Re: [PATCH 0/4] remote-hg: more improvements

Felipe Contreras [off-list ref] writes:

Junio C Hamano wrote:

quoted

In other words, I knew that you are capable enough to track down a
bug in the code you wrote recently that made it violate the
expectation you defined in your own tests.

Wrong. The code in question was not recent, it was introdued in 1.8.3,
more than one year ago.

And wrong, it didn't violate the expectation of my own tests.

The code was simply not exercised in the tests.

quoted

There was no room for differences of opinions to come into play, as it
was just between you and your own code.

OK, I misread the blame output---sorry about that.

But that does not change the fact that your tests caught a bug in
your own code, and the issue was solely between you and your own
code without involving criticism from anybody, does it?

Unless you count a barf from a rather old version of Mercurial as a
criticism, that is.

quoted

Why would I expect otherwise?

Because most people take attacks on their code as personal attacks, and
they don't fix bugs in their code if they don't like the person
reporting it.

But you know I don't take attacks on my code and ideas personally, which
is more that can be said of most people on the list.

Just to make sure new people who may be watching with popcorns in
their hand from sidelines do not get a wrong impression, I do not
share your "most people take attacks ..." observation.

In reviews I have seen over the years around here (and also reviews
at $DAYJOB), I rarely saw such a reaction by the person whose change
is reviewed.  I view this list as very cooperative and productive
environment most of the time.

In any case, there was not even any attack---it was merely your code
not passing your own test on a platform you did not have access to,
which is not something to be upset about.

I don't want to do anything for a "contrib" tool.

It's already broken in v2.0 anyway.

Yes, this is not even an old regression.  If you no longer want to
have it in contrib/, I can drop it in future releases (but not in
v2.0), so that people can find the latest and greatest directly from
you.  Otherwise, queuing a fix on 'pu' and then to 'next' in
preparation for an early graduation for the release after v2.0 (and
as a fix, it may want to go to older maintenance releases) is also
fine by me.

Junio C Hamano wrote:

quoted

I don't want to do anything for a "contrib" tool.

It's already broken in v2.0 anyway.

Yes, this is not even an old regression.

Yes it is. It has nothing to do with with Mercurial v3.0, that's a
separate issue. We've been doing a workaround since v1.8.3, and that
workaround will break things in v2.0.

I already said this multiple times, but let me be clear once more:

MASTER HAS A REGRESSION (for all versions of Mercurial).

If you no longer want to have it in contrib/, I can drop it in future
releases (but not in v2.0), so that people can find the latest and
greatest directly from you.  Otherwise, queuing a fix on 'pu' and then
to 'next' in preparation for an early graduation for the release after
v2.0 (and as a fix, it may want to go to older maintenance releases)
is also fine by me.

Are you saying that the graduation plan is going to continue and they
are going to move out of contrib and be distributed by default?

If that's the case I'll resume the fixes because the current sitution is
not good.

-- 
Felipe Contreras