[interfacekit] Re: A patch for BString

On 2003-11-08 at 02:28:09 [+0100], Ingo Weinhold wrote:
> 
> On 2003-11-08 at 00:05:21 [+0100], Oliver Tappe wrote:
> > 
> > So we should probably go with Ingo's solution as it seems more reliable, 
> > but
> > it may be worthwile to invest some more time into the debugger-stuff.
> 
> I actually like the debugger solution better, for it's less a hack. And one
> can probably live with not hitting the real debugger with the right thread
> and the original cause.

Although I still think the debugger approach would be nicer, I used my symbol 
patcher, since it is really a bit annoying, if not the calling thread but the 
debugger thread enters the real debugger (run the BArchivable tests with the 
other solution and you know what I mean :-).

As proposed by Oliver the CPPUNIT_ASSERT_DEBUGGER macro can now be used to 
execute a statement that is supposed to call the debugger. debugger() 
invocations without a CPPUNIT_ASSERT_DEBUGGER macro will enter the debugger 
as usual (almost: two additional stack frames (of the interception code) will 
appear). The macro has effect only for the calling thread, i.e. another 
thread invoking debugger() will indeed enter the debugger (unless being 
within a CPPUNIT_ASSERT_DEBUGGER call itself).

CU, Ingo

Other related posts: