Pretty printers for versioned namespace

[Jonathan Wakely]

On 13/02/17 21:04 +0100, François Dumont wrote:
>On 31/01/2017 15:42, Jonathan Wakely wrote:
>>
>>>I don't know what is the impact of it but is there a plan to allow 
>>>version 2.69 to work out of the box ?
>>
>>No. It's fairly easy to use the right versions.
>
>    Just for my info, is there some blog/doc explaining why not 
>upgrading to latest version ? Aren't those tools better in latest 
>versions ?

There's nothing wrong with the versions we use, and upgrading can
cause problems. I don't know the details. But it's not difficult to
use the right versions.

>    Otherwise thanks for your tip on how to manage this, it worked fine.
>
>    Attached is the latest patch. It is working fine with and without 
>versioned namespace.
>
>    However to validate it I had to manually change default 
>use_versioned_namespace value in register_libstdcxx_printers function. 
>This is because when running pretty printers tests, in the generated 
>.gdb file we have:
>
>python register_libstdcxx_printers(None)
>
>    So always using default value (False) even if versioned namespace 
>have been activated. Can you advise on how to consider this when 
>running testsuite pretty printers ? Using again configuration ? 
>Looking for some definition in __7:: namespace in gdb-test.exp 
>gdb-test proc and adapt register_libstdcxx_printers call depending on 
>the compilation result ?

Ah, I forgot about that generated file. Something like that should
work.