## LATEX-L@LISTSERV.UNI-HEIDELBERG.DE

#### View:

 Message: [ First | Previous | Next | Last ] By Topic: [ First | Previous | Next | Last ] By Author: [ First | Previous | Next | Last ] Font: Monospaced Font

Subject:

Re: More template experience

From:

Date:

Tue, 6 Feb 2001 23:24:40 +0100

Content-Type:

text/plain

Parts/Attachments:

 text/plain (28 lines)
 At 18:39 +0000 2001/02/06, David Carlisle wrote: > each with a brace group (\csname doesn't mind category 1 and 2 > characters), since even an ingenious fool has work very hard to get braces > mismatched in arguments, whereas with the current naming scheme At 20:18 +0100 2001/02/06, Frank Mittelbach wrote: >this is weird :-) and i thought my learning days on TeX the programs have been >over. > >even > > \expandafter\show \csname ab{cd\endcsname > >does work. TeX uses the character catcodes to tokenize the input, but if the lexer finds a macro name, it must do an additional table lookup in order to stamp an additional, internal token number (not described in the TeX book) for the parser, evidently, as different macros can obey a different syntax. But my guess is that both the lexer regular words syntax and the parser LALR(!) (?) syntax are fixed, only the character catcodes and the macro internal codes can change. (So why is TeX's syntax then commonly called "extensible" if both grammars are fixed?) Has this TeX parser grammar been published somewhere?   Hans Aberg