http://bugs.winehq.org/show_bug.cgi?id=5048
--- Comment #14 from Jon Griffiths jon_p_griffiths@yahoo.com 2011-01-24 21:29:16 CST --- (In reply to comment #10)
TransmitFile may have an implementation, but the ws2_32 side of the coin here is still doing a FIXME("unimplemented TransmitFile").
Didn't the problem just get moved further down the pipe?
It did, this bug isn't fixed, since my patch never made it in.
I've done a bit of digging on a related issue (details here http://forum.winehq.org/viewtopic.php?p=54231#54231), and am working on tying together a possible TransmitFile implementation. Is anyone else (actively) working similar, or otherwise know of a reason why I shouldn't pursue this?
I last had a spate of wine submissions when i had a month or so off work. I'm back at work more than full time now and can't afford the time to massage the patch to get through. I tried to implement tests for this function to get it in and encountered all sorts of unrelated issues in testing that would have taken me far too long to uncover, so i dropped it, hoping that someone might pick it up. A bunch of other patches got ignored at the same time so I decided to enjoy my holiday rather than struggle to get the changes in. It didn't help that the bug was marked fixed when it clearly wasn't. It seemed like the 1.0 series brought out a focus on ticking boxes without focusing on the actual user experience.
I'm happy if you want to pick up my patch and go with it, its still serviceable. The error handling just needs to be fixed according to the last feedback I got, and you could implement overlapped IO but its not needed to make docuwiki work.
Its a shame I think that the barrier for entry for submissions has been raised so high now. I can understand it for existing functionality, but for unimplemented functions a kaizen approach would be better. Checking in the original patch with a FIXME for the error handling would have benefited users more. Someone else could easily have fixed just that issue with the code in tree, rather than letting it bitrot.
I still have patches from 8 years ago that never made it in, and more from a couple of years ago too. But for someone like me with very limited time, the hassle of maintaining a tree full of changes just isn't worth it when there isn't clear feedback on what needs to be done to get patches in, or lots of extra work is demanded before stuff will be accepted. I'll try again when I get more time, but that could be 5 years away at this point. In the future I've thought I might just contribute tests, since they should be easier to get in.
Anyway, the same goes for any patch I've ever sent that didn't make it in. I'm happy for anyone to rework/resubmit them, just please give credit + shared copyright if enough of the original work remains to warrant it. I'm happy to answer any questions about old patches too.
Cheers Jon