On 10-01-18 14:43, Berthold Stoeger wrote:
Hi Jan,
On Mittwoch, 10. Jänner 2018 12:43:53 CET Jan Mulder wrote:
On 09-01-18 10:36, Jan Mulder wrote:
I found out (more or less accidentally) that, for example, the
divelistmodel.h/cpp code is uniquely used in mobile only. It is highly
likely that this is true for more models, or models that are uniquely
used in desktop.
So my question (mainly to Dirk, I think). Is it worthwhile to adapt our
build system to make this more explicit? I would at least help
developers to realize this. Not sure it would help with respect to
footprint of the (primarily mobile) application, as the tool chain might
weed out that unused code.
In the meantime, I added a PR (#1033) to implement a split in the models
(only mobile, only desktop, and both). This is cmake stuff only, and no
other code changes.
That said, with one small code change, I have another PR in queue (that
waits for the decision on #1033), in which I have split-up the models
even further: 3 models generically used, 3 models unique for mobile, and
the remaining 17 models unique for desktop.(when I say model, I mean
.cpp source file containing the model code). Obviously, those 17 desktop
only model code it littered with #ifdef MOBILE blocks, put in in the
early phases of the mobile development, as we did want it to get
compiled. Formally, they are useless know, but notice that the current
split action is based on the current functionality of the mobile app.
And we all know that functionality can change, so models might get moved
over time.
I find both ideas appealing - separating the models into common, mobile and
desktop, as well as removing unused #ifdef-ery.
I was confused here. The mode code is reasonably clean from #ifdef-ery
:-) It is more in the profile widget, which is somewhat understandable,
as the widget is used in both mobile and desktop, but mobile lacks a lot
of features of the widget).
--kam
_______________________________________________
subsurface mailing list
subsurface@subsurface-divelog.org
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface
