[Seaside] about roe update issue

danil osipchuk danil at tsnet.ru
Sat Mar 27 15:38:46 CET 2004


Hello Avi and list
It seems that trouble with updates is reflection of the fact that
insert/update/delete operations are not direct subjects of relational
algebra. These operations in SQL are defined only for tables and I
think roe should behave in the same way. So if #update will be removed
from core RARelation and #update: updateBlock where: selectionBlock will
be added to table relations instead, issue must cease (the same true
for #delete and #deleteWhere: selectionBlock). So there would be no
innatural looking SQL expressions as in my previous post.
I'm attaching another mcz with approximate implementation in
RAPostgresRelation. It breaks testUpdateAlias and
testUpdateAliasWithColumn (I think forever, but this seems to be
unavoidable). I didn't change RAArrayRelation, so it's broken too.
What do you think about all this?
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Roe-dao.37.mcz
Type: application/octet-stream
Size: 15620 bytes
Desc: not available
Url : http://lists.squeakfoundation.org/pipermail/seaside/attachments/20040327/7795c45c/Roe-dao.37-0001.obj


More information about the Seaside mailing list