 |
BorlandTalk.com Borland discussion newsgroups
|
| View previous topic :: View next topic |
| Author |
Message |
vortic Guest
|
Posted: Mon Oct 23, 2006 8:28 pm Post subject: Next compiler: CLI capable? |
|
|
In the next BDS or TurboC++ release:
1. Will it be able to compile using CLI language?
2. Please tell us, will it compile at least 85% of current Boost?
Yes, I know they are working on #2, but it would be nice to get some real
info on this. |
|
| Back to top |
|
 |
vortic Guest
|
Posted: Mon Oct 23, 2006 8:30 pm Post subject: Re: Next compiler: CLI capable? |
|
|
"vortic" <vortic (AT) users (DOT) sourceforge.net> wrote in message
| Quote: |
1. Will it be able to compile using CLI language?
|
I think technically I should say CLI as implemented by MS. |
|
| Back to top |
|
 |
David Dean Guest
|
Posted: Mon Oct 23, 2006 8:55 pm Post subject: Re: Next compiler: CLI capable? |
|
|
In article <453cdf92$1 (AT) newsgroups (DOT) borland.com>,
"vortic" <vortic (AT) users (DOT) sourceforge.net> wrote:
| Quote: | 1. Will it be able to compile using CLI language?
|
I haven't heard anything about that. It doesn't appear to be on the
roadmap, nor are there any QC entries.
| Quote: | 2. Please tell us, will it compile at least 85% of current Boost?
|
What is the best way to find out the percentage? I've been relying on
Nicola and Alisdair's work up until now. I should probably be more
active in finding out how to do it for myself. When Highlander is
released, I'd love to put it through it's paces and find out. (we could
also put in QC entries for the specific libraries where it fails)
--
-David
Quis custodiet custodes ipsos? |
|
| Back to top |
|
 |
mr_organic Guest
|
Posted: Mon Oct 23, 2006 9:13 pm Post subject: Re: Next compiler: CLI capable? |
|
|
"vortic" <vortic (AT) users (DOT) sourceforge.net> wrote in
news:453cdf92$1 (AT) newsgroups (DOT) borland.com:
| Quote: | In the next BDS or TurboC++ release:
1. Will it be able to compile using CLI language?
2. Please tell us, will it compile at least 85% of current Boost?
Yes, I know they are working on #2, but it would be nice to get some
real info on this.
|
1. I doubt it very seriously, unless DTG/DevCo uses Microsoft's own C++
compiler as a plug-in. AFAIK, only Microsoft's C++ compiler can so far
target the .NET CLR. I don't see DTG/DevCo doing this for a long time --
the toolchain has other problems that need attention first (see next
point).
2. I know that Boost is a high-profile area for improvement for the C++
compiler at DTG/DevCo, but I'm not sure that DevCo has any short-term
plans to do anything about it. The low-hanging fruit is gone -- the
compiler needs extensive low-level work in all areas: speed, compliance,
code-generation, etc. (not to mention multi-processor support). And I'm
not even getting into issues with the linker or debugger.
I have a feeling that in order to address all the issues that C++
developers are complaining about, DTG/DevCo is going to have to so
extensively re-work the compiler that it almost counts as a totally new
effort. And since a compiler toolchain IME takes about three to five
years to really mature, I wouldn't hold my breath waiting for a new C++
toolchain out of DTG/DevCo any time soon.
(FWIW, I re-iterate my belief that DTG/DevCo ought to simply fork GCC and
graft on VCL extensions so it can compile the VCL. It would save them a
ton of work in the long run.)
Regards,
mr_organic |
|
| Back to top |
|
 |
Pete Fraser Guest
|
Posted: Mon Oct 23, 2006 10:19 pm Post subject: Re: Next compiler: CLI capable? |
|
|
It's down on the RoadMap as (wrongly) Managed C++ and is due with Longhorn
support in 2007.
http://bdn.borland.com/article/0,1410,33383,00.html
--
Rgds Pete
=========================================
GenHelp, the Component Author's help writer
www.frasersoft.net/genhelp
pete.fraser (AT) frasersoft (DOT) net
"David Dean" <ozchzhq02 (AT) sneakemail (DOT) com> wrote in message
news:ozchzhq02-0875F6.11555023102006 (AT) frylock (DOT) local...
| Quote: | In article <453cdf92$1 (AT) newsgroups (DOT) borland.com>,
"vortic" <vortic (AT) users (DOT) sourceforge.net> wrote:
1. Will it be able to compile using CLI language?
I haven't heard anything about that. It doesn't appear to be on the
roadmap, nor are there any QC entries. |
|
|
| Back to top |
|
 |
Rudy Velthuis [TeamB] Guest
|
Posted: Mon Oct 23, 2006 10:42 pm Post subject: Re: Next compiler: CLI capable? |
|
|
At 17:30:40, 23.10.2006, vortic wrote:
| Quote: |
"vortic" <vortic (AT) users (DOT) sourceforge.net> wrote in message
1. Will it be able to compile using CLI language?
I think technically I should say CLI as implemented by MS.
|
If you mean C++/CLI, that is supposed to become a standard of its own. Of
course MS has the advanatge that they have the compiler and the framework
already. <g>
--
Rudy Velthuis [TeamB] http://rvelthuis.de/
"If FORTRAN has been called an infantile disorder, then PL/I must
be classified as a fatal disease." -- Edsger Dijkstra |
|
| Back to top |
|
 |
vortic Guest
|
Posted: Mon Oct 23, 2006 11:00 pm Post subject: Re: Next compiler: CLI capable? |
|
|
"Pete Fraser" <pete.fraser (AT) frasersoft (DOT) nospam.com> wrote in message
Excellent, Highlander will be released THIS YEAR, and Delphi Vista will be
released in 2007, along with a new Compiler for C++ folks that is capable of
C++/CLI, so we will have C++.NET in the year 2007.
Did you see that everybody, DevCo is releasing C++.NET compiler in the year
2007! |
|
| Back to top |
|
 |
David Dean Guest
|
Posted: Tue Oct 24, 2006 12:08 am Post subject: Re: Next compiler: CLI capable? |
|
|
In article <453cf996$1 (AT) newsgroups (DOT) borland.com>,
"Pete Fraser" <pete.fraser (AT) frasersoft (DOT) nospam.com> wrote:
Thanks Pete. I thought it was there, but then I couldn't find it.
--
-David
Quis custodiet custodes ipsos? |
|
| Back to top |
|
 |
Alisdair Meredith[TeamB] Guest
|
Posted: Tue Oct 24, 2006 4:33 am Post subject: Re: Next compiler: CLI capable? |
|
|
mr_organic wrote:
| Quote: | 1. I doubt it very seriously, unless DTG/DevCo uses Microsoft's own
C++ compiler as a plug-in. AFAIK, only Microsoft's C++ compiler can
so far target the .NET CLR. I don't see DTG/DevCo doing this for a
long time -- the toolchain has other problems that need attention
first (see next point).
|
Note that there were other folks beyond Microsoft also involved in the
C++/CLI standard, who might also be interested in supplying a front
end. Also note that the MS implementation, while the only one that is
currently available, is not yet 100% conforming! (And I don't just mean
because C++/CLI implies you must implement export!)
That said, licensing the MS compiler would be a great starting point,
although I personally have no idea what is being planned inside Borland.
| Quote: | 2. I know that Boost is a high-profile area for improvement for the
C++ compiler at DTG/DevCo, but I'm not sure that DevCo has any
short-term plans to do anything about it. The low-hanging fruit is
gone -- the compiler needs extensive low-level work in all areas:
speed, compliance, code-generation, etc. (not to mention
multi-processor support). And I'm not even getting into issues with
the linker or debugger.
|
I hope to have more to report here after Boost 1.34 is released ...
| Quote: | (FWIW, I re-iterate my belief that DTG/DevCo ought to simply fork GCC
and graft on VCL extensions so it can compile the VCL. It would save
them a ton of work in the long run.)
|
That was effectively tried with the licensed EDG front end for
C++BuilderX. I think the idea of 'simply grafting on the Delphi
extensions' goes deeper than you suspect. Note that they need to be
ABI compatible with Delphi, as well as simply following the semantic
rules. This may contradict the established GCC C++ ABI for Windows.
I still think there is a lot of value in pursuing a GCC based product
as-well-as BDS, but it would be an entirely different product -
essentially a matured CBuilderX. I don't see much activity in this
direction from Borland either, while focussed on giving the best
CBuilder support they can (such as improving language conformance and
passing those darned Boost tests!)
--
AlisdairM(TeamB) |
|
| Back to top |
|
 |
Randall Parker Guest
|
Posted: Tue Oct 24, 2006 6:24 am Post subject: Re: Next compiler: CLI capable? |
|
|
Alisdair Meredith[TeamB] wrote:
| Quote: | I hope to have more to report here after Boost 1.34 is released ...
|
Alisdair,
Wasn't 1.34 supposed to be released months ago? I got the impression back in April
that it was supposed to be very close to release. No?
Also, has Boost compatibility with the latest Borland compiler improved much in
recent months? |
|
| Back to top |
|
 |
Alisdair Meredith[TeamB] Guest
|
Posted: Tue Oct 24, 2006 8:06 am Post subject: Re: Next compiler: CLI capable? |
|
|
Randall Parker wrote:
| Quote: | Wasn't 1.34 supposed to be released months ago? I got the impression
back in April that it was supposed to be very close to release. No?
|
ERm ....
suffice to say we are having a few teething problems finally getting
this one out the door. The number of open issues dropped below 50
while I was away the last two weeks though, so it looks like we could
be getting close to shipping!
Note: the issue list must drop to zero before we can release, although
some of those issue will be resolved by simply marking the test
'expected fail' for a given compiler.
| Quote: | Also, has Boost compatibility with the latest Borland compiler
improved much in recent months?
|
I am hoping to have more news on the latest compiler shortly after 1.34
goes live...
--
AlisdairM(TeamB) |
|
| Back to top |
|
 |
Andre Kaufmann Guest
|
Posted: Tue Oct 24, 2006 8:10 am Post subject: Re: Next compiler: CLI capable? |
|
|
Alisdair Meredith[TeamB] wrote:
| Quote: | mr_organic wrote:
[...]
Also note that the MS implementation, while the only one that is
currently available, is not yet 100% conforming! (And I don't just mean
because C++/CLI implies you must implement export!)
|
Does it really imply to implement export (you mean template export) ? Why ?
By the way template export doesn't IMHO add the value it's expected to
add, true code separation and generally faster build times.
[I would be a great fan of template export if it would - better said
could do this]
| Quote: | I still think there is a lot of value in pursuing a GCC based product
as-well-as BDS, but it would be an entirely different product -
essentially a matured CBuilderX.
|
What value would a GCC based add ? And wouldn't that imply that Borland
would have to release an open source compiler ? A C++ compiler which
doesn't support VCL would IMHO lead to the same fiasco like CBuilderX.
Not enough developers will buy it, because it doesn't support VCL.
Andre |
|
| Back to top |
|
 |
Nicola Musatti Guest
|
Posted: Tue Oct 24, 2006 11:53 pm Post subject: Re: Next compiler: CLI capable? |
|
|
vortic wrote:
[...]
| Quote: | Excellent, Highlander will be released THIS YEAR, and Delphi Vista will be
released in 2007, along with a new Compiler for C++ folks that is capable of
C++/CLI, so we will have C++.NET in the year 2007.
|
Didn't you hear? They had to postpone Highlander till next year because
they found that the name bcb2006 was already taken :-P
Cheers,
Nicola Musatti
--
Nicola Musatti
Team Azzurro
http://nicola.musatti.googlepages.com/home |
|
| Back to top |
|
 |
Nicola Musatti Guest
|
Posted: Wed Oct 25, 2006 12:03 am Post subject: Re: Next compiler: CLI capable? |
|
|
Alisdair Meredith[TeamB] wrote:
| Quote: | mr_organic wrote:
1. I doubt it very seriously, unless DTG/DevCo uses Microsoft's own
C++ compiler as a plug-in. AFAIK, only Microsoft's C++ compiler can
so far target the .NET CLR. I don't see DTG/DevCo doing this for a
long time -- the toolchain has other problems that need attention
first (see next point).
Note that there were other folks beyond Microsoft also involved in the
C++/CLI standard, who might also be interested in supplying a front
end. Also note that the MS implementation, while the only one that is
currently available, is not yet 100% conforming! (And I don't just mean
because C++/CLI implies you must implement export!)
That said, licensing the MS compiler would be a great starting point,
although I personally have no idea what is being planned inside Borland.
|
I'd much prefer that they licensed those other folks' front-end...
Possibly to concentrate on the in-house backend.
| Quote: | 2. I know that Boost is a high-profile area for improvement for the
C++ compiler at DTG/DevCo, but I'm not sure that DevCo has any
short-term plans to do anything about it. The low-hanging fruit is
gone -- the compiler needs extensive low-level work in all areas:
speed, compliance, code-generation, etc. (not to mention
multi-processor support). And I'm not even getting into issues with
the linker or debugger.
I hope to have more to report here after Boost 1.34 is released ...
|
At this point I wonder which will come out first...
| Quote: | (FWIW, I re-iterate my belief that DTG/DevCo ought to simply fork GCC
and graft on VCL extensions so it can compile the VCL. It would save
them a ton of work in the long run.)
That was effectively tried with the licensed EDG front end for
C++BuilderX. I think the idea of 'simply grafting on the Delphi
extensions' goes deeper than you suspect. Note that they need to be
ABI compatible with Delphi, as well as simply following the semantic
rules. This may contradict the established GCC C++ ABI for Windows.
|
I don't think investing in a new technology for win32 is a good idea,
especially as that entails porting the VCL to a new compiler.
On the other hand starting from scratch with a C++/CLI solution makes
much more sense. I guess the real issue is going to be what it's going
to take to make the VCL.NET accessible/usable from C++/CLI.
| Quote: | I still think there is a lot of value in pursuing a GCC based product
as-well-as BDS, but it would be an entirely different product -
essentially a matured CBuilderX. I don't see much activity in this
direction from Borland either, while focussed on giving the best
CBuilder support they can (such as improving language conformance and
passing those darned Boost tests!)
|
That would be interesting especially if they come out with a
multiplatform solution that's not as sluggish as Eclipse. On the other
hand when they launched into CBX JBuilder was all the rage; two years
later it had to join 'em so as not to be licked.
Cheers,
Nicola Musatti
--
Nicola Musatti
Team Azzurro
http://nicola.musatti.googlepages.com/home |
|
| Back to top |
|
 |
Peter Agricola Guest
|
Posted: Wed Oct 25, 2006 8:10 am Post subject: Re: Next compiler: CLI capable? |
|
|
"Nicola Musatti" wrote:
| Quote: | That would be interesting especially if they come out with a multiplatform
solution that's not as sluggish as Eclipse.
|
You should really try
http://www.codeblocks.org/
It's a _native_ CBuilderX and is said to be modelled after Visual Studio. (I
can't compare because I don't use Visual Studio, not even the free one's.
When I removed C#Builder trail including the .NET framework I have never
installed a program again that needs the .Net framework.)
The IDE itself is built using wxWidgets, but there is no need to use that
library yourself. There is a Dialog editor, wxSmith for wx applications but
that one I haven't tried yet. Borland compiler is supported out of the box.
It was auto detected on my machine.
Be sure to pick the nightly build. Version 1 rc2 is depricated.
Peter |
|
| 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
|
|