[RFA] Make access watchpoints work again

Eli Zaretskii eliz@is.elta.co.il
Tue Mar 20 01:54:00 GMT 2001


On Mon, 19 Mar 2001, Andrew Cagney wrote:

> > (Btw, why isn't the case, where an access watchpoint triggers, but the
> > value didn't change, being annotated with a call to
> > annotate_watchpoint?)
> 
> If it wasn't being done in the pre- ui-out code, I'd assume it's because
> no one noticed it missing :-) If it was being done then I'd assume a bug
> in the conversion to ui-out :-(

No, the conversion to ui-out seems to properly follow what's been
there before: any watchpoint which triggers is not annotated if the
watched expression's value is unchanged.

The only way I can understand this is that the users of the annotation
interface did something which doesn't make sense when the value hasn't
changed.  I don't have experience with any application that uses
annotations, so I cannot figure out whether this hypothesis is true or
not.  Anybody?

FWIW, annotate.texi doesn't say anything about read watchpoints being
unannotated.

If no one can find the reason, how about adding the missing annotation
and waiting for someone to scream?

> PS: Assume the ui-out change is an obvious fix.

Commited.



More information about the Gdb-patches mailing list