Jean-Marc Lasgouttes wrote:
> It is not the syntax you showed earlier, is it? I saw
>
> Preamble
> $$XFigPreamble
> EndPreamble
You did see, but I decided to use the machinery of the output format rather
than invent a new machinery of my own.
Preamble XFigInset
will dump a macro definition to the .tex file. Thereafter, we are free to
use this macro as we see fit.
> If the later is what you use, then $$XFigPreamble is not a preamble,
> but rather a variable. So one could have
>
> Variable XFigPreamble
> blah
> EndVariable
>
> And then $$XFigPreamble would be usable everywhere... That may be a
> more useful language feature, isn't it?
I dno't think it is any more poserful than \xfiginput{yourchoiceoffilename}
> No, but you could test the return code of the command that produces
> the output and have an error if it is non 0. This would allow you to
> consider more general cases.
Sure. We could combine both ideas
Product '\\somefancymacro{$$BaseName.pstex_t}'
ProductFailed '\\someothermacro{$$BaseName}'
Preamble DumpBothTheseMacrosPlease
> Angus> Note that the changes I've committed so far (and the patch I
> Angus> have just posted) are almost orthogonal to this discussion.
> Angus> They are fixes and clean-ups of existing functionality.
>
> Are they? Hmmm...
They are. Of course, you could argue that code working as it was probably
envisaged rather than as it did actually is probably an enhancement ;-)
> The question is also whether we want latex to determine whether there
> was an error, or do it at latex generation time. We could also output
> either \input{foo} or \extinseterror{foo} depending on whether
> conversion succeeded.
>
> Thinking aloud...
Me too.
> PS: basically, you write it, so you have the last word :)
Well of course, but it's better to have a discussion leading to an exciting
finale isn't it?
--
Angus
