From: Mark Walters <markwalters1009@gmail.com>
To: Austin Clements <amdragon@MIT.EDU>, notmuch@notmuchmail.org
Subject: Re: [PATCH 0/8] Improve tag change completion
In-Reply-To: <1382471457-26056-1-git-send-email-amdragon@mit.edu>
References: <1382471457-26056-1-git-send-email-amdragon@mit.edu>
User-Agent: Notmuch/0.16 (http://notmuchmail.org) Emacs/23.4.1
	(x86_64-pc-linux-gnu)
Date: Tue, 22 Oct 2013 22:43:53 +0100
Message-ID: <87mwm1x9pi.fsf@qmul.ac.uk>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Precedence: list


This looks good to me +1. It makes the code clearer and nicer to read as
well as giving a better user experience, and it is makes fixing the long
standing tagging races simpler.

I have a couple of docstring comments:

In patch 2 perhaps notmuch-tag-completions could have a docstring.

In Patch 4 I think the docstring for notmuch-search-tag is outdated: it
 is "Change tags for the currently selected thread or region." but 
 beg and end can now be specified by the caller.

and one actual comment:

in patch 3 (for show) delete-dups is called before the list is passed to
notmuch-read-tag-changes whereas it is not for search or pick.
Obviously this is not actually a problem but it might be worth being
consistent.

But that was all I found. All tests pass and everything I try behaves
exactly as expected.

Best wishes

Mark


On Tue, 22 Oct 2013, Austin Clements <amdragon@MIT.EDU> wrote:
> This series improves tag change completion in various ways for
> commands like +, -, and *.
>
> From a user perspective, this provides command-specific prompts like
> "Tag message" and "Tag all" instead of the generic "Tag" prompt, and
> bases tag removal completions on the tags that are in the buffer,
> rather than the current tags in the database, providing a more
> predicable experience.
>
> From an implementation perspective, this new tag removal completion
> behavior improves efficiency and eliminates a road block to fixing the
> tagging race bug (which otherwise results in massive queries just to
> compute removal completions).  The new code is also more "Elispy" and
> predictable because all tag change prompting now occurs at the
> interactive entry points, rather than buried under several layers of
> non-interactive calls.
>
> This is a spiritual successor to
> id:1354263691-19715-1-git-send-email-markwalters1009@gmail.com, though
> it takes a very different approach.  This is also a prerequisite to
> the tag race fix in
> id:1381185201-25197-1-git-send-email-amdragon@mit.edu and I plan to
> send an updated version of that series when this one is accepted.
>
> Patches 1, 5, and 6 could be pushed on their own.  They fix bugs or
> sort of bugs that get in the way of the rest of the series.
>
> _______________________________________________
> notmuch mailing list
> notmuch@notmuchmail.org
> http://notmuchmail.org/mailman/listinfo/notmuch