@franciscojunior, you commented out a section of NpgsqlCommandBuilder a while ago as a fix for pgfoundry bug 1010973, but I'm pretty sure that broke the command builder standard behavior...

I think you are right. As I didn't have a build server and a the test suite in a good shape, I think I inadvertently added that regression. Thanks for heads up and patch!

I'll have a look at it.

I think your code is very good!

DataAdapter and related CommandBuilder are, to me, one of the most complex classes because at the time I started to work on it, I didn't know how it would interact with .net system.

I'm glad you could grok that interaction and fix this annoying bug!

Thanks for the review @franciscojunior, I'm merging.

@franciscojunior, I think it's a good idea to merge this into the 2.2 release?

Em 26/07/2014 05:06, "Shay Rojansky" notifications@github.com escreveu:

@franciscojunior, I think it's a good idea to merge this into the 2.2
release?

Sure!
I'm on it!

—
Reply to this email directly or view it on GitHub.

Applied the patch to release-2.2.0 branch. I had to cherry-pick them because the parent of the pr-294 branch was different and was bringing other commits. But I added all the relevant references.

OK, great @franciscojunior, thanks! I think cherry-picking like you did is the only way to go in this situation.

Great!

I'm glad to hear I got it right.

Thank you too, @roji, for all the help and tips about git. :)