FreeTDS Development Group

["Brian Bruns" <camber AT ais.org>]

> > 
> > Hi all,
> > 
> > I've just checked in some fairly substantial changes to CVS, based 
> > largely off of Frediano's patch, but reworked a bit by moi.  
> 
> Wow.. I never tough this patch was so important to be included in 0.6
> version

It wouldn't have except Bill's stuff is going to need it, and I consider
getting the conversion stuff done once and for all important enough to
thaw the freeze a bit.

[snip]
> I think this is not good. context mantain information per library, msg_info
> is not per library (but can be per socket). First strange thing I note in
> patch was msg_info in TDSCONTEXT.

actually some (client) messages can be per context...I'm leaning towards
tds_client_msg() having this as an automatic variable.  Don't pay any
attention to the man behind the curtain, this patch is not complete yet.

> Perhaps best solution is to pass to libraries two parameters
> message(TDSSOCKET* ,TDSMSGINFO*), (TDSCONTEXT can be retrieved from
> TDSSOKET). Added library information per socket can be retrieved by
> tds_socket->parent. 

Unless you have a NULL tdssocket, which is the entire reason for this
patch going in.

> I think library must be full-multithreaded for socket (I cannot read from a
> socket from 2 thread).

Agreed.

> Mmmm... 
> - tds all work (I have also written some fix for problem deleted unexisting
> table or other see patch #580272)

Is this a big deal? It's been that way forever, and just generates a not
found message which is summarily ignored by the unittest.  I'll slate it
for 0.61

> - none of ctlib work (see patch #581224 for fix all)
> - none of dblib work (see patch #581224 and also as stated below fix calling
> to tds_willconvert in dbbind and 17 of them work)

> PS: I'm glad if someone can check patch #580072 (IMHO another required patch
> for 0.6)

I'll take a look at these.

brian