On Sep 25, 2011, at 10:52 PM, Simon Slavin wrote: > > On 25 Sep 2011, at 9:46pm, Petite Abeille wrote: > >> On Sep 25, 2011, at 10:29 PM, Roger Binns wrote: >> >>>> Would there be a way, in a future version of SQLite, to provide more >>>> informative constraint violation messages? Or is that just not feasible, >>>> implementation wise? Or? >>> >>> http://www.sqlite.org/src/tktview?name=23b2128201 >> >> Right. That ticket is sitting there, unloved, since 2006-Jan-30 :( >> >> http://www.youtube.com/watch?v=rwp60eYuie0 >> >> Any reason for the lack of love? E.g. "If it was hard to validate, it should >> be hard to understand what failed"?!? > > That is the reason which has been previously given. The various constraints > are combined and expressed in a way which is considerably different from the > form they were specified in. Figuring out which one was violated requires a > whole other procedure.
Hmmm... in the current embodiment of SQLite, is it even possible to figure it out, technically speaking? Or? I'm a bit puzzled by the care and effort that went into implementing all the constraint plumbing, everything from parsing the DDL, propagating to the engine, validating the DML operations, etc, etc, and then... nothing... just "constraint failed"... quite an anticlimax, no? Considering all the work that went into it? Baffling. _______________________________________________ sqlite-users mailing list [email protected] http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
