Re: What's cooking in git.git (Oct 2017, #06; Fri, 27)

From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Oct 29, 2017, 13:13 UTC
Message-ID: <alpine.DEB.2.21.1.1710291409390.6482@virtualbox>
In-Reply-To: <xmqq7evhc7nw.fsf@gitster.mtv.corp.google.com>

Hi Junio,

On Fri, 27 Oct 2017, Junio C Hamano wrote:

Show 16 quoted lines

> * cc/git-packet-pm (2017-10-22) 6 commits
>  - Git/Packet.pm: extract parts of t0021/rot13-filter.pl for reuse
>  - t0021/rot13-filter: add capability functions
>  - t0021/rot13-filter: add packet_initialize()
>  - t0021/rot13-filter: improve error message
>  - t0021/rot13-filter: improve 'if .. elsif .. else' style
>  - t0021/rot13-filter: refactor packet reading functions
> 
>  Parts of a test to drive the long-running content filter interface
>  has been split into its own module, hopefully to eventually become
>  reusable.
> 
>  Expecting a reroll.
>  The original unfortunately is quite broken in error detection
>  cases to make the result sub-par as a standalone module before it
>  gets fixed.

Would you terribly mind adding a SQUASH??? with my suggested fix for PERL_NO_MAKEMAKER?

Otherwise, the Windows build will keep failing for `pu`.

Show 7 quoted lines

> * jr/ssh-wrappers (2017-10-24) 5 commits
>  - ssh: 'simple' variant does not support --port
>  - ssh: 'simple' variant does not support -4/-6
>  - ssh: 'auto' variant to select between 'ssh' and 'simple'
>  - connect: split ssh command line options into separate function
>  - connect: split git:// setup into a separate function
>  (this branch uses bw/protocol-v1.)

Likewise, this needs a SQUASH??? to remove the uplink$X before (over-)writing uplink, to pass the test suite on Windows.

Of course, if we could eventually have Continuous Testing of individual *branches* (instead of `pu`), then we would not have to see those test failures in every effing single build of `pu`. I mean, there is no need to test things after they were tested already, and failed, right?

Ciao, Dscho