 |
BorlandTalk.com
Borland discussion newsgroups
|
| View previous topic :: View next topic |
| Author |
Message |
JD
Guest
|
 Posted: Fri Nov 04, 2005 1:43 pm Post subject: Re: Dexter and the C++ standard |
 |
|
Edward Diener <eddielee_no_spam_here (AT) tropicsoft (DOT) com> wrote:
Those of us who are beta testing Dexter are not even permitted
to say that we are a beta tester. You'll get no information
from any one who is not a Borland employee. I would suggest
send an email to Tim Del Chiaro and ask him instead.
[email]tdelchiaro (AT) borland (DOT) com[/email]
~ JD
|
|
| Back to top |
|
 |
Edward Diener
Guest
|
| Posted: Fri Nov 04, 2005 2:01 pm Post subject: Dexter and the C++ standard |
|
|
I am interested to know if the C++ compiler for the next version of C++
Builder, C++ Builder 2006, has improved significantly in its adherence
to the C++ standard. Does anybody know ? Can the many Boost libraries
which currently can not be compiled with C++ Builder 6 now be compiled
successfully with the C++ compiler in Dexter aka C++ Builder 2006 ?
I listened to Eli Boling in the 24 hours to Dexter and had the distinct
impression that the entire issue of C++ Builder's adherence to the C++
standard was ignored and the entire issue deceptively avoided.
I would like to think that after 4+ years of doing nothing with C++
standards compliance in C++ Builder 6 and the 2 years before that of
doing next to nothing with C++ standards compliance in C++ Builder 5,
essentially 6 years of almost absolute stasis in this area, that Borland
would have made a major attempt at getting their compiler approaching
closely to the C++ standard, so that advanced C++ libraries could be
compiled with the C++ compiler in Dexter without the usual headaches.
But almost total avoidance of this issue in the Dexter hype has led me
to believe that, once again, Borland has totally ignored their C++
customers who might actually want, hard to believe, a C++ compiler that
actually compiles 98%+ of the C++ language.
|
|
| Back to top |
|
|
Zach Saw
Guest
|
| Posted: Fri Nov 04, 2005 4:00 pm Post subject: Re: Dexter and the C++ standard |
|
|
I asked a question regarding the compliance of the new C++ compiler, and the
answer given was Borland will test the compiler using the plum hall C++ test
suite. Looking at the plum hall related bugs that are deferred to the next
release (i.e. High-Lander) in QC, it is quite simple to answer your
question.
** disclaimer: this is my own deduction from QC and therefore is completely
my own opinion -- my 2 cents **
There'll still be a lot of compliance issues with the new C++ Compiler. In
fact, looking at QC, there're still going to be quite a few syntax that will
break the compiler (read: cause an internal error in the compiler).
Regarding Boost though, this is what I gather from my near 24 hours of
attendance during the 24h of Dexter broadcast: It will require a few tweaks
in the Boost library to get the new compiler to compile it (someone said
this during the broadcast, don't remember who though).
In all fairness, Borland nearly dropped BCB completely (at one point,
everyone agreed it has been dropped and majority of us went for MSVC). It
wasn't until recently BCB was picked up again, and therefore the same
applies for the C++ compiler. I heard Eli saying something on the broadcast
version 6 was put on hold, and they brought back version 5, and it'll
probably be version 5.8 or something. It will be a version that contains
mostly fixes, while version 6 will probably contain more upgrades in terms
of compliance and performance (my own guess here).
The Borland guys are probably busy preparing for DevCon as we speak, so you
might not get a prompt reply from Tim. But whatever is the reply, hope you
will share with us in this thread.
|
|
| Back to top |
|
|
Alisdair Meredith [TeamB]
Guest
|
| Posted: Fri Nov 04, 2005 5:37 pm Post subject: Re: Dexter and the C++ standard |
|
|
Edward Diener wrote:
| Quote: | But almost total avoidance of this issue in the Dexter hype has led me
to believe that, once again, Borland has totally ignored their C++
customers who might actually want a C++ compiler that actually
compiles 98%+ of the C++ language.
|
All I can say is 'trust me, they know'.
Efforts are definitely being made in this direction. Quality is one of
the main focusses of the DeXter release, and surely that must include
fixing the compiler? [Personally I don't see language conformance as a
feature - rather every failure to conform is a bug report]
I don't know how much progress we will see in the final release, but I
am sure the answer will be 'some'. Not 100% this time round, maybe not
as much progress as some of us would like (after all, I would like
100%!) but certainly more than 'zero' progress as well.
I would worry about the tendency to use Boost as the benchmark that
counts though. It is a library collection, nothing more and certainly
nothing less. It is demanding and stresses compilers in many ways, but
it is not an exhaustive conformancy test.
Why does that matter? The bug that most affects me personally, and
would probably block our own adoption of DeXter, does not show up in
the boost tests (QC2629) I would much rather this bug was fixed than
the compiler passed the boost::Spirit test, although naturally I would
prefer both ;¬)
--
AlisdairM(TeamB)
|
|
| Back to top |
|
|
Dennis Jones
Guest
|
| Posted: Fri Nov 04, 2005 5:48 pm Post subject: Re: Dexter and the C++ standard |
|
|
"JD" <nospam (AT) nospam (DOT) com> wrote
| Quote: |
Those of us who are beta testing Dexter are not even permitted
to say that we are a beta tester.
|
Does anyone else find this hilarious? Need I say more?
- Dennis
|
|
| Back to top |
|
|
Edward Diener
Guest
|
| Posted: Fri Nov 04, 2005 5:49 pm Post subject: Re: Dexter and the C++ standard |
|
|
Zach Saw wrote:
| Quote: | I asked a question regarding the compliance of the new C++ compiler, and the
answer given was Borland will test the compiler using the plum hall C++ test
suite.
|
If Borland can not give a straight, simple answer then it is obvious to
me they have done nothing, and have done their best through the whole 24
hour Dexter circus not to admit to their failing to improve the compiler
conformance in any significant way. They've wasted 6 years because they
have no C++ compiler expertise among their developers and it looks like
they will continue to delay and waste as many years as they can, while
still selling their product, not doing anything about the C++
conformance. What a sorry state of affairs !
|
|
| Back to top |
|
|
Rudy Velthuis [TeamB]
Guest
|
| Posted: Fri Nov 04, 2005 6:05 pm Post subject: Re: Dexter and the C++ standard |
|
|
At 18:48:12, 04.11.2005, Dennis Jones wrote:
| Quote: | Those of us who are beta testing Dexter are not even permitted
to say that we are a beta tester.
Does anyone else find this hilarious?
|
I don't.
No, no need.
--
Rudy Velthuis [TeamB] http://velthuis.homepage.t-online.de
"Nietzsche was stupid and abnormal."
- Leo Tolstoy (1828-1910)
|
|
| Back to top |
|
|
Edward Diener
Guest
|
| Posted: Fri Nov 04, 2005 6:10 pm Post subject: Re: Dexter and the C++ standard |
|
|
Dennis Jones wrote:
| Quote: | "JD" <nospam (AT) nospam (DOT) com> wrote in message
news:436b739b$1 (AT) newsgroups (DOT) borland.com...
Those of us who are beta testing Dexter are not even permitted
to say that we are a beta tester.
Does anyone else find this hilarious? Need I say more?
|
No, your point is well taken.
What is even more hilarious, and absurd, is that in order to find out
about the C++ conformance of Dexter one must listen carefully only to
Borland's obviously slanted hype about the product only to find out that
nothing whatsoever has been said about it AND no one else who may know
anything about it is allowed to say anything about it either !!!
But this tells me all I need to know anyways.
|
|
| Back to top |
|
|
Chris Uzdavinis (TeamB)
Guest
|
| Posted: Fri Nov 04, 2005 6:36 pm Post subject: Re: Dexter and the C++ standard |
|
|
Edward Diener <eddielee_no_spam_here (AT) tropicsoft (DOT) com> writes:
.... snip...
| Quote: | They've wasted 6 years because they have no C++ compiler expertise
among their developers
|
I think this is a grossly unfair and insulting mischaracterization of
Borland's employees. Are you trying to get a rise out of us by saying
such things?
--
Chris (TeamB);
|
|
| Back to top |
|
|
Edward Diener
Guest
|
| Posted: Fri Nov 04, 2005 7:47 pm Post subject: Re: Dexter and the C++ standard |
|
|
Chris Uzdavinis (TeamB) wrote:
| Quote: | Edward Diener <eddielee_no_spam_here (AT) tropicsoft (DOT) com> writes:
... snip...
They've wasted 6 years because they have no C++ compiler expertise
among their developers
I think this is a grossly unfair and insulting mischaracterization of
Borland's employees. Are you trying to get a rise out of us by saying
such things?
|
I am trying to get a rise out of Borland, and vent my anger that after
years of stasis they still avoid even talking about any efforts to fix
compiler bugs or improve their compiler standards adherence. This
refusal to even address the issue, and to present Dexter hype which so
carefully avoided addressing it, leads me to believe that they just do
not have the expertise among their developers to improve their compiler
and/or do not feel it worhtwhile to do so or even discuss the issue.
They are the ones who should feel ashamed, not me.
|
|
| Back to top |
|
|
Zach Saw
Guest
|
| Posted: Fri Nov 04, 2005 7:52 pm Post subject: Re: Dexter and the C++ standard |
|
|
| Quote: | If Borland can not give a straight, simple answer then it is obvious to me
they have done nothing, and have done their best through the whole 24 hour
Dexter circus not to admit to their failing to improve the compiler
conformance in any significant way. They've wasted 6 years because they
have no C++ compiler expertise among their developers and it looks like
they will continue to delay and waste as many years as they can, while
still selling their product, not doing anything about the C++ conformance.
What a sorry state of affairs !
|
Actually, I think they made it pretty clear that it will not be a bug-free
product during the 24h of Dexter.
In fact, you can search for the known bugs in QC, or listen to the QA
manager's session. I think this time around, they have actually been pretty
open about their approach of selling the new product.
I don't agree with everything they have said in the broadcast myself,
especially in the part where they said they're the first users of their own
product -- that only applies for Delphi. We C++ developers have long been
neglected by Borland because we are not the cash cow. Borland is a mid-size
company -- and by that, I mean it does not have the vast resources of (for
example) Microsoft, where they can deliver products losing money every
quarter, with a simple vision of being the leader in the market after 15 or
20 years. The big size companies have products that provide the mainstream
income, which funds the rest of the losing businesses. This is a luxury
Borland does not have.
Do not forget that Borland left us high and dry long ago -- if you search
the old posts in the newsgroup, you'll find a lot of faithful BCB developers
left for Microsoft. In fact, when we tried to sell our VCL components to
software houses, and persuaded them to use BCB instead, they simply brushed
us off -- "Oh! There are actually people still using C++ Builder?" Most of
the software houses use Microsoft products now, and it was simply impossible
to try to convert them. I personally have been a victim of Borland's
negligence -- we nearly went out of business (actually, we that part of our
business did vanish with BCB's discontinuation) because our clients simply
dumped BCB and went for Microsoft. Since all our VCL components are written
in BCB, our client base dropped to nearly zero in a few weeks after
C++BuilderX was announced to have no support for VCL. I believe Borland had
no plans to resurrect BCB until Paul gathered the signature of a few
hundreds of developers in renowned corporates and government agencies. It
wasn't until then BCB was in Borland's plans. With that in mind, the BCB
community might have lost its mass and Borland will need to spend a lot more
into marketing Dexter, if they want to gain back the market share they used
to have, which is ironic. Bringing back BCB actually hurts Borland more
financially in the long run, than to cut it off and ignore us. With that,
they definitely deserve some credits. It is a bold move, something that I
personally hope will pay off for Borland and make BCB a cash cow for
Borland, finally. It may have been too late in our case, but I hope it is
not for the rest.
The table has definitely turned -- Borland used to be the leader in software
development products, Microsoft only had BASIC. I can't help but believe in
the forecast made by some analysts that in the future, all software
businesses will converge, and the strongest will be the only one left. I
don't like the thought of it, as I run my own software house. Definitely
something we can't ignore, given the events that have been happening in
recent years.
Sorry for the length of this post. Just trying to point out the obvious.
|
|
| Back to top |
|
|
Edward Diener
Guest
|
| Posted: Fri Nov 04, 2005 8:05 pm Post subject: Re: Dexter and the C++ standard |
|
|
Alisdair Meredith [TeamB] wrote:
| Quote: | Edward Diener wrote:
But almost total avoidance of this issue in the Dexter hype has led me
to believe that, once again, Borland has totally ignored their C++
customers who might actually want a C++ compiler that actually
compiles 98%+ of the C++ language.
All I can say is 'trust me, they know'.
Efforts are definitely being made in this direction. Quality is one of
the main focusses of the DeXter release, and surely that must include
fixing the compiler? [Personally I don't see language conformance as a
feature - rather every failure to conform is a bug report]
|
Then why was this completely avoide and skirted around in the 24 hours
to Dexter, and why has Borland avoided saying anything about it in any
announcements ? Because they have made significant strides ? I doubt it.
| Quote: |
I don't know how much progress we will see in the final release, but I
am sure the answer will be 'some'. Not 100% this time round, maybe not
as much progress as some of us would like (after all, I would like
100%!) but certainly more than 'zero' progress as well.
|
You are being much too 'nice' about it. I know you chided Borland about
it in your segment of the 24 hours but in such gentle terms that one
doubts that it made a dent in their consciousness.
| Quote: |
I would worry about the tendency to use Boost as the benchmark that
counts though. It is a library collection, nothing more and certainly
nothing less. It is demanding and stresses compilers in many ways, but
it is not an exhaustive conformancy test.
|
Completely agree, which is why reason why Boling's vague remarks about
the C++ Compiler "perhaps" working better with Boost meant nothing. But
Boost has brought out many compiler bugs in BCB6 so paying attention to
them would have provided an immediate list of things which needed to be
fixed in their C++ compiler. How about a remark that they fixed all the
C++ compiler bugs reported on QC for BCB6, as well as all of the C++
issues reported on the Boost NGs and in Boost tests regarding Borland's
latest compiler, and had upgraded the conformance of their compiler
significantly ? Their reticence to say anything, since after all this
can't be a company secret they want to keep, leads me to the obvious
conclusion.
| Quote: |
Why does that matter? The bug that most affects me personally, and
would probably block our own adoption of DeXter, does not show up in
the boost tests (QC2629) I would much rather this bug was fixed than
the compiler passed the boost::Spirit test, although naturally I would
prefer both ;¬)
|
I got tired of reporting QC issues long ago, and seeing nothing get done.
|
|
| Back to top |
|
|
Zach Saw
Guest
|
| Posted: Fri Nov 04, 2005 8:10 pm Post subject: Re: Dexter and the C++ standard |
|
|
| Quote: | I don't know how much progress we will see in the final release, but I
am sure the answer will be 'some'. Not 100% this time round, maybe not
as much progress as some of us would like (after all, I would like
100%!) but certainly more than 'zero' progress as well.
|
Actually, I do think the compiler needs to be the most solid piece of
software in the whole product. Without a near bug-free compiler, we'll just
keep on producing software that are going to work most of the time, and fail
some of the time. I'm not only talking about conformance -- it should
include all the VCL extension support as well, since majority of the BCB
code is tightly coupled with the VCL, which is written in Pascal. It is due
to those pieces of code (and the _only_ reason) that we are still using BCB.
Getting all these right should be of utmost importance to Borland. For
example, QC #15800 have plagued nearly every piece of software compiled with
BCB, and it should continue to plague those compiled with Dexter (I believe
so, since the bug remains in "Open" stage as of the time of this post).
Those who have came across the QC report will know how to avoid it, but most
do not even know the bug exists, let alone workaround it.
Personally, I don't care about the new things Borland keeps bringing into
their products like MIDAS (back in BCB 5?), ECO, or any of those new fancy
technologies. I still believe they should get the fundamentals right, before
they try building on it. That brings us back to the question -- will anyone
pay for a bug fix release of, say, Dexter?  And I guess that is what
places Borland in a tough spot.
|
|
| Back to top |
|
|
Gillmer J. Derge [TeamB]
Guest
|
| Posted: Fri Nov 04, 2005 8:19 pm Post subject: Re: Dexter and the C++ standard |
|
|
Edward Diener wrote:
| Quote: | Their reticence to say anything, since after all this
can't be a company secret they want to keep, leads me to the obvious
conclusion.
|
I can't see the sun, which leads me to the obvious conclusion that the
world is ending. You say it's just cloudy outside? Ridiculous! It's
obviously the end of the world.
Consider that Dexter is not a released product that you can buy yet.
They're still working on it. Why is it so surprising that they don't
want to officially commit in a recorded session to exactly what portions
of a library they have no control over can be compiled by a compiler
that isn't finished yet?
--
Gillmer J. Derge [TeamB]
|
|
| Back to top |
|
|
Edward Diener
Guest
|
| Posted: Fri Nov 04, 2005 10:26 pm Post subject: Re: Dexter and the C++ standard |
|
|
Gillmer J. Derge [TeamB] wrote:
| Quote: | Edward Diener wrote:
Their reticence to say anything, since after all this can't be a
company secret they want to keep, leads me to the obvious conclusion.
I can't see the sun, which leads me to the obvious conclusion that the
world is ending.
|
It may be night time.
| Quote: | You say it's just cloudy outside? Ridiculous! It's
obviously the end of the world.
|
But I can look and see if the sun is shining or not. If I was blind and
not able to look and see if the sun is shining, I could ask someone else
and I would get an answer to that question. If someone skirted around
that answer, or refused to give me an answer, it would not be wrong to
assume that it was night and the sun is not shining at the time I asked,
else why would someone not answer a simple question.
| Quote: |
Consider that Dexter is not a released product that you can buy yet.
They're still working on it. Why is it so surprising that they don't
want to officially commit in a recorded session to exactly what portions
of a library they have no control over can be compiled by a compiler
that isn't finished yet?
|
My OP did not ask that or perhaps you can not read. Try reading it
again. It's not so hard.
What a sorry state it is for a company when a simple question regarding
a feature of a soon to be released product gets stonewalled by not only
the company but by its supporters and spokesmen on its NGs.
Nonetheless I have my answer and know I will not be foolish enough to
buy their upcoming product. Good luck to y'all !
|
|
| Back to top |
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
FAQ
Search
Memberlist
Usergroups
Register
Profile
Log in to check your private messages
Log in
