[prev in list] [next in list] [prev in thread] [next in thread] 

List:       git
Subject:    Re: [PATCH v1 0/4] advice: remove usage of `advice_*` global variables
From:       Ben Boeckel <mathstuf () gmail ! com>
Date:       2021-08-03 0:45:36
Message-ID: YQiRsEzF8psNTqRR () erythro ! dev ! benboeckel ! internal
[Download RAW message or body]

On Tue, Aug 03, 2021 at 00:15:06 +0200, Johannes Schindelin wrote:
> On Fri, 30 Jul 2021, Ben Boeckel wrote:
> > When looking at global variable usage for my `branch.default*` settings,
> > I found the `advice_` variables which were simple enough to resolve.
> 
> Even better, it concludes the journey started in c4a09cc9ccb (Merge branch
> 'hw/advise-ng', 2020-03-25).
> 
> I reviewed the entire series and left a few comments I believe to be
> constructive.

Thanks, they've been helpful. I'll work on updating my topic in the
coming days.

> Since patch 2/4 resolves a problem introduced by merging divergent changes
> (one adding `advice_settings`, the other adding two entries to
> `advice_config`), an obvious concern with this patch series is: How can we
> guarantee that we're not introducing a similar problem when removing
> `advice_config`? A future branch could easily add entries to that array,
> and a merge of this here topic could potentially forget to add those
> entries to `advice_settings`.
> 
> However, such a future merge would always cause merge (add/remove)
> conflicts in the `advice_config` array, i.e. it will be much easier to
> notice such a divergence, and hence it will be much more likely that the
> `advice_setting` array will be adjusted accordingly in such a hypothetical

Yes, Git should be able to notice a conflict of this nature.

Thanks,

--Ben
[prev in list] [next in list] [prev in thread] [next in thread]