[tex-live] Is TeXk from TeXlive 2003 TRIP certified?

Olaf Weber olaf at infovore.xs4all.nl
Wed Mar 17 09:03:01 CET 2004

Petr Olsak writes:

> On Tue, 16 Mar 2004, Fabrice Popineau wrote:

>> > I checked out this part properly:
>> You need to use the texmf.cnf provided in the texk/web2c/trip directory
>> of the source tree, not the default one.

> In another words: the default TeXk with default configuration file is not
> TRIP compatible. It is out of my interrest that the TeXk binary is able to
> be TRIP compatible by some special configuration file and some special
> operation environment when the default configuration is not TRIP
> compatible. This was not the reason of TRIP test.

> It is your intend to make TeXk with default configuration file TRIP
> incompatible? Why?

Read up on the trip test, please.  You're supposed to either build a
special-purpose binary to run the test, or have a way to "tune"
various parameters of a generic binary to values that make it
completely useless for any real work.  (You're hurt by the memory size
differences in particular.)

>>> PS. The difference in last digists in dvitype output is not explained
>>> by this.
>> No, but it has always been regarded as acceptable.

> Yes it is acceptable, but why _the same_ operating system, _the same_ TeX
> distribution (web2c only in different versions) has these differences?

I find the move from 'w' to 'x' commands also interesting.  It
suggests something inside is doing things in a different fashion.

> Yes, it is acceptable, but the two pages lost by TeXk (14 pages versus 16
> pages) are not acceptable.

Line lengths may be the issue there.  Do an eyeball compare of the
results first to see whether that can be issue.

Olaf Weber

               (This space left blank for technical reasons.)

More information about the tex-live mailing list