On Fri, Jul 28, 2023 at 06:39:58PM +0100, David Carlisle wrote:
> On 28/07/2023 18:25, LARONDE Thierry wrote:
> > Why in this case not having requested a \filefind primitive, with an
> > argument specifying a required extension (an empty string if to be
> > taken as is), and using the return value with other primitives (size,
> > moddate, md5, dump) doing no search and no extension (very "primitive"
> > primitives)?
>
> you are asking on a latex list, but this is not a latex issue.
>
> \pdffilesize has had this behaviour in pdftex for decades, and another
> primitive isn't needed.
>
> It does not seem surprising that all the file primitives use the same
> logic which is documented for web2c tex in the kpathsea manual section
> 6.2 so it is not specially documented in the pdftex manual where each
> file primitive searches.
>
> Conversely in prote currently there is the documented difference that
> \filesize uses a different handling than \input. That only needed
> documentation as it an unexpected difference.
Yes, it is documented in Prote precisely because I had a rough time
finding what the primitives were supposed to do (and fortunately
Phelype helped me a lot by providing test cases and "reference"
behaviors, because without this and only the "doc" available, it
was hopeless).
I have tried to make Prote orthogonal from kerTeX, while I'm the
developer of both, and I have tried to document what the primitives are
doing so that someone else can whether take the prote.ch as is, in
another distribution (Martin Ruckert has used it for HiTeX, that is
now in TeXLive and using kpathsea and the like; and he had strictly
nothing to change to prote.ch), or one can use the documentation
to implement things totally independantly.
But it is not difficult if the API is specified. The problem is here.
--
Thierry Laronde <tlaronde +AT+ polynum +dot+ com>
http://www.kergis.com/
http://kertex.kergis.com/
Key fingerprint = 0FF7 E906 FBAF FE95 FD89 250D 52B1 AE95 6006 F40C
|
