lilypond-user
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: svg output


From: Graham Percival
Subject: Re: svg output
Date: Mon, 18 Jan 2010 10:20:04 +0000

On Mon, Jan 18, 2010 at 9:47 AM, Dmytro O. Redchuk
<address@hidden> wrote:
> У сб, 2010-01-16 у 15:33 +0000, Graham Percival пише:
>> It would be nice if somebody could check the regtest comparison...
>> it's too late to recall 2.13.11 if it broke anything, but if it *did*
>> break something, the sooner we find out about it, the faster we can
>> fix it.
>>
>> More info in the Contributor's Guide 7.4 Checking and verifying issues
>> http://lilypond.org/doc/v2.13/Documentation/contributor/checking-and-verifying-issues
>>
>>
>> After that, it would be nice if somebody could check all the regtests.
>
> 7.6 Finding the cause of a regression
>
> Please, which job, _what_ can i do?-) What should i do to "identify"
> "problematic commit"?

Sorry, I was unclear -- I'm not asking random people to find the
*cause* of a regression.  Just find a regression.

1.  Comparisons: go here, and select the version you want to check:
http://lilypond.org/test/
for example, v2.13.11-1

You'll see the difference in regtests between v2.13.10 and v2.13.11.
For any change, ask yourself if it's a change for the worse or better.
 (or just no real change at all)

2.  Checking all regtests:
http://lilypond.org/doc/v2.13/Documentation/web/development
and find the link for "regression tests"  (warning: downloads a large
number of images, so be patient)

Each regtest has a short description, then an image.  Does the image
match the description?  If not, send an email to bug-lilypond like

"In the ancient-accidental.ly regtest, the accidentals for hufnagel
and vaticana look exactly the same.  Shouldn't they be different?"

(this isn't true in 2.13.11, but if they *did* look the same, it would be a bug)

> (Can this regtests checking be automated?.. Yes, i do have
> bash/sed/awk/python and some idea how to use them, it this can help.
> Yes, i'm lazy too, but not so delicious and bad, probably.)

The regtest comparison is automated -- humans only need to check on
average a dozen images, instead of 300 or whatever.

Cheers,
- Graham




reply via email to

[Prev in Thread] Current Thread [Next in Thread]