On Tue, 3 Jul 2007, Martin Schreiber wrote:
> On Tuesday 03 July 2007 12.57, Michael Van Canneyt wrote: > > I'd rather see that the whole is coordinated a bit. I don't think all this > > duplication is a good idea; One solid set of components makes much more > > sense: the number of developers available for it's maintenance would > > double. The end user would also benefit, as we can give 1 solid set of > > components, no doubts possible. > > > > There are not so many people available who understand the inner workings > > of TDataset and friends, so all this scattering and dividing is > > counterproductive. > > > > It would be much better if you, me, Joost and Bram (who work on SQLDB) > > had a single set of components to work with: speed would not be an issue > > then, and we could guarantee a working and reliable set of components for > > everyone. > > > You are right, I will think about your offer. :-) That's good, thank you :-) > Michael, Joost and Bram please look into my code, can you imagine to merge > with FCL? > The relevant files are in > https://mseide-msegui.svn.sourceforge.net/svnroot/mseide-msegui/trunk/lib/common/db Checking out now. I'll consult with Joost and Bram separately. I am a bit worried about the widestring issue. While good for msegui which uses widestrings internally, it's also slower and more memory consuming. A fixed buffer has it's advantages. The change-log handling and indexes are a Good Thing (tm), and I remarked on that some time ago already :-) We'll see. Let me consult, and then I'll get back to you. Michael. _______________________________________________ fpc-pascal maillist - fpc-pascal@lists.freepascal.org http://lists.freepascal.org/mailman/listinfo/fpc-pascal
