Dear Remy,
In message 3efb10970902190815v2fc3010dg5574891a9ab26f07@mail.gmail.com you wrote:
I know you do not want build failure messages on the list, but this
You are absolutely right - instead of build failure messages I prefer to see patches that fix build failures :-)
Instead of sending patches, I would prefer such patches not hitting mainline at all...
You are right, that would be best.
Next best is to provide fixes.
I think it is the job of the one who breaks things to repair the damage, but that person sometimes needs to be told that he broke something, because I assume it is not broken on purpose. That's why I posted this build log. Do not shoot the messenger...
The thing is that patches get posted here for review. People are supposed to have a look at the patches, and eventually even try these out. If there have no complaints been posted after a couple of days, patches will get applied.
Even for a very careful submitter it is usually impossible to compile all boards on all architectures. Speaking for myself, I don't even have toolchains installed for BlackFin, AVR32, Leon, MicroBlaze, NIOS and some other archictures - do you?
So we have to rely on co-operation of as many developers in the community as posible.
And if the breakage was small, I would already have made a patch. But the breakage is that huge here, that I would suggest to revert those patches, until the committer has solved the bugs he introduced.
What if the problem doesn't show up on the boards he can build, but shows up on the boards you can build?
(If fixing this takes a long time from now, the tree is no longer proper bisect-able, therefore reverting could be wiser.)
So please help out...
Best regards,
Wolfgang Denk