Boost support for Boalrnd to be dropped.

[John Grabner]

I have just seen on the boost developer list that they intend to drop
support for borland.

news://news.gmane.org:119/uy7ulu...@boost-consulting.com

Alisdair, Nicoli do you have any information on this?

John.

[Dennis Jones]

"John Grabner" <grab...@ihug.co.nz> wrote in message
news:44c4...@newsgroups.borland.com...

Well, the question I would ask is, "What does 'support for Borland' mean
anyway?" Presumably, it means that the Boost developers will no longer be
required (or willing) to spend inordinate amounts of time trying to find
workarounds to allow the compiler to successfully compile code that also
executes as expected. If the compiler were more up-to-date with regard to
standards compliance, or if Borland were at least able to demonstrate a
commitment toward achieving standards compliance, this wouldn't even be an
issue -- but it isn't and they haven't (at least not satisfactorily).

It seems to me that if Borland had put as much effort into bringing the
compiler into compliance with current standards as they did trying to make a
whiz-bang, wow-the-audience sort of IDE, we wouldn't be in this sorry
predicament. Of course, IDE improvements are always nice, and usually
welcome - just not at the expense of stability and language compliance, both
of which (if these newsgroups are any indication) have been the chief
complaints of BCB users for the last several releases (I'm thinking back to
about BCB4, which was, I think, the first time I ever saw what looked like a
step backwards for Borland -- and it's been 'two steps forward, one step
back' with every release since). Borland is now stating that language
compliance will be a top priority for Highlander. Well, gee, that's
great -- but dang it, it could be too little, too late this time.

The way I see it, in a worst-case scenario, Boost discontinues support for
Borland. Fortunately, that doesn't necessarily mean we can't use Boost
anymore -- just that we may have to work a little harder. If we're lucky
however, Borland will make good on their promise of better language
compliance with Highlander. If they do that and Boost should happen to
discontinue support for Borland in the meantime (and assuming that Borland
hasn't also fallen too far behind additional language changes in the
meantime), the Boost developers will probably re-instate support for
Borland.

If we're _really_ lucky, Boost won't discontinue support for Borland at all.
Even better, if Borland could do something more immediate to improve
BDS2006's language compliance, and therefore its ability to successfully
pass the Boost tests right now -- oh nevermind, that would be the equivelent
of winning the California state lottery -- but we can dream, right? After
all, isn't that why you buy those lottery tickets? ;-)

- Dennis

[Alisdair Meredith[TeamB]]

John Grabner wrote:

Well, support for older compilers will be dropped after the NEXT Boost
release, so there should be no issues in 1.34.

After that, Boost will drop official support for MSVC6 and 7.0,
GCC2.95, CodeWarrior 8.3, and BCB6. BCB2006 will be supported until
the next Borland compiler ships, although I would expect support to
drop shortly after that (as the new compiler takes over)

Dropping support means the compiler is no longer a required part of the
regression tests for each release, and library authors may be less
inclined to accept patches that only fix issues with these older
compilers. Boost will not suddenly stop working, but may degrade over
time.

I personally think it would be better to really break support after
1.34 with a #error in the config pointing to the last supported
release, but we'll see what happens with 1.35, which at this rate is at
least 12 months away, given 1.34 hasn't been release yet!

--
AlisdairM(TeamB)