Re: [PATCH 2/3] merge-recursive: fix logic ordering issue

From: Johannes Schindelin <johannes.schindelin@gmx.de>
Date: Jan 5, 2018, 21:22 UTC
Message-ID: <nycvar.QRO.7.76.6.1801052221210.1337@wbunaarf-fpuvaqryva.tvgsbejvaqbjf.bet>
In-Reply-To: <20180105202001.24218-3-newren@gmail.com>

Hi Elijah,

On Fri, 5 Jan 2018, Elijah Newren wrote:

Show 11 quoted lines

> merge_trees() did a variety of work, including:
>   * Calling get_unmerged() to get unmerged entries
>   * Calling record_df_conflict_files() with all unmerged entries to
>     do some work to ensure we could handle D/F conflicts correctly
>   * Calling get_renames() to check for renames.
> 
> An easily overlooked issue is that get_renames() can create more
> unmerged entries and add them to the list, which have the possibility of
> being involved in D/F conflicts.  So the call to
> record_df_conflict_files() should really be moved after all the rename
> detection.

I cannot actually see how the *current* rename detection would result in D/F conflicts, as it is really only about detecting the renames, the conflicts should have happened earlier.

But I could imagine that your "detect directory renames" patch series changes this behavior.

In any case, the reordering does not hurt.

Ciao, Dscho