Re: [Monotone-devel] [PATCH] _MTN/log pre-specified magic line quickie

[Ben Walton]

I actually looked at implementing this inside the hook, but chose the
c++ route for the following reason.  I may be off base here with my
reasoning, so let me know if you agree/disagree:

Doing this in the c++ code leaves a hook writer the responsiblity of
explicitly removing the 'magic' line if they wish to disable the
feature.  Implementing this in the lua hook leaves the hook writer the
implicit resposibility of providing the feature if they override the
hook.  Now in reality, the difference is small.  I just felt that from
a UI perspective, it's better to make a user responsible for explicit
rather than implicit actions.

I see that the mkdir patch is failing too, at line -1, which is odd.

If tests need updating to work with the 'magic' line, I don't mind
doing that, if the code is sound.

Thoughts?
-Ben

On 12/15/06, Nathaniel J. Smith <address@hidden> wrote:
> On Mon, Dec 11, 2006 at 11:34:23AM -0500, Ben Walton wrote:
> > 2006-12-11  Ben Walton  <address@hidden>
> >        * cmd_ws_commit.cc,
> >        tests/commit_using__MTN_log/commit_log_modified_return.lua,
> >        tests/commit_using__MTN_log/__driver__.lua:
> >        Added the requirement that when _MTN/log is pre-specified by
> >        the user, that the user remove a 'magic' line prior to the
> >        commit being accepted.  Modified the units tests to verify this.
> >        Updated texinfo to match this new behaviour.
>
> It looks like this broke a few tests; e.g. search for "FAIL" in
>   http://venge.net:9000/linux_suse_9.3_x86_unopt/builds/50/test/0
> They mostly look like the tests need updating, rather than that the
> actual code is broken.
>
> I'm not sure how annoying it will be to fix some of them; I almost
> wonder if the actual ******DELETE THIS LINE***** handling should move
> to be more of the hook's responsibility?  (Like we tell the hook when
> we think some extra confirmation might be needed, and the hook tells
> us when it thinks that that extra confirmation has been given.)  Just
> a random thought, though.
>
> -- Nathaniel


--
---------------------------------------------------------------------------------------------------------------------------
Ben Walton <address@hidden>

Unanswered questions are far less dangerous than unquestioned answers.
- The Roadside Pulpit
---------------------------------------------------------------------------------------------------------------------------