Breakpoints

[Andreas Pakulat]

On 15.04.09 23:09:18, Vladimir Prus wrote:
> On Wednesday 15 April 2009 22:37:08 Andreas Pakulat wrote:
> > The breakpoints already have a "set
> > but not yet active" visual state, so we should make use of that.
> 
> That's how it's done *now*, but this is not solution to scales to the
> case where you have 100 different breakpoints, half of which cannot
> be inserted.

Good point, no idea right now.
 
> > > Anyway, it seems like we're not getting closure via email -- can we schedule
> > > a real time chat -- say IRC or Skype?
> > 
> > You mean regarding a "Current launch config"? 
> 
> Not only that -- generally, how to keep 100 breakpoints without showing them
> all and informing the user there's some problem with 95 of them.

I could be available the weekend of 25th/26th, basically any time on sunday
or saturday afternoon.
 
> > We won't reach a conclusion
> > on IRC over that anyway. IMHO its wrong and I won't change my mind and
> > there's at least one other IDE out there that doesn't have a "current
> > launch config" or "current debug session". And debugging in that IDE works
> > quite well, even if its not java.
> 
> Do you mean Eclipse? I happen *work* on CDT, so I both deal with a number of
> bug reports regarding how that IDE works quite well (at least for ordinary users),
> and also know how it's implemented.

Yes and I'm working with the JDT only, so maybe thats something thats
different between CDT and JDT? And I just as well know how debug support is
implemented in Eclipse as I happen to have done exactly that for a custom
debugger ;) Though in that case we don't have to care about multiple debug
sessions as thats simply not supported :)

> The handling of breakpoints is definitely not something I want to have
> cloned in kdevelop.

It definetly has downsides and as I'm not going to implement any of that
anyway in KDevelop, maybe I should just shut up completely... 

Andreas

-- 
You will overcome the attacks of jealous associates.