Board index » cppbuilder » Re: Thinking about precompiled headers and fast parallel compiles

Re: Thinking about precompiled headers and fast parallel compiles


2005-08-23 05:29:00 AM
cppbuilder34
Chris Uzdavinis (TeamB) wrote:
Quote
"I haven't heard anything about the preview compiler in a while.
Hmmmm, I wonder why?"
Even the link on the download page disappeared after the homepage update.
--
Regards,
Andreas Hausladen
 
 

Re:Re: Thinking about precompiled headers and fast parallel compiles

Andreas Hausladen wrote:
Quote
I don't know. But bcc32 5.6.4 can't be seen as a state of the art compiler
anymore with it's many bugs (I found at least 5 bugs I had to work around
last weekend while porting Qt 4.0.1 for bcc32).


I'm not suggesting that 5.6.4 is the one that will ship...
 

Re:Re: Thinking about precompiled headers and fast parallel compiles

Alex Bakaev [TeamB] < XXXX@XXXXX.COM >wrote:
Quote
Andreas Hausladen wrote:

>
>The bccx 6.0 preview compiler doesn't have this bug. So there shouldn't be
>any problems. (except that there is no POSIX support [in the Preview
>compiler]).

Are you sure next version of BCB will have that compiler?
Are you really and honestly suggesting
they won't...err...they would even
/consider/ to not to...use an EDG-based
compiler as their new C++ compiler???
Oh boy, then Dinkumware is out, too, I
suspect?
Quote
.a
Schobi
--
XXXX@XXXXX.COM is never read
I'm Schobi at suespammers dot org
"Coming back to where you started is not the same as never leaving"
Terry Pratchett
 

{smallsort}

Re:Re: Thinking about precompiled headers and fast parallel compiles

Chris Uzdavinis (TeamB) < XXXX@XXXXX.COM >wrote:
Quote
[...]

Thinking out loud to myself:

"I haven't heard anything about the preview compiler in a while.
Hmmmm, I wonder why?"
Please, please tell me you're not suggesting
the EDG-based compiler isn't anymore.
Schobi
--
XXXX@XXXXX.COM is never read
I'm Schobi at suespammers dot org
"Coming back to where you started is not the same as never leaving"
Terry Pratchett
 

Re:Re: Thinking about precompiled headers and fast parallel compiles

"Hendrik Schober" < XXXX@XXXXX.COM >wrote:
Quote
Oh boy, then Dinkumware is out, too, I
suspect?
I think in one of their online discussions they indicated that
Dinkumware is in.
- Leo
 

Re:Re: Thinking about precompiled headers and fast parallel compiles

Chris Uzdavinis (TeamB) < XXXX@XXXXX.COM >wrote:
Quote
"I haven't heard anything about the preview compiler in a while.
Hmmmm, I wonder why?"
Could it be that it never included support for VCL extensions? Seems
like the whole VCL thing is what [killed] CBX, and the preview
compiler was to be included in CBX.
- Leo
 

Re:Re: Thinking about precompiled headers and fast parallel compiles

"Hendrik Schober" < XXXX@XXXXX.COM >writes:
Quote
Chris Uzdavinis (TeamB) < XXXX@XXXXX.COM >wrote:
>[...]
>
>Thinking out loud to myself:
>
>"I haven't heard anything about the preview compiler in a while.
>Hmmmm, I wonder why?"


Please, please tell me you're not suggesting
the EDG-based compiler isn't anymore.
Nice weather we're having, isn't it?
--
Chris (TeamB);
 

Re:Re: Thinking about precompiled headers and fast parallel compiles

Leo Siefert < XXXX@XXXXX.COM >wrote:
Quote
Chris Uzdavinis (TeamB) < XXXX@XXXXX.COM >wrote:

>"I haven't heard anything about the preview compiler in a while.
>Hmmmm, I wonder why?"

Could it be that it never included support for VCL extensions? Seems
like the whole VCL thing is what [killed] CBX, and the preview
compiler was to be included in CBX.
Well, wasn't there a joined Borland/EDG
proposal for adding porperties to C++?
Quote
- Leo
Schobi
--
XXXX@XXXXX.COM is never read
I'm Schobi at suespammers dot org
"Coming back to where you started is not the same as never leaving"
Terry Pratchett
 

Re:Re: Thinking about precompiled headers and fast parallel compiles

Chris Uzdavinis (TeamB) < XXXX@XXXXX.COM >wrote:
Quote
[...]
>Please, please tell me you're not suggesting
>the EDG-based compiler isn't anymore.

Nice weather we're having, isn't it?
No. Actually it's cloudy and raining.
(If you were looking out of my windows here,
you'd think I live in PDX or even in the Lake
District.) Too bad. And DW is rained on, too?
Is this something only TeamB'ers are knowing
or has the field test already started?
Oh, the weather, I know...
Schobi
P.S.: You know, I do, as everybody else who is
interested, know pretty well when VC8 is
going to be released, what bugs are going
to be fixed, what will be new in the std
lib etc.
Yes, MS might slip, in which case we'll
have to adjust plans. But we are able to
plan at all. (One thing this gained us is
that I am /not/ spending weeks on trying
to circumvent a particular {*word*193} compiler
bug that prevents important code to get
compiled on Windows. Product management
agreed to defer the Windows version a few
month in exchange for getting something
else done earlier.)
--
XXXX@XXXXX.COM is never read
I'm Schobi at suespammers dot org
"Coming back to where you started is not the same as never leaving"
Terry Pratchett
 

Re:Re: Thinking about precompiled headers and fast parallel compiles

Leo Siefert wrote:
Quote

I think in one of their online discussions they indicated that
Dinkumware is in.

- Leo
According to bdn.borland.com/article/0,1410,32958,00.html
<quote>
jkaster: DavidS: "Will the new C++ Identity be based on the existing
BCB compiler / linker?" - It's had a lot of revisions since then. It's
not a ground up new compiler and linker, but it's had lots of changes
since it was last released.
</quote>
So it is a revision of the current compiler by the sounds of it. And
later on:
<quote>
jkaster: derbas: "Will anything be done to reduce executable size?" -
Outside of including the dinkumware STL, we don't have anything on the
plate right now
</quote>
So Dinkumware is in.
Worth a read if you haven't already.
Cheers
Russell
 

Re:Re: Thinking about precompiled headers and fast parallel compiles

Russell Hind < XXXX@XXXXX.COM >wrote:
Quote
According to bdn.borland.com/article/0,1410,32958,00.html

<quote>
Thanks for tracking that down - I was too lazy.
- Leo
 

Re:Re: Thinking about precompiled headers and fast parallel compiles

Hendrik Schober wrote:
Quote
Well, wasn't there a joined Borland/EDG
proposal for adding porperties to C++?
In a word- No!
Although both Borland and EDG staff have separately proposed properties
<g>
Borland put forward a 'PME' proposal, Properties, Methods and Events, a
few years back. That did not generate much interest:
www.open-std.org/jtc1/sc22/wg21/docs/papers/2002/n1384.pdf
At the Sydney meeting last year, Daveed Vanadevoorde put forward a
different paper for introducing properties into C++ based on work done
for C++/CLI standards. This was largely written on the basis he did
not personally have to believe in or defend the proposal. It was quite
an interesting debate ;?) I think it was Dave Abrahams finally
persuaded Daveed that his paper had merit, for a purpose totally
different than normally used (far away from Object Inspectors and RAD)
- writing richer DSELs.
www.open-std.org/jtc1/sc22/wg21/docs/papers/2004/n1600.html
Lois Goldthwaite got rather wound up in the whole debate, and turned up
next day with a counter-paper, proposing a library based alternative
that she didn't want either - but preferred to hacking the core
language if we were sure we needed properties!
www.open-std.org/jtc1/sc22/wg21/docs/papers/2004/n1615.pdf
I would not say there was any clear consensus on whether properties
were a good thing or a bad thing, the whole spectrum of views was seen
from strongly in favour to strongly against, with probably more against
than for, but most falling into the 'weakly held' view, so may be
persuaded with further work. Not sure anyone is doing that work at the
moment though.
AlisdairM(TeamB)