msi: Implement the DROP TABLE sql command.

[Austin English]

On Thu, Oct 2, 2008 at 3:47 PM, Steven Edwards <winehacker at gmail.com> wrote:
> On Thu, Oct 2, 2008 at 4:42 PM, Austin English <austinenglish at gmail.com> wrote:
>> You'd still be sending the patch twice to wine-patches, and by that
>> logic, for _EVERY_ patch, not just ones that were rejected.
>
> No. you would only have to send it again if the first one was rejected
> by patchwatcher. By default no ones work flow would change, just the
> email you are sending it to for retries saving everyones inbox on
> spam. If your patch was bad you would still have to do extra leg work
> but all of that would be done in private. If we get make test working
> in more places then its possible the number of first time failures to
> wine-patches will go down also if we can encourage people to run make
> test before submitting patches.
>
> --
> Steven Edwards
>
> "There is one thing stronger than all the armies in the world, and
> that is an idea whose time has come." - Victor Hugo
>

Here's how I understand your idea, correct me if I'm wrong:

WinedevA submits a patch to wine-patches. Crashes on patchwatcher,
etc. WinedevA gets a private e-mail from patchwatcher saying "you're a
failure, learn to code better, fix this, etc.". WinedevA fixes patch
up, resubmits to wine-patches-testing multiple times, finally gets it
right. He does a little dance, celebrates, etc. Patchwatcher then
forwards this good patch to wine-patches.

WinedevB submits a patch to wine-patches, works fine. Passes
patchwatcher, no valgrind warnings, etc. Because it passed, it then
forwards it to wine-patches. Unless patchwatcher somehow keeps track
of all patches that have failed and been resubmitted, which seems like
extra work for nothing.

I thought we had a consensus on the process I mentioned earlier...

-Austin