[haiku-commits] Re: r39846 - in haiku/trunk: headers/os/kernel headers/private/debug headers/private/kernel src/kits/debug src/system/kernel ...

  • From: "Axel Dörfler" <axeld@xxxxxxxxxxxxxxxx>
  • To: haiku-commits@xxxxxxxxxxxxx
  • Date: Tue, 14 Dec 2010 19:11:41 +0100

Rene Gollent <anevilyak@xxxxxxxxx> wrote:
> On Tue, Dec 14, 2010 at 8:50 AM, Axel Dörfler <axeld@xxxxxxxxxxxxxxxx>
> > wrote:
> > It's not really relevant debugging information, so using the
> > debugger
> > interface to publish that info indeed doesn't feel right.
> > While an extra flag could be more convenient to use, another
> > possible
> > solution besides polling would be to introduce a message based
> > generic
> > notification possibility (based on the existing notification
> > mechanism).
> It just seemed like the logical way to do it to me at the time since
> it helps keep the debugger in sync with the state of the threads it's
> tracking.

Sure, it's very convenient.

> While the name is purely cosmetic (but still annoying as
> pointed out in the case of Tracker and a few others), knowing the
> priority seems like valid debug information since that may come into
> play in various situations where you have a concurrency bug such as a

find_thread() by name might also be important in a debugging situation,
and presenting current data is definitely the way to go - it's just not
debugging specific information.

> priority inversion. In any case, reverted for now, will look into the
> notification mechanism if that's what's preferred.

At least something like ProcessController could make use of it as well,
and it would be the cleanest solution IMO. I wouldn't mind the
convenient flag version Ingo suggested, or to keep your solution
temporarily either, though.


Other related posts: