[epiar-devel] The successes and failures of a night with MSVC

[Doug Mirro]

  some of the code i have that needs to be tweaked before submitting 
might provide better debug info for VS users.  i will try and get it in 
tomorrow

Doug Mirro
Easy Tech Talk - Technology made simple <http://easytechtalk.com/>

On 10/6/2010 12:42 AM, Димитър Станчев wrote:
> Hi,
>
> On what version of MSVS did you managed to compile it? I tried cmake 
> build with 2010 express and something gave an error. (2008 works)
>
> It comes out like a "undefined behavior" rigth? I watched a 
> presentation on the subject "Here Be Dragons: C++ Undefined Behavior". 
> Check it out, if it bring you an idea, its useful to know it.
>
> Since MinGW didn`t detect this, I think it`s not a significant thread. 
> May be MSVC knows better but we can leave it with #ifdef until we 
> solve it betterwise.
>
> I can ask my lectors about this problem but won`t be on my PC to try 
> it for 2 weeks.
>
> -Dido
>
>
> >-------- Оригинално писмо --------
> >От: Matthew Zweig
> >Относно: Re: [epiar-devel] The successes and failures of a night with 
> MSVC
> >До: epiar-devel at epiar.net
> >Изпратено на: Сряда, 2010, Октомври 6 09:37:11 EEST
>
> >I've created Trac Ticket #96 for this issue.
> >
> >I suggest that we add a #ifdef for MSVC that removes this line. That 
> will mean that that Windows builds will leak memory, but at least they 
> will run. The memory that's being leaked really isn't that much, but 
> it's still a leak.
> >
> >~Matt (knowknowledge)
> >
> >On Oct 5, 2010, at 11:14 PM, Christopher Thielen wrote:
> >
> >> Gather 'round children young and old, for we have a story to tell,
> >>
> >> (okay, I'll drop that pretense)
> >>
> >> Matt and I spent about 2-3 hours working on the MSVC Windows crashing
> >> bug, the one that only appears if you compile via MSVC and not MinGW.
> >>
> >> We were able to find a single line, which, when removed, caused the
> >> problem to go away. Unfortunately, this line was simply a memory leak,
> >> so our massive memory issue (I say massive, it's really just a double
> >> delete or something - my best guess) was of course contained by 
> avoiding
> >> entirely. Here's the line:
> >>
> >> UI/ui_container.cpp ::DelChild() -> comment out the delete (*i) 
> line in
> >> the for loop. Recompile and enjoy.
> >>
> >> We're not sure what to do about this right now. If Arke or somebody 
> else
> >> really experienced with MSVC (MSVS?) debugging wants to tackle 
> this, we
> >> would be more than glad to help out. We're not MSVS experts and at 
> this
> >> point, we've decided to go ahead and rely on other paths for the time
> >> being for Windows builds. This is not to say MSVC does not work (it
> >> compiles fine and obviously runs without that) but we're reaching the
> >> pragmatic limits of our debugging abilities.
> >>
> >> (We're sending this to the list so people can see this and to document
> >> it for future reference, but please, feel free.)
> >>
> >> -topher (and a disgruntled knowknowledge)
> >> _______________________________________________
> >> epiar-devel mailing list
> >> epiar-devel at epiar.net
> >> http://epiar.net/mailman/listinfo/epiar-devel
> >
> >
> >_______________________________________________
> >epiar-devel mailing list
> >epiar-devel at epiar.net
> >http://epiar.net/mailman/listinfo/epiar-devel
> >
>
>
> _______________________________________________
> epiar-devel mailing list
> epiar-devel at epiar.net
> http://epiar.net/mailman/listinfo/epiar-devel
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://epiar.net/pipermail/epiar-devel/attachments/20101006/a8588209/attachment.html