Re: Proposal: move write-contents-functions higher up in basic-save-buffer

[Eric Abrahamsen]

Stefan Monnier <address@hidden> writes:

> I haven't looked at basic-save-buffer recently, but in the worst case we
> could keep the current code and add a
>
>     (if (null buffer-file-name)
>         (run-hook-... 'write-contents-functions)
>       ...)
>
> but admittedly, it's better if we can move the single call to
> write-contents-functions so it's shared by the file and the
> non-file cases.

Eli Zaretskii <address@hidden> writes:

>> From: Eric Abrahamsen <address@hidden>
>> Date: Wed, 24 May 2017 12:55:53 +0800

>> I can try to produce a patch, if this is acceptable in principle.
>
> Yes, please.

Richard Stallman <address@hidden> writes:

> Please try using write-contents-functions for this, and you'll see if it
> does the job.  If not, you'll see what more change is needed.

Okay, here's a first stab at it. I think it should work correctly: all
the short-circuit hooks get a chance to run in all cases, but the
function only insists on the presence of a file if
`write-contents-functions' are not present, or if they fail with a nil
value. I'd like to specify in the docs that those functions should fail
with an error.

If this looks okay I'll spend a bit more time testing it, then make
docstring and manual edits.

Eric

0001-First-whack-at-write-contents-functions-for-non-file.patch
Description: Text Data