Bug 687775

Summary:	relative paths in cidfmap
Product:	Ghostscript	Reporter:	Igor Melichev <igor.melichev>
Component:	Resource	Assignee:	Alex Cherepanov <alex>
Status:	RESOLVED WORKSFORME
Severity:	enhancement	CC:	christinedelight.top85, gsview, htl10
Priority:	P4
Version:	master
Hardware:	All
OS:	All
Customer:		Word Size:	---

Description Igor Melichev 2004-10-26 04:16:01 UTC

When specifying a CID font emulation with a True Type file in lib/cidfmap, a 
relative file path would be comfortable. Currently we require an absolute path. 
This request is moved from Bug 687767.

I suggest to implement a new special command line option for this purpose.
Alternatively we can put it into cidfmap as a special record type. Don't know 
which way is better.

Comment 1 leonardo 2006-09-07 06:10:18 UTC

Folks, please bring suggestions how to do ths better.

Comment 2 Hin-Tak Leung 2006-09-07 09:58:18 UTC

I agree -I is a bit overloaded.

How about using GS_FONTPATH properly? but I hate having configurations in two
places - so ideally one should be able to specify that in the cidfmap file as
well? e.g.

(/some/extra/place/to/look/for/fonts) setextrafontpath
... relative path entries ...
<< ... /Path (msmincho.ttc) ...>> % search under the extra place

Comment 3 leonardo 2007-05-21 01:34:05 UTC

Passing to Alex since he owns gs\lib.
Uniting with GS_FONTPATH isn't a priority because it will need a font file 
prescan to distinguish fonts from CID fonts.

Comment 4 Alex Cherepanov 2011-11-27 03:45:55 UTC

The string specified by /path attribute is used as:
    dup /Path get (r) file
So the relative path is supported.

cidfmap file is interpreted as a PS program. It's trivial to
append strings if so desired.

The stock cidfmap file has an example that uses getenv and concatstrings.

Comment 5 Hin-Tak Leung 2011-11-27 21:26:47 UTC

(In reply to comment #4)
> The string specified by /path attribute is used as:
>     dup /Path get (r) file
> So the relative path is supported.

How does this interact with the new security-related default (-dSAFER equivalent)? Documentation probably could be updated to indicate how this is supposed to work.