On Thu, Dec 03, 2015 at 10:58:34AM -0800, K. "pestophagous" Heller wrote:
> On Wed, Dec 2, 2015 at 9:41 PM, Dirk Hohndel <[email protected]> wrote:
> >
> >> On Dec 2, 2015, at 9:20 PM, K. pestophagous Heller 
> >> <[email protected]> wrote:
> >>
> >> The DiveEventItem(s) need not store a pointer to memory
> >
> > "need not" is odd. "should not" maybe?
> 
> haha. oddness seems to be a side-effect as i try to internalize the
> discipline of using imperative tense in commit messages.  i never
> developed any habits about verb tense in commit messages, and now that
> i am hyper-aware of this concept, it's making me write real weird.
> getting afraid to use verbs in general. help!

Relax. Deep breath. It's all good. If I really hate it, I'll rephrase it.
Consider this "style suggestions", nothing more.

> >> DiveEventItem::recalculatePos is a slot.  As many have said, one faces
> >> an uphill battle (a losing battle) to try to reason cleanly about
> >> ordering and dependency graphs and such between all the various
> >> signals and slots.
> >
> > That's the beauty and the curse of it. It makes things presumably
> > easier when writing code - but then it comes back to bite you.
> 
> i am open to the beauty of it.  it offers the promise of some great
> potential decoupling of components, but it is hindered (for now)
> because many subsurface widgets that subclass Qt classes don't act
> "object oriented" in terms of encapsulating (hiding) and protecting
> their own state.

It's a mixed bag. Some do, some don't. Some can't because of weird
dependencies which might mean that we haven't abstracted things out in the
best possible way.

> > Would you re-edit this and re-submit?
> 
> will do tonight. my pleasure.

Thanks!

/D
_______________________________________________
subsurface mailing list
[email protected]
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface

Reply via email to