[Fontconfig] Multiple values in <test> - Simple Solution?

Raimund Steger rs at mytum.de
Mon Sep 3 16:23:17 PDT 2012

Infinality wrote:
> Hi,
> I have some configurations that make gratuitous use of multiple values
> in <test>, to group certain families of fonts together, as shown below.
> My understanding is that the only "correct" way to do this now is to
> have an entire
> <match><test><string>NAME</string></test><edit></edit></match> for every
> single family!  Is there any simpler way to do what I would like?  If
> not, is it planned?

fontconfig's behavior is unchanged; only the warning is new. There have 
been a few discussions on this I think, ones that I recollect are 
http://thread.gmane.org/gmane.comp.fonts.fontconfig/3741 and 
http://thread.gmane.org/gmane.comp.fonts.fontconfig/3991, including the 

Multiple families in <alias> or <match> only work as expected if you do 
not use the position of the matching value for <edit>'s, i. e. only use 
the <edit> modes prepend_first or append_last or edit another property. 
For other modes, the position where the edit is applied is erratic. That 
was the rationale behind the warning.

Since that frequently came up I think the change was a good idea and 
I've written my configs in a way that I use individual <match> blocks. 
(I only assign some marker property in these blocks and don't do other 
edits so that I don't have to repeat them every time.)

As to whether there is The Solution planned for this, I don't see it... 
The problem is you can have any number of matching positions in such 
cases, and then where would you apply the edit? What would you do for 
qual's other than "any"? It just wouldn't make sense. OK maybe the 
warning could be more precise and only warn if the match position is 
used in the edit and/or if there is more than one actual match per 
property for a given pattern...


Worringer Str 31, Duesseldorf 40211 Germany   home: <rs at mytum.de>
+49-179-2981632 icq 16845346                  work: <rs at interface-ag.de>

More information about the Fontconfig mailing list