Richard Frith-Macdonald wrote: > > On 31 Oct 2008, at 20:20, Fred Kiefer wrote: > >> Samantha Rahn wrote: >>> What types of syntax can and cannot be used in NSPredicate? >>> >>> When I use >>> >>> companyName CONTAINS 'Corp' >>> >>> I get: >>> >>> NAME:NSInvalidArgumentException REASON:Format string contains >>> extra characters: "companyName CONTAINS 'Corp'" >>> >> >> I had a look at the source code and it looks like we are missing code >> for the parsing of string constants. Somebody will have to add this. > > Looked simple, so I added it ... try svn trunk to see if it works. > >> The whole file NSPredicate.m needs a make over, we should split it up >> and replace the current scanner with something more powerful. > > What is needed? The current NSScanner based implementation has lots of > method call overheads and is obviously going to be slower than something > similar written in pure C, but it's not obvious to me that it's less > powerful (and it has the advantage of re-using reliable code). > I'm not familiar with NSPredicate though, so I could easily be missing > some disadvantage.
Thank you. Looks like you are more familiar with NSScanner than me and if you think it is a supportable way of implementing this, we should stay with it. _______________________________________________ Discuss-gnustep mailing list [email protected] http://lists.gnu.org/mailman/listinfo/discuss-gnustep
