FreeTDS Development Group

["Thompson, Bill D (London)" <bill_d_thompson AT ml.com>]

Hi Freddy,

> Very big patch !!!

I don't do things by halves :-)

> I'm revising some of it (I avoided ctlib/dblib code)

feel free.

> - tds_alloc_cursor contain a buffer overflow calling strcpy, allocate
> namelen/querylen + 1 or use memcpy instead of strcpy

Thanks for spotting.

> - why you not send data (tds_flush_packet) within some tds_cursor_*
> functions ? You should also test result of tds_flush_packet

because in ct-library you can batch up a declare cursor, cursor setrows and
cursor open into one send

> - some more tests fail (tds/flags, ctlib/rpc_ct_param,
> ctlib/rpc_ct_setparam)

oops! didn't notice those failures. I'll take a look at them

Bill



> -----Original Message-----
> From: ZIGLIO Frediano [SMTP:Frediano.Ziglio AT vodafone.com]
> Sent: 31 October 2003 14:13
> To:   FreeTDS Development Group
> Subject:      RE: [freetds] Bill's patch submitted
> 
> > 
> > Hi all,
> > 
> > I've finally submitted my patch.
> > 
> > This contains a number of features.
> > 
> > 1) reworked ct-lib results processing. Hopefully the "empty 
> > result set" problem is in the past now.
> > 
> > 2) The basics of server side cursors. All the wire stuff is 
> > implemented for TDS 5.0 , 7.0 and 8.0
> >     The ct-library functions necessary to use cursors are 
> > implemented, and a test program is available.
> >     I've left undone all the issues surrounding our handling 
> > of multiple results metadata, as I consider it needs more thought.
> >     In any meaningful sense, the cursor functionality is 
> > useless until we sort this out, but I wanted to commit what I 
> > had so far.
> > 
> > 3) started work on moving bcp out of dblibrary
> > 
> > 4) some other small patches including ct_con_props CS_SERVERNAME
> > 
> > 
> > Any questions - I may be around from time to time this weekend.
> > 
> > Bill
> > 
> 
> I Bill!
> Very big patch !!!
> 
> I'm revising some of it (I avoided ctlib/dblib code)
> - tds_alloc_cursor contain a buffer overflow calling strcpy, allocate
> namelen/querylen + 1 or use memcpy instead of strcpy
> - some functions in query.c should call tds_to_quering to avoid reuse
> - why you not send data (tds_flush_packet) within some tds_cursor_*
> functions ? You should also test result of tds_flush_packet
> - some more tests fail (tds/flags, ctlib/rpc_ct_param,
> ctlib/rpc_ct_setparam)
> 
> freddy77
> _______________________________________________
> FreeTDS mailing list
> FreeTDS AT lists.ibiblio.org
> http://lists.ibiblio.org/mailman/listinfo/freetds