Re: [Boost-docs] [debugger_visualizers] wiki update

Subject: Re: [Boost-docs] [debugger_visualizers] wiki update
From: Filip Konvička (filip.konvicka_at_[hidden])
Date: 2007-08-06 20:44:08

Cédric Venet 6.8.2007 21:18:
>> Also, I think that the visualizers we've created so far are quite
>> satisfactory, so I don't know why would you want anything more
>> "advanced".
>>     
>
> I didn't says they weren't, just you wrote on the wiki:
>
> Microsoft Visual Studio 2005 is the first of the Visual Studio versions to support user-customizable debugger visualizers. For CLR languages, such as C#, the possibilities are even more powerful, and documented. Native C++ debugger visualizers are not documented, though it was claimed by some people from Microsoft that some docs are coming - so if you know of some, please let us know!
>
> So I posted.
>   
Yeah, I actually meant some docs on autoexp.dat's [Visualizer] section 
(that the wiki is about), and this is not documented AFAIK.
So you are of course right, those visualizers that you mention are 
indeed documented and available since VC 6.
>> Did you read the discussion "[Boost-users] [multi_index] MSVS 2005
>> visualizers?" at the boost-users mailing list, BTW?
>>     
>
> I haven't subscribed to the users mailling list. Perhaps I should, but I was fearing the quantity of mail and a very bad signal/nose ratio.
>   
Don't fear, there's very little noise and some discussions are very 
interesting. Just look at the archives.
>> Thanks for the links anyway, I haven't tried writing a "DLL" visualizer
>> yet. We might use it for some Boost classes, such as ptime. However, we
>> must first examine how does it integrate with the IDE (whether it works
>> in the watch window, etc.). It might be that this feature no longer
>> works in new VS editions, as there is a brand new visualizer API for
>> CLR debugging.
>>     
>
> The API has been upgraded for VS .NET and is reported to work on VS .NET 2005. And the CLR api will probably never be usable with c++ (not CLI) because it use CLR and associated technology (speculation) 
>   
That's very good news. I guess that some classes will need some C++ code 
executed to display something useful. (I succeeded running a function to 
display ptime as a string in a watch window, but sadly this does not 
work with autoexp.dat.)

You're right about the new CLR API, this uses reflection and all that stuff.

Cheers,
Filip