[XeTeX] How to manually create the xelatex.fmt?

[Chris Travers]

On Wed, Oct 19, 2011 at 6:10 AM, Peter Dyballa <Peter_Dyballa at web.de> wrote:
>
> Am 19.10.2011 um 12:19 schrieb Chris Travers:
>
>> If RHEL 6 (released about a year ago) is sticking to TeXLive 2007, we all have problems.
>
> The only problem is that of understanding. It's like the fifth wheel or the tool to change wheels that come with new car. They're not really usable, they're more kind of alibis. And that's the situation of TeX in Linux.

Ummmm....  Not necessarily.  Anyway, aside from the xelatex.fmt issue,
which I have found (and documented) a workaround for, it's working
well enough for me, and it's certainly working well enough to
predictably generate invoices in an accounting system.

Hint:  Maybe if that hammer isn't useful it's a crescent wrench instead?

>
> Because it's not necessary to build a second rail of distribution via DEB or RPM packages. TeX Live comes with its own package manager and in packages and in meta-packages. Use this opportunity!

But if you do that, you lose the ability to tie the application built
in Perl to its own dependencies in a package manager.  So you have a
package manager, Perl has a package manager, and the OS has a package
manager and none of them talk to eachother.  The result becomes a
dependency tracking nightmare.
>
> In Mac OS X, Solaris, MS Windows we have the freedom to use third parties as suppliers of software packages. It's time that Linux grows up to that level. Even when it becomes complicated for the users and administrators to send bug reports to a thousand authors of TeX Live packages.
>
You know, that's kind of unnecessary.  I could just as easily point
out that I came here looking for help on what I felt is an important
development on LaTeX, but now feel like it's pretty clear that XeTeX
hasn't outgrown it's shiny-new-iMac roots, or that it's time for XeTeX
to realize that real productivity occurs in the area of server-side
document processing where stability is far more important than folks
here seem to want to acknowledge.

To be honest, I am pretty discouraged here.  I've long used LaTeX for
document processing because it is a stable technology and unlikely to
change out from under my documents.    I understand that XeTeX hasn't
reached that level of maturity yet and may never.  However, it seems
to me that this community here doesn't really care about the kinds of
environments where this sort of document processing occurs.

>
> Have you thought of keeping exactly one (maybe only virtual) alibi server with long-term support software? In case someone asks you have something to present... (and when it's virtual you can easily add some CPU power and/or disk space, if needed)

Not really an option.   The goal is to get people up and running fast,
with their package managers if they prefer.  That's not negotiable.
>
>
> BTW, the packages supplied by CTAN are *stable* packages. (It's also possible to get preliminary test software.)

I think "stable" in terms of "you can safely use this to render your
documents" and "stable" in terms of "no unnecessary changed so we know
the software using this clearly and predictably works every time" are
different senses of the word "stable."  I need the latter once the
software is installed, you are talking the former.

The point is that changing upgrading software underneath fairly
critical systems just because there is a newer version out with bug
fixes that don't affect you will *always* cause more harm than good.

Best wishes,
Chris Travers