vm-delete-postponed-message is too heavy handed

[Mark Diekhans]

I have recently lost copys messages being composed in steps using vm-pine . I believe this is caused by somethingcausing vm-delete-postponed-message to be run.

vm-delete-postponed-message does

                (vm-expunge-folder)
                (vm-save-folder)

This seems very heavy handed to purge all deleted message and then save the folder. It seems far better behavior
to just mark the message as deleted.

Happy to submit a patch, however I would like to understand the motivation of expunging. I would just like to drop these two lines, or at very least, the expunge

[Mark Diekhans (markd-kermodei)]

this seems to be related to this comment, however it's not clear why this is true:

;; If you have postponed messages you will be asked if you want to continue
;; composing them, if you say "yes" you will visit the `vm-postponed-folder'
;; and you can select the message you would like to continue and press "m"
;; again! However be aware this works currently only if you expunge all
;; messages marked for deletion and save the postponed folder.

[Mark Diekhans (markd-kermodei)]

vm-continue-what-message also expunges.

[Mark Diekhans (markd-kermodei)]

This patch removes the automatic expunge of the postponed buffer. I have been using it extensively with both curses and OS/X native emacs. I can't find a reason for the expunge and assume the reason no longer eists.

[Uday Reddy (reddyuday)]

Hi Mark, I always felt that this kind of forcible expunge was a bad idea, but this is how vm-pine had been from the earliest revision in the version control.

Recall that this is called `vm-pine'. So the behaviour is expected to be the same what was in pine I suppose. Having never used pine myself, I can't say how it should work.

Incidentally, I haven't been getting the behavior you describe because there was an undefined function `frame-of-buffer' in Gnu Emacs. So, the entire expunging code was getting skipped. I am wondering how it worked for you.

[Mark Diekhans (markd-kermodei)]

Hi Uday,

My guess from the comments was that this behavior was to work around some technical restriction. However I can't find it. Email to original author was not answered.

It looks like the called to frames-of-buffer needs to be conditioned on vm-xemacs-p. I will update the patch to do this. I have no idea how it worked for me; my guess is the other conditions were not satisfied.

I have never used pine; the only reasons for using this facility is to postpone messages in a way that integrates well with VM. Is there any other approach?

[Mark Diekhans (markd-kermodei)]

updated version of patch that avoids frames-of-buffer in GNU emacs

[Mark Diekhans (markd-kermodei)]

looks like this was fix with addition of vm-auto-expunge-postponed-folder, however ticket
not updated. Am I correct? trying to test trunk.

[Uday Reddy (reddyuday)]

Fix committed in rev. 1513. (Forgot how it got done!)