init: API

Nicolas Mailhot nicolas.mailhot at laposte.net
Sun Nov 20 11:52:07 UTC 2005


Le dimanche 20 novembre 2005 à 01:11 +0000, Dave Mitchell a écrit :
> On Sun, Nov 20, 2005 at 01:55:47AM +0100, Nicolas Mailhot wrote:
> > Le samedi 19 novembre 2005 à 22:59 +0000, Dave Mitchell a écrit :
> > 
> > > This is a bit hand-wavy because I'm not familiar with the font.conf
> > > schema, but for example
> > > 
> > >     substitute name=family /mono/ "monospace"
> > 
> > Perfect example of subtle flat-file syntax no one can understand without
> > the man page under its eyes. Push a few more lines like this, each with
> > its own rules hidden in magic characters like /, and you get something
> > pretty much impossible to administer (let alone write a robust parser
> > for)
> 
> And you'd be able to understand the XML version without having the
> "part of the schema for font name substitutions within font.conf" manual
> in front of you?

Yes, I can understand that :

        <match target="pattern">
                <test qual="any" name="family">
                        <string>mono</string>
                </test>
                <edit name="family" mode="assign">
                        <string>monospace</string>
                </edit>
        </match>

means

« For a font which got the string mono in its family metadata field
(qual="any" is not defined but I'd guess any means as global as you can)
assign the string monospace as the family field »

And this is only by reading the xml tags (pattern with test + associated
action), without going into the schema or anywhere else. Of course you
need to know beforehand fonts declare metadata, which include "family". 

In your example I have to guess the second field is a pattern, the third
is the substitution, it does not permit changing one field based on
another, you've used space as separator which will bite you later, etc
etc.

> XML is good for text markup. It's really crap for anything else.
> (IMHO of course).

Except configuration files need to be readable and explicit tags are way
better than magic "//" expressions for this (or implicit arguments based
on ordering)

Regards,

-- 
Nicolas Mailhot
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 197 bytes
Desc: Ceci est une partie de message num?riquement sign?e
URL: <http://listman.redhat.com/archives/fedora-devel-list/attachments/20051120/e6f78d45/attachment.sig>


More information about the fedora-devel-list mailing list