On Thu, Sep 19 2013, Mark Walters <markwalters1009@gmail.com> wrote: > I have looked at this but haven't fully worked out what is going on. > > Tomi found (on irc) that putting a (select-active-regions nil) to line > 212 of notmuch-mua.el solves the problem. You found that it was select-active-regions what to try ... > My testing shows that putting a (let ((select-active-regions nil)) > around the lines 224-227 in notmuch-mua.el also solves the problem. > > However, when trying to narrow the problem I see some odd behaviour: if > you put the let just around the message-cite-original line then the > following happens: > > if the primary selection is in some non-emacs window (eg an xterm) then > the bug does not show up but if the primary selection is in the emacs > window and then you initiate the reply then the bug does show up. > > We could just try one of the above two fixes but it could be papering > over something. Yes, such duct-taping something is probably not the best approach; we could use that as a "temporary" solution but you know what happens with these "temporary" solutions. Also, doing these ad-hoc testing switching between emacs 23 & 24 and something doing "mouse-draggin" or "shift-selection" (*) and other times C-SPC and then moving cursor (and forgetting which approach is tried get things confusing). (*) mouse-dragging & shift-selection taken from the documentation string of variable select-active-regions. It seems that 'only is the value I personally want to use there.... ... have to try the 'only option -- and also disabling transient-mark-mode -- how to do that is a question (the docstring is a bit confusing and it seems although I have (transient-mark-mode t) it is enabled -- how to know transient-mark-mode is enabled is also a bit of mystery to me :/ Forget my mml-quote-reqion -comment in my previous mail. although the diagnose was wrong the cure was working ;/ > Best wishes > > Mark Tomi > > > > > On Wed, 18 Sep 2013, Tomi Ollila <tomi.ollila@iki.fi> wrote: >> On Wed, Sep 18 2013, Jameson Graef Rollins <jrollins@finestructure.net> wrote: >> >>> I've just started noticing that when I reply to messages from the emacs >>> UI, my X clipboard is filled with the body of the reply message, >>> displacing whatever was in there previously. I'm not sure if this is >>> related to my personal emacs configuration, or some non-ideal way we're >>> inserting body parts in replies. However, even if it is a configuration >>> issue in my emacs, I suspect that there should be a better way to insert >>> the reply body such that there's not even a possibility that it would >>> leak into the users X clipboard, regardless of emacs settings. Any >>> thoughts? >> >> env -u DISPLAY emacs -f notmuch ;) >> >> >> .. OK, I'm marking this as a bug -- I witness the same behaviour. >> >> Originally I thought this relates directly to the fact that mark >> is set at the end of quoted text and (point) is at the beginning of >> buffer -- so that simple c-w can be used to remove the contents in >> case desired... but the X paste provides the original text, not the >> quoted one. >> >> Someone™ has to dig (deeper) into the code for this... >> >>> jamie. >> >> Tomi >> >> _______________________________________________ >> notmuch mailing list >> notmuch@notmuchmail.org >> http://notmuchmail.org/mailman/listinfo/notmuch > _______________________________________________ > notmuch mailing list > notmuch@notmuchmail.org > http://notmuchmail.org/mailman/listinfo/notmuch