FreeTDS Development Group

["Ludo Brands" <ludo.brands AT free.fr>]

> 
> Planned, no.  Possible, possibly.  
> 
> I had to write a little parser to understand your 
> capabilities, output attached.  TDS_REQ_LARGEIDENT is bit 83, 
> which you'll note is byte 11, and FreeTDS sends 9 capability bytes.  
> 
> I haven't looked into it, but I would hope that you could add 
> two bytes to the defaultcaps variable defined in src/tds/mem.c.  
> 
> There also might be a baked-in assumption about 30-character 
> identifier names, but I doubt it.  DBMAXNAME is defined as 30 
> but used in only two places.  (Don't name your database with 
> more than 30
> characters.)  Otherwise you should be OK because most of the 
> code doesn't care which version of the protocol you're using.  
> 
Thank you for the reply.

I have tested already with a modified defaultcaps that sends 12 bytes
including TDS_REQ_LARGEIDENT (and set TDS_MAX_CAPABILITY to 25 in tds.h). It
works fine for 30+ identifiers. I didn't test up to 256 chars but since the
odbc driver reports 128 char identifiers for SQLServer >= 7, I assume that
at least 128 would be OK. I haven't tested this against older sybase
servers. 
This is a straightforward mod but it doesn't cover the case where
ENABLE_EXTRA_CHECKS is defined. When this is defined, the capabilities
packet is checked in mem.c against a set of capabilities. It hardcodes in
several places that the request_capabilities and response_capabilities
blocks are both 9 bytes or, worse, TDS_MAX_CAPABILITY / 2 - 2. 
I guess ENABLE_EXTRA_CHECKS is defined only in a test/debug environment. 

Ludo