5 Sep
2019
5 Sep
'19
2:50 p.m.
some updates for those who are not closely following
github or IRC where most development is communicated ...
iOS and Desktop UX efforts are heavily and very nicely progressing ...
there are theming, localization, refined group-creation and chat views,
and much more, along with severe speedups and cleanups on all fronts.
More on this when we get to release them ... still a bit off but
hopefully not too much ;)
This core library porting work (126 merged PRs in the last months,
despite several vacationings and CCC camp intervening) has been done
by a growing team of existing and new developers, see here
https://github.com/deltachat/deltachat-core-rust/pulse/monthly
As to our super-scientific memory-management "evilness" metrics,
here is what happened between Jul 26, Aug 13 and today (Sept 16):
total unsafe: 567 -> 514 -> 287
total free: 326 -> 289 -> 186
total gotoblocks: 225 -> 147 -> 0
LOCs src/*.rs: 31296 -> 31150 -> 26970
The biggest source of manual memory management is now mime generation
and mime parsing. It is also often the cause for some bugs and issues
especially with incoming messages.
While the sudden development of porting the core library fro C to Rust
wholesale (started April 26th in Kyiv) has somewhat reduced our UI release frequency
it's going to be a big win in the slightly longer run. There are many new
features and improvements we want to tackle and our new Rust-core development
situation provides an excellent basis for this.
Nevertheless, Delta Chat remains a very usability focused project and
there were equally many efforts spent in improving and fixing UX issues
even though they didn't become visisble in releases yet. Next week
several of the UX-focused developers, researchers, activists and
designers are to hang out together in Freiburg and draw up plans for the
next half year of UX feature and streamlining happenings. But first
of all we'll focus on getting releases out, hopefully in September!
If you will, the new Rust-core work is also aiming at better usability, namely
for developers as they can now change more things with more confidence even if
they you don't have years of low-level programming experience. Special
thanks again here go to Friedel who carefully spear-headed and guided the porting work,
and to Alexander/link2xt and Kaction/Dimitry who joined these efforts (and the overall
DC team) six weeks ago and have made many significant contributions without which
we wouldn't be near where we are now.
Meanwhile, Google's new OAUTH verification policies drove away quite some DC users --
it's never nice if an app that you used suddenly stops working :(
The "fun" thing is that the OAUTH dev verficiation process seems to be geared
at "platform" apps that store user data on their own servers. However,
when you use DC with gmail then the only server used for storing user
data is Gmail and your end-device. There are no other servers involved
and we manage zero/zilch/nada user-data. This at least should enable us to avoid them
requiring between $15K-70K USD for a special security review as to how our (non-existing)
platform handles user-data. However, we have been bot-informed a few days ago that
we now need to produce youtube-videos showing and explaining OAUTH authentication
workings and this is what we are doing next (also quite some work).
Never mind, that Delta Chat worked nicely and caused 0 issues with
Gmail/OAUTH between Feb and July 2019 and never mind that DC comes with
a crystal-clear fully GDPR-compliant privacy policy, and its source code
is fully visible. Oh well, yandex OAUTH DC integration still works fine ;)
holger