On 16 May 2010 14:35, Karl Hammar <[email protected]> wrote:

> Does it matter, if not, why is in the contributor manual ?

I think the instructions were lifted directly from the old (2.10)
docs, so there may well have been changes to the test suite since they
were written.  TBH, when I started doing regression test checks, I
bumbled along until I found something which worked.

The important point is this comment:

## redo files differing from baseline

If you haven't done `make check' after applying a patch, then the only
files which will be retested will be the ones showing up in the
initial test.

I think the point of `make test-redo' is to speed up the process when
working on a patch, since it only rebuilds the snippets which have
changed.

> It has an "make check" after the test-baseline which you don't have, is
> it redundant?

I've never followed that step; it seems like complete waste of time.

> git-apply and patch -p1 produces the same result, so that is not the
> cause of our different output:

Correct.

> Isn't the diff a complete statement of what you want to change?
> Also you *did* cite the diff.

Yes, but it was easier for me to pluck the relevant parts out than
link to the individual files on Rietveld.

Cheers,
Neil

_______________________________________________
lilypond-devel mailing list
[email protected]
http://lists.gnu.org/mailman/listinfo/lilypond-devel

Reply via email to