BOOST_FOREACH review results
Hi, everybody BOOST_FOREACH macro by Eric Niebler is accepted into boost. The library was received pretty good by boost community (I counted at least 13 positive reviews). The only two negative reviews based their rejection on principle: "Macros are evil and so should not be used". While this maybe interesting point in itself, as I see things now, boost practice supports using macros where necessary and macro nature of the tool could not be a compelling enough reason to reject the submission. Would this trend be ever changed in a future we could change a policy for future submissions. While there were not so many specific issues discovered during review, I would like to list couple: * documentation should contain more examples and special pitfalls section * testing should be split by compatibility sections * it would be good to make this facility work on more compilers Congratulation Eric for the excellent work and thanks to all the reviewers for participation. Gennadiy Rozental BOOST_FOREACH review manager
"Russell Hind" wrote:
I take it this will be for the 1.34.0 release of boost as 1.33 is in feature freeze?
Perhaps exception could be made - it is mature, used in apps and the documentation can be updated quickly. It may serve as starting point for people exploring other libraries. /Pavel
Pavel Vozenilek wrote:
Perhaps exception could be made - it is mature, used in apps and the documentation can be updated quickly.
It may serve as starting point for people exploring other libraries.
I would like to see it in there too. AFAICT, it doesn't require any changes to other libraries (the bcc32 addressof problem has already been patched in CVS) so its impact should be minimal to none on regression tests. I guess its would still then be up to Eric to decide if the current tests and docs are where he'd like them to be for general release after the review. Cheers Russell
"Russell Hind" <rh_gmane@mac.com> wrote in message news:d5d08i$99a$1@sea.gmane.org...
Pavel Vozenilek wrote:
Perhaps exception could be made - it is mature, used in apps and the documentation can be updated quickly.
It may serve as starting point for people exploring other libraries.
I would like to see it in there too. AFAICT, it doesn't require any changes to other libraries (the bcc32 addressof problem has already been patched in CVS) so its impact should be minimal to none on regression tests.
I guess its would still then be up to Eric to decide if the current tests and docs are where he'd like them to be for general release after the review.
The Release Manager would also have to agree. And there should be a posting with a specific subject line so others not reading this thread can comment. --Beman
participants (5)
-
Beman Dawes -
Eric Niebler -
Gennadiy Rozental -
Pavel Vozenilek -
Russell Hind