Board index » cppbuilder » Re: Borland Bashing- is it productive at this point?

2005-04-08 08:28:50 PM
cppbuilder108
"Hendrik Schober" < XXXX@XXXXX.COM >wrote in message
templates to my projects. Some didn't work with
BCB, some worked but caused my build times to
sky rocket or my exe size to increase too much.
Some worked sometimes, depending on how the
linker felt that day. Additionally, it caused no end
of problems with "cannot create precompiled header"
stuff.
It's not so much that I think Borland should shoot for
100% boost compatability, but if they do, they will
more than likely have to handle templates a lot
better.
 

Duane Hebert < XXXX@XXXXX.COM >wrote:
is pretty good. Nevertheless, that it does
compile a certain state of the boost lib
does not mean it won't {*word*88} on templates
that are as complex as the ones found in
boost.
Schobi
--
XXXX@XXXXX.COM is never read
I'm Schobi at suespammers dot org
"The presence of those seeking the truth is infinitely
to be prefered to those thinking they've found it."
Terry Pratchett
 

Daniel James < XXXX@XXXXX.COM >wrote:
is, that in general it should allow some
optimizations, while with VC using it is a
pessimization.
implemented.
about 'export', but I have seen his article
from one or two years ago, which he also
presented to the std committee. FWIW, this
article seems pretty biased. Also, it was
voted down by almost everyone at that
meeting.
You might want to read something about how
useful 'export' could be in this thread:
groups.google.com/groups
types for about 10 years because people
actually never used them? Or do you mean
holding people back from actually using
boost or other great libs?
Or are you talking of more current
decisions -- like the one not to support
two-phase lookup? This alone causes us a
lot of pain and eats our precious time.
I'd call that stupid, not pragmatic.
--
XXXX@XXXXX.COM is never read
I'm Schobi at suespammers dot org
"The presence of those seeking the truth is infinitely
to be prefered to those thinking they've found it."
Terry Pratchett
 

{smallsort}

Daniel James wrote:
implemented in MSVC last time I checked, although it did accept the
syntax.
Similar to MS approach to export, they have taken the pragmatic route
of focussing on perceived need rather than blind 100% conformance. It
has paid off very handsomely, in that VC7.1 is a world apart from
previous versions.
a popular misconception that empty throw allows compilers to make added
assumptions. In fact it is quite the opposite - it forces them to
verify your assumptions at runtime. Hopefully we will be able to do
something about this in C++0x - whenever that is <g>
I brought the subject up they were quite to to explain to me why I
really didn't want them anyway - that they would bring me more trouble
than help and break far more existing code written without this feature
that 'just works' today.
They were clear that they had as much intention at the time of
supporting this as the export keyword. They are always open to
changing their stance due to overwhelming customer demand - otherwise
there were far more useful things to be working on.
AlisdairM(TeamB)
 

Hendrik Schober wrote:
compiler.
AlisdairM(TeamB)
 

In article < XXXX@XXXXX.COM >,
mr_organic < XXXX@XXXXX.COM >wrote:
in the shuffle, but Intel is definitely more conformant than VC++.
Additionally, Comeau C++ is the most conformant on any platform,
including Windows.
--
Greg Comeau / Comeau for the Mac? Stay tuned.
Comeau C/C++ ONLINE ==>www.comeaucomputing.com/tryitout
World Class Compilers: Breathtaking C++, Amazing C99, Fabulous C90.
Comeau C/C++ with Dinkumware's Libraries... Have you tried it?
 

In article <42558826$ XXXX@XXXXX.COM >,
mr_organic < XXXX@XXXXX.COM >wrote:
However, it's my belief that VC++ will never be ISO conformant.
--
Greg Comeau / Comeau for the Mac? Stay tuned.
Comeau C/C++ ONLINE ==>www.comeaucomputing.com/tryitout
World Class Compilers: Breathtaking C++, Amazing C99, Fabulous C90.
Comeau C/C++ with Dinkumware's Libraries... Have you tried it?
 

In article < XXXX@XXXXX.COM >,
Alisdair Meredith <alisdair.meredith@ XXXX@XXXXX.COM >wrote:
job and should be commended. However, I think adding in
about Boost in a same discussion about compliance is changing
the topic, as although Boost does stress compilers, especially
with templates, much of Boost is not directly standards related,
and a good bit of it has nothing to do with it at all.
There has also been aspects of focusing on one compiler and
not another. Note: _None_ of this is to negate the accomplishments
of either MS or the Boost folks.
--
Greg Comeau / Comeau for the Mac? Stay tuned.
Comeau C/C++ ONLINE ==>www.comeaucomputing.com/tryitout
World Class Compilers: Breathtaking C++, Amazing C99, Fabulous C90.
Comeau C/C++ with Dinkumware's Libraries... Have you tried it?
 

In article < XXXX@XXXXX.COM >,
Randall Parker < XXXX@XXXXX.COM >wrote:
that was no the case for any compiler, so IOWs, why should it be
the case for Borland? That said, I do know that the Boost folks
made considerable effort to get rid of many such concerns, and IMO
were successful to the extent that that is possible.
but management. Go back and you will find similar discussions
there as you are finding here.
And not only has C++ not gone away, but after a while of
plateau'ing, it seems to be on the rise again by all inidcations,
it's a good time again for any company to re/jump on board,
if not even too late already.
answer to a questions like that.
--
Greg Comeau / Comeau for the Mac? Stay tuned.
Comeau C/C++ ONLINE ==>www.comeaucomputing.com/tryitout
World Class Compilers: Breathtaking C++, Amazing C99, Fabulous C90.
Comeau C/C++ with Dinkumware's Libraries... Have you tried it?
 

In article < XXXX@XXXXX.COM >,
Alisdair Meredith <alisdair.meredith@ XXXX@XXXXX.COM >wrote:
--
Greg Comeau / Comeau for the Mac? Stay tuned.
Comeau C/C++ ONLINE ==>www.comeaucomputing.com/tryitout
World Class Compilers: Breathtaking C++, Amazing C99, Fabulous C90.
Comeau C/C++ with Dinkumware's Libraries... Have you tried it?
 

In article < XXXX@XXXXX.COM >,
vavan < XXXX@XXXXX.COM >wrote:
and it's not. Of course I'm not saying it has nothing to do with it,
it clearly has lots to do with it.
--
Greg Comeau / Comeau for the Mac? Stay tuned.
Comeau C/C++ ONLINE ==>www.comeaucomputing.com/tryitout
World Class Compilers: Breathtaking C++, Amazing C99, Fabulous C90.
Comeau C/C++ with Dinkumware's Libraries... Have you tried it?
 

In article <42564c7f$ XXXX@XXXXX.COM >,
Hendrik Schober < XXXX@XXXXX.COM >wrote:
made significant differences. That said, they also definitely
have their own motivations (but who doesn't), but also their
own thought processes on the issues which I either don't
agree with or understand, but it's not for me to do so probably.
--
Greg Comeau / Comeau for the Mac? Stay tuned.
Comeau C/C++ ONLINE ==>www.comeaucomputing.com/tryitout
World Class Compilers: Breathtaking C++, Amazing C99, Fabulous C90.
Comeau C/C++ with Dinkumware's Libraries... Have you tried it?
 

In article <42565294$ XXXX@XXXXX.COM >,
Hendrik Schober < XXXX@XXXXX.COM >wrote:
necessarily equate to less manpower per se. It's nice
to think maybe of being X times more productive than others,
but cutting through red tape etc has its place as well.
In the end, it's still decades of man years and millions
of dollars of expenses.
--
Greg Comeau / Comeau for the Mac? Stay tuned.
Comeau C/C++ ONLINE ==>www.comeaucomputing.com/tryitout
World Class Compilers: Breathtaking C++, Amazing C99, Fabulous C90.
Comeau C/C++ with Dinkumware's Libraries... Have you tried it?
 

"Hendrik Schober" < XXXX@XXXXX.COM >wrote in message
work with the boost stuff, it will mean making templates
work better for non boost stuff. It won't mean that all template
problems will be solved but it should mean that many
would be.
 

In article < XXXX@XXXXX.COM >,
Daniel James < XXXX@XXXXX.COM >wrote:
I think Alisdair is as correct. IOWs, hard to process
and conflicting statements have been offered by MS.
Have you seen www.comeaucomputing.com/iso/promises.html and others?
Also, most of Herb's articles on export has nothing to do with
export and more to do with C++ in general and every feature
in general (IOWs, the article does not actually offer what it claims
to be doing either, and yet most people believe that it does).
of all features. :)
I doubt anything will ever disappear. In fact, the committee
has made no efforts to even deprecate these things,
and furthermore, really doesn't even want to talk about them.
--
Greg Comeau / Comeau for the Mac? Stay tuned.
Comeau C/C++ ONLINE ==>www.comeaucomputing.com/tryitout
World Class Compilers: Breathtaking C++, Amazing C99, Fabulous C90.
Comeau C/C++ with Dinkumware's Libraries... Have you tried it?
 

XXXX@XXXXX.COM (Greg Comeau) wrote in news:d366s5$lhv$1
@panix3.panix.com:
guess I was trying to make a point about commercial/big-name compilers,
not *any* compiler out there.
But I do think that constantly harping on "conformancy" misses a lot of
the point -- a compiler can be very conformant to the standard, and yet
still suck. It can generate bloated/slow code, it can generate
*erroneous* code, it can have limited configurability, and so on. That's
why it grates on me a bit when people tend to use the conformancy
argument as an indicator of the *quality* of a given compiler. It's one
measure, sure; but it's only one of many. And -- I'd argue -- not even
the most important one.
Just my .02, anyway.
mr_organic