Re: What's cooking in git.git (Feb 2019, #03; Wed, 13)

From: Junio C Hamano <gitster@pobox.com>
Date: Feb 14, 2019, 05:49 UTC
Message-ID: <xmqqzhqy9axy.fsf@gitster-ct.c.googlers.com>
In-Reply-To: <20190214035944.GB7209@sigill.intra.peff.net>

Jeff King <peff@peff.net> writes:

Show 19 quoted lines

> On Wed, Feb 13, 2019 at 07:33:22PM -0800, Junio C Hamano wrote:
>
>> [Graduated to "master"]
>> [...]
>> * js/smart-http-detect-remote-error (2019-02-06) 3 commits
>>   (merged to 'next' on 2019-02-06 at ec1a6f67c1)
>>  + t5551: test server-side ERR packet
>>  + remote-curl: tighten "version 2" check for smart-http
>>  + remote-curl: refactor smart-http discovery
>> 
>>  Some errors from the other side coming over smart HTTP transport
>>  were not noticed, which has been corrected.
>
> Hmm, this was maybe a risky one to merge as part of -rc1. It is just
> enforcing the rules from the spec document, but it's possible that it
> may trigger on some poorly-behaved server.
>
> I'm not worried enough to suggest pulling it back, but it's something we
> should keep an eye on during the -rc period.

True. I actually think that it is probably a good idea to unleash this kind of thing as quickly as possible to the real world. Those outside the project, or even inside it, tend to only discover issues after a feature release is made, so while I am generally in favor of merging a new and potentially disruptive change in an early batch after a feature release to 'master' (which would give us enough time to fix things up until the next release), I do not think that approach would help third-party interop issues very much.