Re: [O] partial-completion-mode error when refiling

[Carsten Dominik]

On Jun 30, 2011, at 2:40 PM, Nick Dokos wrote:

> Bastien <address@hidden> wrote:
> 
>> Hi Sebastian,
>> 
>> "Sebastien Vauban" <address@hidden> writes:
>> 
>>> When I was trying to refile an extract of an email, I got this:
>>> 
>>> Getting targets...done
>>> funcall: Symbol's function definition is void: partial-completion-mode
>> 
>> thanks for reporting this -- this is indeed something wrong with the fix
>> I made to `org-without-partial-completion' (see my other message to Paul
>> Sexton).  
>> 
>> I reverted his patch so you won't see this error again.
>> 
> 
> I'm not sure that't the problem though: the org-without-partial-completion
> macro is called in a couple of places, once in org-remember.el and twice
> in org.el. I'm not sure how many people still use org-remember, but I suspect
> quite a few. The macro basically says: execute the body while mmaking sure
> that partial-completion-body is off during the execution. At least, that's
> the intent but I haven't thought through the quoting change that Paul made.
> 
> The calls:
> 
> o org-remember-apply-template: called in the g or G case to complete tags.
> o org.el: in org-icompleting-read.


> o org.el: in org-set-tags *around* org-icompleting-read.
> 
> The last one seems superfluous at first sight, but I haven't thought about
> it yet.

Yes, this one is superfluous.


> 
> In any case, these seem fairly common situations so I think it is likely
> that the macro has been called hundreds of times (over the whole org 
> population)
> without ill effects.
> 
> OTOH, partial-completion-mode is called explicitly in org-refile-get-location,
> like this:
> 
>     (partial-completion-mode nil)

This is not a function-calling form, but this is part of a let form,
so it just sets the variable partial-completion-mode to nil.
In effect, this does indeed turn off partial-completion-mode for
the body of the form.

> 
> Could it be that it is really meant to turn *off* partial completion mode?
> In which case, it would be better to call the org-without-partion-completion
> macro here to do the work.
> 
> In any case, this explicit call seems to be more problematic than the macro.
> After all that's what Seb hit.
> 
> Nick
> 
> 

- Carsten