4.7.5 next week

[Dirk Hohndel]

> On Dec 1, 2017, at 4:51 AM, Jan Mulder <jlmulder at xs4all.nl> wrote:
> 
> Dirk Hohndel schreef op 2017-12-01 04:43:
>>> My request for 4.7.5 is a merge of Jefs libdivecomputer work into our branch. Some OSTC work related to pSCR diving is recently added (and some other stuff that I have no opinion about).
>> libdivecomputer master contains quite a few changes for Jef's approach to
>> refactoring the IO subsystem - which unfortunately moves it even further
>> away from what we are doing (and based on what I've seen so far, this
>> isn't going to be a useful approach for us, so I'm not sure what we can do
>> here). I'm definitely NOT in favor to merge this right before a release...
> 
> Yes, I fully understand this. However, with a monthly schedule of Subsurface releases, things are getting even more difficult to get in, as the window for testing gets small.

My guess is that this will force me to think through the development model once again.
With a libdc submodule (HIGH on my todo list) it should be possible to work on nothing but the iostream issues in separate branches on both Subsurface and libdc for a while and then merge that back in once we are happy with things.
But with the releases for Subsurface, Subsurface-mobile for Android and Subsurface-mobile for iOS I really don't want to introduce things into master that are risking to have the code broken for many weeks.

>> I talked to Linus about this (who is traveling this week) and he was going
>> to take a look and merge all the non-iostream related commits into our
>> branch. I can see about eight or ten commits in there that we should
>> pick... everything up to the 0.6.0 release, and then today's commit that
>> detects invalid profile data on an OSTC3.
> 
> Yes, this might be a way on at this point in time.

I haven't seen a comment from Linus - he may have spent today on a plane, I don't have his exact travel schedule...

>> Will the SCR patch from today be useful for us without further changes
>> in Subsurface?
> 
> Well, I will add a commit after the actual add of the DC_DIVEMODE_CCR/SCR change to get pSCR for OSTC fully working (with correct divemode). Without this, the pSCR import from OSTC3 is simply broken (cannot parse header). But that can go after 4.7.5 (cosmetic error). So yes, very usefull as it  unbreaks things. That extra change obviously does not build before the change in libdivecomputer..

Which brings me back to submodules. DIRK GET YOUR ACT TOGETHER

/D