Namespace association: ODR violation made easy

Gabriel Dos Reis gdr@integrable-solutions.net
Mon Mar 1 22:22:00 GMT 2004


Paolo Carlini <pcarlini@suse.de> writes:

| Gabriel Dos Reis wrote:
| 
| >A recent PR about "user forward declaring" entities in namespace std
| >reminded me about a message I sent to Benjamin, Matt and Jason more
| >than two months ago.  Since 3.4.0 is imminent, I think something
| >should go in our docs; I don't what, but I believe people may soon
| >start reporting PRs about.
| >
| I completely agree with you, given, in particular, that the PR you are
| hinting to (14370), was distilled from "Boost" and that the EDG front
| end accepts it in strict mode.... Ergo the issue seems not completely
| trivial ;)

Well, EDG in strict mode still accepts codes that trigger undefined
bahaviour :-) [and it does not come with a standard library, therefore
it ought to support those declarations in some ways]
But, yes, people at Boost should not have been writing codes like that
-- they know better ;-)  


I'm still looking for comments on the mail I forwarded though.
(Given that it was written somewhere near Xmas, I can't guarantee it
being completely accurate :-))

-- Gaby



More information about the Libstdc++ mailing list