Web lists-archives.com

Re: [PATCH 3/3] tests: optionally skip `git rebase -p` tests




Johannes Sixt <j6t@xxxxxxxx> writes:

> Am 01.11.18 um 07:12 schrieb Junio C Hamano:
>> "Johannes Schindelin via GitGitGadget" <gitgitgadget@xxxxxxxxx>
>> writes:
>>
>>> The `--preserve-merges` mode of the `rebase` command is slated to be
>>> deprecated soon, ...
>>
>> Is everybody on board on this statement?  I vaguely recall that some
>> people wanted to have something different from what rebase-merges
>> does (e.g. wrt first-parent history), and extending perserve-merges
>> might be one way to do so.
>
> Maybe you are referring to my proposals from a long time ago. My
> first-parent hack did not work very well, and I have changed my
> workflow. --preserve-merges is certainly not a feature that *I* would
> like to keep.

Thanks, that reduces my worries.

> The important question is whether there are too many users of
> preserve-merges who would be hurt when it is removed.

Yes, and the claim this series makes is that there is none and all
existing users should be able to happily use the rebase-merges,
which also means that we need to commit to improve rebase-merges to
support them, if there were some corner cases, which we failed to
consider so far, that are not yet served well.

As I said, as long as everybody agrees with the plan (e.g. we'll
know when we hear no objections to the planned deprecation in a few
weeks), I am perfectly OK with it.

Thanks.