Re: RTEMS-5 integration

[Heinz Junkes <junkes@fhi-berlin.mpg.de>]

Hello Benjamin,
Do not make the effort.

I will create a new merge request for the ‘dereorganized' repo.

I will also try to address the issues raised by Michael. Especially the question about osdThread.c from the posix branch
For safety's sake, I created my own osdThread.c for RTEMS5. There was a serious bug (https://devel.rtems.org/ticket/3243)
which has been fixed in the meantime. I'll try it now with the existing posix/osdThread.c. But I'm on the road for the next 1.5 
weeks and will come to it later.
Heinz

> On 26. Jun 2018, at 17:34, Ralph Lange <Ralph.Lange@gmx.de> wrote:
> 
> The repositories of the modules (core, libcom, database, ca) that were separate branches on the epics-core repo have been merged (recombined), without changing the file structure.
> The resulting new-focus-of-attention branch is called '7.0' (in line with the other release series' branch names).
> 
> You can try to merge the request for core straight into 7.0 - in that case the structure has not changed.
> 
> For the libcom request, the underlying structure has changed, so you would either have to move everything into modules/libcom on the proposal branch before trying to merge, or create a set of patch files that you can try to apply to a 7.0 clone while being in modules/libcom.
> 
> Lots of overhead, I know. But it is the only merge request that is affected.
> 
> Cheers,
> ~Ralph
> 
> 
> On Tue, Jun 26, 2018 at 5:24 PM Benjamin Franksen <benjamin.franksen@helmholtz-berlin.de> wrote:
> On 06/26/2018 01:40 PM, Ralph Lange wrote:
> > Merge Proposals are at 
> > https://code.launchpad.net/epics-base/+activereviews
> 
> Thanks.
> 
> > Heinz' RTEMS branch needs some clean-up and minimization - there are 
> > comments in the merge proposals that name some places.
> 
> Okay, I have read Michael's comments.
> 
> > Note that the two MPs are actually one (again), as we recently
> > dropped the split-up of repo-branches for the modules formerly known
> > as EPICS V3.
> 
> Nice. It looks as if the rtems5 branch is a bit behind, though, it still
> has submodules for these parts.
> 
> > I was volunteering to help him do this, but it would actually be
> > great if you could jump in. My arguments (same language and time
> > zone) cover you in the same way, and you have a much better
> > justification (as ITER does not intend to use RTEMS).
> 
> The first thing to do would be to bring the rtems5 branch up to date
> with core/master. I tried to merge hj/rtems5 which succeeded but of
> course that gave me only a few changes in configure. I then tried to
> merge hj/rtems5libcom but that failed horribly with thousands of
> rename/delete conflicts. My git-fu is rather weak so I guess my naive
> attempts were doomed from the beginning...
> 
> Cheers
> Ben
> --
> "Make it so they have to reboot after every typo." ― Scott Adams
>

Attachment: smime.p7s
Description: S/MIME cryptographic signature