Source-Changes-D archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: Pre-branch commits



On 12.01.2019 13:54, Martin Husemann wrote:
> On Sat, Jan 12, 2019 at 05:11:31AM +0700, Robert Elz wrote:
>> I'd like to suggest a possible solution:   Go back to the old way,
>> and announce the branch date in advance (with a reasonable
>> lead time, not just a day or so, which would change nothing.
>> Reasonable here is likely to be something like a month.)
> 
> Yes, I'd like to do that.
> 
> I was trying to come up with an official "must have" feature list, but
> by now most of what I would have put there already made it into -current.
> 
> The only thing that is still open and not exactly a chekpoint feature
> is my personal perception that the aarch64 port (which will have its debut
> with 9.0) could use a small bit of more stability. Due to lack of something
> better my simple metric would be: number of ATF failures <= numer of ATF
> failures on e.g. alpha or sparc64, and additionaly (quite egoistic) a fully
> working XFce4 setup on my pinebook ;-)
> 
> After having everything in place, releng would pick some arbitrary date
> and if a that point all builds are green, create the branch. This date
> should be published as early as possible, with a request to developers
> NOT to rush in things shortly before the deadline.
> 
> Martin
> 

I'm investigating sigbus & vfork(2) issue with generation of core(5)
file (it's not just a syscall called in a vfork(2)ed child that could be
UB). Next I will work on aarch64 boot issues reported on qemu and push
the integration of LLVM sanitizers into src/.

I was asked to address the mentioned issues and I will rebase and finish
my code for LLVM sanitizers.

Attachment: signature.asc
Description: OpenPGP digital signature



Home | Main Index | Thread Index | Old Index