FreeTDS Development Group

["James K. Lowden" <jklowden AT schemamania.org>]

We have 31 functions (in db-lib and ct-lib, combined) that are stub
implementations.  

I think we should remove them from the header files and split them off
into stub.c.  I don't see the point of deferring the error until runtime. 


Is there some concrete benefit to keeping them?  If they're acutally
useful to you, maybe we could preserve them as long as TDS_INCLUDE_STUBS
is defined, or something like that.  

The functions are listed below for your convenience.  

Thanks.  

--jkl

$ perl doc/api_status.txt | grep stub | awk -F'\   ' '{print $4}' | uniq
bcp_moretext
dbcurcmd
DBCURCMD
dbcurrow
DBCURROW
dbmny4divide
dbmny4mul
dbmnyadd
dbmnydivide
dbmnydown
dbmnyinit
dbmnymul
dbmnyndigit
dbmnyscale
dbmnysub
ct_poll
cs_conv_mult
cs_manage_convert
cs_objects
cs_set_convert
cs_setnull
cs_strcmp
cs_time
blk_colval
blk_default
blk_getrow
blk_gettext
blk_rowalloc
blk_rowdrop
blk_sendrow
blk_sendtext
blk_srvinit
blk_textxfer