Re: [O] Bug? Encoding trouble in org-id-locations-load

2017-11-04 Thread Nicolas Goaziou
Hello,

Anders Johansson  writes:

> I tried changing the call to ~insert-file-contents-literally~ in
> ~org-id-locations-load~ to just ~insert-file-contents~ and this seemed
> to fix the behaviour (as correct decoding is done then?). I don’t know
> if there are other unwanted effects from using ~insert-file-contents~,
> but otherwise this seems to me to be a more correct solution.

Agreed. I applied your suggestion. Thank you.

Regards,

-- 
Nicolas Goaziou



Re: [O] Bug? Encoding trouble in org-id-locations-load

2017-11-03 Thread Jorge Morais Neto
On 3 November 2017 at 12:11, Anders Johansson  wrote:
> I use org-id and got some surprising reports of duplicate IDs. It seems that
> the issue is that one of my files containing IDs has a filename consisting
> of some non-ascii characters (a Swedish ä).

This may be related to two problems I had with
~org-icalendar-combine-agenda-files~.  When I invoked that command,
Emacs created some duplicates of buffers for Org files with names
containing non-ASCII Latin-1 characters.  In the names of the
duplicates, each non-ASCII Latin-1 character was replaced by two octal
escape sequences (probably the character code in UTF-8).

The second problem was Emacs interrupting the export to ask me which
character encoding to use for certain "problematic characters".
Strangely, the "problematic characters" were octal escape sequences
which I verified to be correct UTF-8 codes of non-ASCII Latin-1
characters.  Emacs was apparently trying to "reencode" "text" that was
actually UTF-8 code.  Each time I invoked
~org-icalendar-combine-agenda-files~ and Emacs asked that, I chose raw
text encoding so the UTF-8 codes would be left as is.

I finally decided to investigate and traced both problems to some id
links pointing to headlines in files with names containing non-ASCII
Latin-1 characters.  I worked around by setting Org Mode not to export
those links.

Regards

-- 
- I am Brazilian.  I hope my English is correct and I welcome feedback
- Please adopt free formats like PDF, ODF, Org, LaTeX, Opus, WebM and 7z
- Free (as in free speech) software for Android: https://f-droid.org/



[O] Bug? Encoding trouble in org-id-locations-load

2017-11-03 Thread Anders Johansson

Hi,
I use org-id and got some surprising reports of duplicate IDs. It 
seems that the issue is that one of my files containing IDs has a 
filename consisting of some non-ascii characters (a Swedish ä). 
When this filename is read in from ~org-id-locations-file~ in 
~org-id-locations-load~ the ä is interpreted as “\303\244”.
But ~org-id-files~ and ~org-id-locations~ is also populated from 
currently open files, so I usually get that file represented twice 
as two different files in ~org-id-files~. So I get both 
“j-allmänt.org” and j-allm\303\244nt.org“. Both seem to be scanned 
correctly by ~org-id-update-id-locations~ and this results in 
duplicate IDs.


I tried changing the call to ~insert-file-contents-literally~ in 
~org-id-locations-load~ to just ~insert-file-contents~ and this 
seemed to fix the behaviour (as correct decoding is done then?). I 
don’t know if there are other unwanted effects from using 
~insert-file-contents~, but otherwise this seems to me to be a 
more correct solution.


Cheers,
Anders Johansson