Re: [Comp-hist-devel] Updated (again) Parser

[Kurt R. <ku...@ra...>]

Should this be put in CVS?
-Kurt

> From: "Christophe Fergeau" <Chr...@fn...>
> Reply-To: com...@li...
> Date: Sat, 26 Aug 2000 23:15:07 +0200
> To: <com...@li...>
> Subject: Re: [Comp-hist-devel] Updated (again) Parser
> 
> I fixed a few things in the parser :
> 
> 1. On my machine, it could not compile (type casting pb in parse_line)
> 2. The file sinclair was not correctly parsed because it contains lines more
> than 512 characters long
> 3. Now, it correctly deals with nodes which have several Successor to fields
> 
> Now that we have a working C parser, we can easily convert our database to
> an  sql database and allow people to consult our database using php. If you
> want such an access, I can do it but not before mid september
> 
> Next week, I'll be away so I won't be able to do any work for the project.
> At least, I'll try to check my mails regularly.
> 
> Chris
> 
> ----- Original Message -----
> From: Hans B Pufal <ha...@di...>
> To: <com...@li...>
> Sent: Saturday, August 26, 2000 9:56 AM
> Subject: [Comp-hist-devel] Updated (again) Parser
> 
> 
>> OK, I updated the parser again!
>> 
>> It is now a module which can be easily integrated into a larger program.
>> 
>> The parser code is also independent of the node structure (little bit of
>> tricky coding there!)
>> The node structure is defined in node.h and the preser gets its info
>> from there. So adding a new field is simply a matter of editing two
>> tables in node.h and recompiling.
>> 
>> I removed identifiers.* (moved the one required routine to parser.c) and
>> keywords, that data
>> is now in node.h
>> 
>> Compiling parser.c alone will generate an obj module ready to be linked,
>> compiling with a define of TESTING will create program which parses all
>> its input files and prints out the
>> nodes read.
>> 
>> -- Hans
> 
> 
> 
>