Carsten Dominik wrote:
> Hi Kai,
> 
> 
> On Oct 23, 2009, at 9:01 AM, Kai Tetzlaff wrote:
> 
>> Hi,
>>
>> i noticed that when using the org-mode clock persistence, the stored
>> clock data gets deleted when i start emacs and exit again without
>> turning on org-mode in between.
>>
>> When looking at org-clock-persistence-insinuate it looks like org-clock
>> load will only run after org-mode gets started whereas org-clock-save
>> will always be called when exiting emacs:
>>
>> (defun org-clock-persistence-insinuate ()
>>  "Set up hooks for clock persistence"
>>  (add-hook 'org-mode-hook 'org-clock-load)
>>  (add-hook 'kill-emacs-hook 'org-clock-save))
>>
>> Not running org-mode-hook (i.e. not starting org-mode) thus does not
>> load clock data but org-clock-save overwrites any prviously saved data
>> when exiting emacs.
>>
>> An easy fix for that would be to just add org-clock-load to e.g.
>> emacs-startup-hook. But this will only work if the code in
>> org-clock-load does not depend on any org-mode initialization code (or
>> would require loading org-mode).
>>
>> So org-clock-save should probably check if org-clock-load has been
>> running during the current emacs session (or if clock persistence was
>> just enabled) and only then save clock data when exiting emacs. I tried
>> to add this to the code in org-clock-save:
>>
>> diff --git a/lisp/org-clock.el b/lisp/org-clock.el
>> index c7ebbf8..c0fe4e6 100644
>> --- a/lisp/org-clock.el
>> +++ b/lisp/org-clock.el
>> @@ -1803,7 +1803,8 @@ This function is made for clock tables."
>>   "Persist various clock-related data to disk.
>> The details of what will be saved are regulated by the variable
>> `org-clock-persist'."
>> -  (when org-clock-persist
>> +  (when (and org-clock-persist
>> +             (or org-clock-loaded (not (file-exists-p
>> org-clock-persist-file))))
>>     (let (b)
>>       (with-current-buffer (find-file (expand-file-name
>> org-clock-persist-file))
>>     (progn
>>
>>
>> /Kai
>>
> 
> I see you point, but I am not convinced that you fix is the right one. 
> What one really wants to do is save the clock and history if and only if
> the clock has been used in the current session.  So I guess it is better
> to create a special variable org-clock-used-in-this-session which then
> will be set to t each time clock-in starts a clock somewhere.  I believe
> this would be a better insurance policy.
>
> Would you like to try your hand at a patch in this spirit?

Ok, i've attached a new path which uses the suggested approach. I wasn't
exactly sure about where to set org-clock-used-in-this-session to t.
It's now done right at the end of org-clock-in, just before running
hooks in org-clock-in-hook.

I've done a couple of (simple) tests and it works fine for me.

/Kai

> - Carsten
> 
>> _______________________________________________
>> Emacs-orgmode mailing list
>> Remember: use `Reply All' to send replies to the list.
>> Emacs-orgmode@gnu.org
>> http://lists.gnu.org/mailman/listinfo/emacs-orgmode
> 
> - Carsten
> 
> 
> 
>