[gs-devel] [RFC] revive runlibfileifexists function. xefitra

mpsuzuki at hiroshima-u.ac.jp mpsuzuki at hiroshima-u.ac.jp
Thu Nov 22 05:22:41 PST 2007

On Thu, 22 Nov 2007 21:51:23 +0900
"CHIKAMA Masaki" <masaki.chikama at gmail.com> wrote:
>> The CIDFnmap implementation in original gs-cjk was designed
>> to provide F1 (so the case of missing reference was considered
>> by design), but I don't think cidfmap was designed so.
>> The refering of other cidfmap file is not designed but
>> occasionally possible at present, so fallbacking is not
>> designed.
>I'm surprised to hear this. Then are lib/Fontmap and lib/FAPIcidfmap
>also not designed to read other files from them?

The original syntax of lib/Fontmap was designed long ago,
its syntax is different from PostScript. I guess, the author
of lib/Fontmap may think about the reference of external file
at that time. But FAPIfontmap, FAPIcidfmap, cidfmap were
introduced since gs8.x, I think the idea of Fontmap author
was not carried over.

>The Fedora (one of Linux distribution, you know) uses this mechanism
>since about 2.5 years ago on ghostscript 8.15.

Yes, it was occasionally possible to refer other cidfmap
file from a cidfmap. It was occasionally possible. If
you read the documentation stored in doc/, you will find
that such mechanism is not described even for Fontmap.
I remember Fedora maintainer wrote a patch to avoid the
crash caused by reference of missing cidfmap.

>If it is not designed and there is no plan to design,
>I have to think another way.

I guess Artifex people may not be against the idea
(refering other cidfmap in a cidfmap, etc) itself.
The refering of cidfmap seems to be well restricted
purpose and its side effect won't be serious.

What I had asked and refused was more generic mechanism
to use Koji Otani's patch via user configuration.


More information about the gs-devel mailing list