[vwnc] What is Store's "trace" column used for?

The "trace" attribute appears in TW_ClassRecord, TW_NameSpaceRecord,
TW_Package, TW_ParcelRecord, TW_Method, TW_Blessing, TW_Bundle,
TW_DataElement, TW_FileRecord, and TW_PropertyRecord.

It looks like it should be an FK somewhere.  Possibly self referencing
to show heritage?

--
Visit <http://tggagne.blogspot.com/> for more great reading.

_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc

Thomas:

> The "trace" attribute appears in TW_ClassRecord, TW_NameSpaceRecord,
> TW_Package, TW_ParcelRecord, TW_Method, TW_Blessing, TW_Bundle,
> TW_DataElement, TW_FileRecord, and TW_PropertyRecord.
>
> It looks like it should be an FK somewhere.  Possibly self referencing
> to show heritage?

If the version is a new version then trace points back to the parent.

_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc

Samuel S. Shuster wrote:
> <snip>
> If the version is a new version then trace points back to the parent.
>
MySQL's constraint seems to work OK, though.  I replaced the 0 with NULL
and all is well.  I also modified #trace: to use NULL when 0 is sent.

--
Visit <http://tggagne.blogspot.com/> for more great reading.

_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc

At 08:04 PM 4/23/2008, Thomas Gagné wrote:

Samuel S. Shuster wrote:
> <snip>
> If the version is a new version then trace points back to the parent.
>
MySQL's constraint seems to work OK, though.  I replaced the 0 with NULL
and all is well.  I also modified #trace: to use NULL when 0 is sent.

I think you may want to be careful about that. I don't have anything concrete, but I have a vague recollection that that might be a problem.



_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc

Alan Knight wrote:
> <snip>
> I think you may want to be careful about that. I don't have anything
> concrete, but I have a vague recollection that that might be a problem.
Initially it did hiccup a few places.  But I've been playing with it now
(loading/publishing) and all seems happy.  Right now I'm publishing Base
to see if I can find any errors during the inserts.

--
Visit <http://tggagne.blogspot.com/> for more great reading.

_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc

While in there, I noticed three different methods update the instance
variable, 'trace.'  They are #trace:, #dbtrace:, and #setTrace:.  I
overrode the first to set nil when 0, then modified the other two to
send 'self trace:.'

I remember noticing that earlier while getting the StoreForMySQL working
I needed to put a breakpoint somewhere to see who was doing what and
later discovered there were multiple setters.  Maybe some refactoring is
due?

--
Visit <http://tggagne.blogspot.com/> for more great reading.

_______________________________________________
vwnc mailing list
[hidden email]
http://lists.cs.uiuc.edu/mailman/listinfo/vwnc