[tex-live] STOP: Showstopper on Mac OS X

Gerben Wierda Gerben.Wierda at rna.nl
Sun Sep 21 01:25:22 CEST 2003

On Sunday, Sep 21, 2003, at 00:03 Europe/Amsterdam, Thomas Esser wrote:

>> Something seriously goes wrong here with the newly created programs
>> (source Sep 20, 21:30). There is a problem with POOLSIZE in pdftex and
>> pattern memory as well (note: nothing changed in my variables and
>> texmf.cnf!) and suddenly fmtutil -all starts doing an endless loop of
> Hm... Can you please retry to build just one format, e.g. latex and
> set KPATHSEA_DEBUG=4 and see if some wrong texmf.cnf file is used?
> Or is some essential file is not found?
> E.g. trie_size should be 262000 in the first texmf.cnf file that sets 
> it.

The texmf.cnf that is installed by 'make install' is severely limited 
it turns out.

The only variables for sizes it has are:

> % Max number of characters in all strings, including all error 
> messages,
> % help texts, font names, file names, control sequences.
> % These values apply to TeX and MP.
> max_strings      =              100000 % max number of strings
> nest_size =                     500 % simultaneous semantic levels 
> (e.g., groups
> )
> param_size       =              10000 % simultaneous macro parameters
> save_size        =              50000 % for saving values outside 
> current group
> stack_size       =              5000 % simultaneous input sources
> % Context's metafun
> main_memory.metafun      =              2500000
> pool_size.metafun        =              1000000

Apparently, this has changed recently. I know Sebastian was planning to 
update the texmf.cnf that was installed by make install, but this seems 
a downgrading more than an upgrading ;-)

I can work around this, but I think the texmf.cnf installed by 'make 
install' should have decent values (in fact, I think the one installed 
by 'make install' should be the only one.

"To be or not to be, that is the question" -- Parmenides

More information about the tex-live mailing list