few sample minimal examples that
> produce potentially problematic midi files to test against.
We do but I am fuzzy on how midi2ly figures into it. Basically the test
suite is found in input/regression and is executed with "make test"
(often preceded by "make test-clean"). There is also "make
test-baseline" for making a test and recording the results (which are
not cleared by "make test-clean") and "make check" for comparing to the
recorded results semi-graphically (the full graphic comparison is only
done for results with different metrics) after which "firefox
out/test-results/index.html" shows the comparison.
No idea if midi2ly figures into any displayed difference or just appears
in the log output.
--
David Kastrup