Re: What tests are run by patchy?

[Frédéric Bron]

> I would at least normally expect that you do a 'make' to see that it builds
> against current tree - sometimes on the morning of the PATCH countdown, that
> is hard because someone may push a patch and the new tree is merged before
> the test scripts are run and you submitted a patch based on the tree before
> the push. That can cause make problems, but those are usually rare.

before posting a patch, I always rebase against the most recent master.

> Doing a make check is quite time consuming for some users and slow, so as
> long as you don't mind the potential wait for your patch to be tested (by me
> - others also run patchy but may not do the full make doc) think of it like
> just 'good manners' to do at _least_ a make ;)

of course I check make!
In particular, what could arise is that I change foo(string s) to
foo(const string &s) in derived class and not in base class. This
would cause bad behaviour.
I am very careful for this but of course, it is better to have tests.
Also I check for compiler warnings but lilypond builds with a lot of
compiler warnings which does not help. I think we should try to remove
all warnings.

Frédéric