[Patch] libstdc++/14493

[Benjamin Kosnik]

>> it looks like mainline is currently wrong: what's in 
>> GLIBCXX_3.4.10 should be in GLIBCXX_3.4.9 (and those changes should go 
>> to gcc-4_2-branch) and the new changes that you propose should also go 
>> in GLIBCXX_3.4.9 (and also go in gcc-4_2-branch.)
> 
> We never discussed in the open this kind of policy. In my opinion, it 
> makes the life of the maintainers simpler, but, on the other hand, it's 
> risky and/or blocks unrestrained development in newer branches.

I agree that it's simpler for the maintainers. This is good, and the 
only sane solution IMHO. Versioning is hard: why make it harder?

I don't think it blocks unrestrained development in new branchers: how?

>  I don't 
> think this is the obvious, general policy of GCC. 

Oh yeah, the policy.

;)

 > Indeed, in this
> specific case, this kind of sync implies that slightly experimental 
> stuff we have in mainline (e.g., the debug-mode multithreading fixes, 
> useful but not a regression) must necessarily go immediately in 4.2.0 
> branch and the present fix too. 

Yeah. I think the debug-mode MT fixes are fine and stand behind your 
work: that should be in 4.2.0 IMHO. Same for the rest of the added symbols.

 > Since you are of this opinion (and,
> supposedly, will share the responsability for the corresponding action 
> ;) I will take care of the whole sync work, but still, I'm not 
> completely happy with this kind of general policy (for the future)

Of course I am willing to share responsibility.... merely stating my 
opinion on the current state.

I'm not necessarily suggesting this as a future, blanket policy. Or even 
that there is any blanket policy at the moment! Or, should be.


> P.S. That "Capiche" sounded totally weird to me, then I found this:
> 
>    http://en.wikipedia.org/wiki/Capiche
> 
> Actually, the spelling is wrong, would be "Capisce?", or, confidential, 
> "Capisci?". Well, in general, it's slightly more polite to say, 
> "D'accordo?", because it means "Agreed?", therefore there is no 
> ambiguity that the listener is asked whether he *understands* the matter 
> vs agrees about the best solution.

Ah. Serves me right to spread confusion when I, now corrected, say: 
D'accordo?

Thanks for this: quite informative.

best,
benjamin

>