cfrees at imapmail.org
cfrees at imapmail.org
Sun Jul 11 03:18:08 CEST 2010
On Sun 11th Jul, 2010 at 00:04, Karl Berry seems to have written:
> What if the new glyph is named Copyleft instead of copyleft?
I would do one of two things:
1) if Copyleft occurred very rarely (ie not an issue likely to affect
many fonts), I'd put the Copyleft in a font-specific encoding file and
2) I'd copy ts1.etx or ts1-euro.etx to ts1-copyleft.etx or
ts1-euro-copyleft.etx if lots of fonts used Copyleft rather than
copyleft and edit the file as I did for the euro/Euro case.
I can confidently say that I would not change copyleft to Copyleft in
> I.e., the
> whole problem with euro vs. Euro can repeat ad infinitum, with any glyph
> in any font at any time. Or new glyphs might be created which are
> conveniently added in an empty slot, not in a whole new font. In
> general, pretty much every font I know of which started off using
> "common" definitions eventually ended up diverging and needing its own
Mostly, I'm using ts1-euro.etx in conjunction with font-specific
encoding files. Only if all of the glyphs needed for TS1 and available
in the font have the names TS1 expects, except for the Euro/euro, do I
use ts1-euro.etx for the TS1 encoding on its own.
> In any event, I suppose it ultimately depends on how careful you want to
> be. If you are confident you will never change ts1-euro in the future,
> then there's no technical problem with just using it and repeating it as
> many times as you like. As long as the files are identical, it doesn't
> matter which one is found, so (true) duplicates aren't a problem.
I don't really have to be careful. Not changing it is simply part of
the way I've set things up. I have certain .etx files which are
"standard" (customised but standard) which I reuse for new fonts but
don't change in ways which would change the corresponding .enc files. I
keep these entirely separate from the font-specific .etx files. If I
update the support for a font and the glyph names have changed, I'd
edit the font-specific .etx files and regenerate the .enc files, but I
wouldn't change the .etx files I've put into the "standard" group. (I
might copy them and create additional files for the standard group or I
might copy them and turn the copies into font-specific encoding files.
But then I'd rename them.)
Maybe I've just set things up oddly, but being careful really doesn't
come into it.
> I still think each font should have its own copy to avoid dependencies.
> Getting back to your original point, I think that duplication is better
> than having a separate package with the common files, because
> package-to-package dependencies in TL are not perfect and not
> Up to you. You asked for advice, well, that's my advice :).
OK. Thanks. I might have to think about it.
> I would have thought it would make sense for that to be the same
> across fonts just because the encoding is the same.
> I was talking about the name, not the definition. Yes, the glyph names
> in the definition have to match whatever the fonts have. So if the
> fonts use the same glyph names, the enc definition will be the same too.
> (And, more likely, if the fonts use different glyph names, the enc
> definitions have to be different.)
I also meant the name :). I was thinking a 1-1 mapping from names to
definitions was most straightforward even though a many-1 mapping would
If the /fontinst... name is changed in the map and enc file, is that
sufficient? Do the vf and tfm files also contain references? (The
pl/vpl files do but I'm not sure if they're just comments.)
More information about the tex-live