Board index » cppbuilder » Re: BorCon 2004: A New Hope for C++Builder?

Re: BorCon 2004: A New Hope for C++Builder?


2004-09-25 10:22:00 AM
cppbuilder92
Randall Parker < XXXX@XXXXX.COM >wrote:
Quote
Leroy,

Yes, I hear generally bad things about CodeWarrior. This is why I keep hoping that
Borland will make BCBX support other processor targets such as ARM, MicroChip
MicroPIC, TI DSPs, and Moto MPC 555. Borland then could offer a unified IDE and
de{*word*81} front-end across different processor targets. That'd be nice.
Doesn't GCC support some of these?
 
 

Re:Re: BorCon 2004: A New Hope for C++Builder?

Hear, absolutely *&!$% hear. Duane and Oscar have summed up exactly what is
wrong with CBX and Borland's fruitless pursuit of trying to make it work.
Borland seems to be based on "pie in the sky" promises currently, in terms of
any product apart from Delphi! Perhaps they'll have something ready by the time
Longhorn and Novell/Mono are fully fledged, but it will be for XP and Wine
instead! Jeez.
Why didn't the CBX support person have a copy running on his machine, ready to
verify whether the BOE bugs had been fixed, there and then on the phone?!?
That's like a tech support person saying, "Go and find out for yourself."
<SEETHE>
I hate this *&!$% hierarchical consumer society we live in. Let's base things on
technical excellence instead of financial gain. Oh, sorry, that would mean a
complete turnround to the status quo, and the satanic powers that be can't have
that, can they?
</SEETHE>
--
Mark Jacobs
"Duane Hebert" < XXXX@XXXXX.COM >wrote in message
|
| "Oscar Fuentes" < XXXX@XXXXX.COM >wrote in message
|>IMO, CBX only appeals to those who didn't bother to learn how to use
|>the tools at hand. Some marketers thinks
|>there is a significant profit waiting on the group of programmers who
|>suffer from "bad laziness", that is, the ones who prefer to save work
|>today and sacrifice tomorrow's productivity instead of working hard
|>now for being effective on the future (that's good laziness).
|
| Well I certainly agree given what they seem to be offering. To me
| it just seems like an expensive way to create make files.
|
|>(Personally, I prefer hard-to-use de{*word*81}s. I've observed
|>that those encourage the production of better code, making the
|>de{*word*81} a rarely needed tool and thus less relevant. A hard to use
|>de{*word*81} is a productivity boost :-)
|
|>Support is a word incompatible with Borland. Just after leaving
|>Borland I realized how bad it is. MinGW, which is a project maintained
|>by a bunch of people on their free time, is much more well supported
|>and "customer friendly" than Borland. GCC releases improved versions
|>on a steady basis. Intel does the same. You talk to a human when you
|>deal with GCC or Intel, which most of the time addresses your problem
|>and fixes it. As a former Borland customer, this amazed me to no end
|>for a long time, as it was a new experience :-)
|
| I have to agree with this as well. For Borland, most support has
| consisted of "Buy the next version and hope that the bugs you
| need fixed are fixed and the new ones won't be so bad" I've never
| heard of a software company that doesn't even list the fixes/improvements
| in a new release. I asked in the CBX group whether 1.5 fixed any of
| the BOE bugs. A Borland employee asked me to download the
| trial edition and let him know! I don't blame him - he was trying
| to be helpful but how is it that even within Borland they didn't know
| that none of the bugs were fixed?
|>>that may be a reason but if the IDE is just a shell that allows
|>>me to use my own tools I may as well write the macros in
|>>Emacs.
|
|>I would require this too: several years of demonstrated commitment to
|>quality and good support.
|>
|>Even if Borland releases a CBX that can compete with my tools right
|>now on terms of productivity, I will not consider it.
|
| Unless I change companies, my current management
| is not going to even consider another Borland product. When CBX
| was announced, we discussed it as we needed to do something
| for the cross platform market. It seemed that CBX would take
| less time converting (since it was a Borland product) so we
| waited a while. If we had not decided to believe what was
| said about CBX, we'd be MUCH better off now. As it is
| we're fairly pressed to meet our release schedule.
|
| I have a lot of respect for the Borland developers. It seems that
| they're just as frustrated with management decisions (we can
| all relate to that at some point). But it's Borland as a company
| that generates their sales. We're either left with the impression
| that they don't know what they want to do or that they are intentionally
| misleading us with promises of open letters and future announcements.
|
| Some people have actually sounded upbeat because at BorCon they
| mentioned that they would tell us something in December! Then we
| hear that they need an ROI of 10x to keep a product going. I think
| may of us went with Borland initially because we didn't like
| the corporate practices of MS. At this point, the only difference
| that I see between Borland and MS is that MSVC is cheaper than
| cbx, works better, is supported, a free beta 2005 release is available
| and they seem to announce their future plans and more or less follow
| up on them. They have even suggested that we send them letters telling
| them how important our customers are. If Borland doesn't want to support me,
then
| my customers won't even notice when I change vendors. They could
| care less if I use Borland or MS or whoever. And yes, some are DOD
| contracts, aerospace companies, auto manufacturers etc. I think that
| a good company would support personal developers to the same extent
| as large software groups. Any time I contact Qt with support questions
| or upgrade requests, they don't ask me if I have high profile customers.
| I find that crass and unprofessional that someone would even mention
| that.
 

Re:Re: BorCon 2004: A New Hope for C++Builder?

"Randall Parker" < XXXX@XXXXX.COM >wrote in message
| I repeat myself: There is a huge world of programming beyond portable GUI
stuff
| across operating systems. Borland could address it. You just go on about
Windows. CBX
| could address a whole different set of programmers who never use Windows and
who ever
| write an X Windows-based GUI program on the *nix operating systems.
Please read what hands-on experience of CBX from Duane and Oscar has yielded.
CBX really is waste of time.
"Duane Hebert" < XXXX@XXXXX.COM >wrote in message
|
| "Oscar Fuentes" < XXXX@XXXXX.COM >wrote in message
|>IMO, CBX only appeals to those who didn't bother to learn how to use
|>the tools at hand. Some marketers thinks
|>there is a significant profit waiting on the group of programmers who
|>suffer from "bad laziness", that is, the ones who prefer to save work
|>today and sacrifice tomorrow's productivity instead of working hard
|>now for being effective on the future (that's good laziness).
|
| Well I certainly agree given what they seem to be offering. To me
| it just seems like an expensive way to create make files.
|
|>(Personally, I prefer hard-to-use de{*word*81}s. I've observed
|>that those encourage the production of better code, making the
|>de{*word*81} a rarely needed tool and thus less relevant. A hard to use
|>de{*word*81} is a productivity boost :-)
|
|>Support is a word incompatible with Borland. Just after leaving
|>Borland I realized how bad it is. MinGW, which is a project maintained
|>by a bunch of people on their free time, is much more well supported
|>and "customer friendly" than Borland. GCC releases improved versions
|>on a steady basis. Intel does the same. You talk to a human when you
|>deal with GCC or Intel, which most of the time addresses your problem
|>and fixes it. As a former Borland customer, this amazed me to no end
|>for a long time, as it was a new experience :-)
|
| I have to agree with this as well. For Borland, most support has
| consisted of "Buy the next version and hope that the bugs you
| need fixed are fixed and the new ones won't be so bad" I've never
| heard of a software company that doesn't even list the fixes/improvements
| in a new release. I asked in the CBX group whether 1.5 fixed any of
| the BOE bugs. A Borland employee asked me to download the
| trial edition and let him know! I don't blame him - he was trying
| to be helpful but how is it that even within Borland they didn't know
| that none of the bugs were fixed?
|>>that may be a reason but if the IDE is just a shell that allows
|>>me to use my own tools I may as well write the macros in
|>>Emacs.
|
|>I would require this too: several years of demonstrated commitment to
|>quality and good support.
|>
|>Even if Borland releases a CBX that can compete with my tools right
|>now on terms of productivity, I will not consider it.
|
| Unless I change companies, my current management
| is not going to even consider another Borland product. When CBX
| was announced, we discussed it as we needed to do something
| for the cross platform market. It seemed that CBX would take
| less time converting (since it was a Borland product) so we
| waited a while. If we had not decided to believe what was
| said about CBX, we'd be MUCH better off now. As it is
| we're fairly pressed to meet our release schedule.
|
| I have a lot of respect for the Borland developers. It seems that
| they're just as frustrated with management decisions (we can
| all relate to that at some point). But it's Borland as a company
| that generates their sales. We're either left with the impression
| that they don't know what they want to do or that they are intentionally
| misleading us with promises of open letters and future announcements.
|
| Some people have actually sounded upbeat because at BorCon they
| mentioned that they would tell us something in December! Then we
| hear that they need an ROI of 10x to keep a product going. I think
| may of us went with Borland initially because we didn't like
| the corporate practices of MS. At this point, the only difference
| that I see between Borland and MS is that MSVC is cheaper than
| cbx, works better, is supported, a free beta 2005 release is available
| and they seem to announce their future plans and more or less follow
| up on them. They have even suggested that we send them letters telling
| them how important our customers are. If Borland doesn't want to support me,
then
| my customers won't even notice when I change vendors. They could
| care less if I use Borland or MS or whoever. And yes, some are DOD
| contracts, aerospace companies, auto manufacturers etc. I think that
| a good company would support personal developers to the same extent
| as large software groups. Any time I contact Qt with support questions
| or upgrade requests, they don't ask me if I have high profile customers.
| I find that crass and unprofessional that someone would even mention
| that.
 

{smallsort}

Re:Re: BorCon 2004: A New Hope for C++Builder?

"Mark Jacobs" < XXXX@XXXXX.COM >wrote in message news: XXXX@XXXXX.COM ...
Quote
Why didn't the CBX support person have a copy running on his machine, ready to
verify whether the BOE bugs had been fixed, there and then on the phone?!?
That's like a tech support person saying, "Go and find out for yourself."
Just to clarify, there are no "tech support" persons available
on these groups AFAIK except for the TeamB guys. I'm
not sure the position of that particular guy, but I got the
impression that he was a developer. This was not a
phone call but a posting in a newsgroup. I've never
actually called Borland for support except intially with
some problems with licensing and registration. I think it's
been a long time since Borland offered free tech support
though I may be wrong...
Quote
<SEETHE>
I hate this *&!$% hierarchical consumer society we live in. Let's base things on
technical excellence instead of financial gain. Oh, sorry, that would mean a
complete turnround to the status quo, and the satanic powers that be can't have
that, can they?

</SEETHE>
Yep. I think that we stayed with Borland for so long because we remembered
that at some point in some long forgotten history they were more concerned
with technical excellence and small developer groups. Now it seems that
if I don't have a couple of contracts with Lockheed, they don't want to
support me. I think that's the bit that set me off. It's like they don't mind
dumping their customers but if they in turn have some high profile customers,
maybe it would give Borland a bad name and affect their stock so maybe
they should spend a few minutes pacifying their customers. But then again,
maybe not. It will depend on whether the bean counters decide that the
extra costs are valid. At no point do they seem to consider that they
should fix BCB because it's broken. I really feel sorry for the guys that
bought CBX enterprise.
 

Re:Re: BorCon 2004: A New Hope for C++Builder?

Please direct your browser at info.borland.com/newsgroups/ and read
the newsgroup guidelines. One of them asks us not to quote entire posts we
are following up to; instead, please trim the quotes to the parts relevant
for your reply. Thanks!
 

Re:Re: BorCon 2004: A New Hope for C++Builder?

Actually, this is exactly what I have done. Not much got trimmed, though,
because it was nearly all mostly relevant to the CBX anti-argument.
--
Mark Jacobs
"Thomas Maeder [TeamB]" < XXXX@XXXXX.COM >wrote in message
|
| Please direct your browser at info.borland.com/newsgroups/ and read
| the newsgroup guidelines. One of them asks us not to quote entire posts we
| are following up to; instead, please trim the quotes to the parts relevant
| for your reply. Thanks!
 

Re:Re: BorCon 2004: A New Hope for C++Builder?

Leroy,
I am told that that problem is that the BCBX de{*word*81} front-end with other processors
doesn't know how to deal with different byte sexes, different register sets, and the
like. It needs more smarts for dealing with different processor architectures to
better be able to walk the stack, show locals, and similar things.
Leroy Casterline wrote:
Quote
Randall Parker < XXXX@XXXXX.COM >wrote:
>Leroy,
>
>Yes, I hear generally bad things about CodeWarrior. This is why I keep hoping that
>Borland will make BCBX support other processor targets such as ARM, MicroChip
>MicroPIC, TI DSPs, and Moto MPC 555. Borland then could offer a unified IDE and
>de{*word*81} front-end across different processor targets. That'd be nice.


Doesn't GCC support some of these?
 

Re:Re: BorCon 2004: A New Hope for C++Builder?

Randall Parker < XXXX@XXXXX.COM >wrote:
Quote
Leroy,

I am told that that problem is that the BCBX de{*word*81} front-end with other processors
doesn't know how to deal with different byte sexes, different register sets, and the
like. It needs more smarts for dealing with different processor architectures to
better be able to walk the stack, show locals, and similar things.
This isn't an area with which I'm familiar, but if CBX interfaces with
GDB wouldn't that take care of it? Or perhaps GDB doesn't support these
processors...or perhaps I'm not familiar enough with the issues<g>.
 

Re:Re: BorCon 2004: A New Hope for C++Builder?

If one is running on processor X and talks to processor Y where a de{*word*81} back-end
is running and the de{*word*81} back-end sends bytes over you in response to an examine
request then on the client side the app has to know what the byte sex is for those bytes.
Also, if that other processor does not use the same floating point format or integer
format (e.g. a DSP that has 48 bit variables) then, again, you have to know on the
client side what data representations the target side is using.
Keep in mind that CBX has its own GUI front-end into back-end de{*word*81}s. A fair
amount of the formatting of the data and the logic for how the data is represented is
coded for on the front-end side. The de{*word*81} back-end expects the front-end to know
a fair amount about the target architecture.
Leroy Casterline wrote:
Quote
This isn't an area with which I'm familiar, but if CBX interfaces with
GDB wouldn't that take care of it? Or perhaps GDB doesn't support these
processors...or perhaps I'm not familiar enough with the issues<g>.
 

Re:Re: BorCon 2004: A New Hope for C++Builder?

Tamas Demjen wrote:
Quote
CodeGuard is usually right (read it: practically always)
The only time we've known it be wrong is occasionally with strcpy().
Sometimes for no obvious reason it starts giving out spurious errors.
You can tell they are wrong because the range information is blatantly
valid. Thankfully it doesn't do it often and we usually find that a
complete ground-up (deleting OBJs /and/ EXE/DLL) build fixes it.
--
Andrue Cope [TeamB]
[Bicester, Uk]
info.borland.com/newsgroups/guide.html
 

Re:Re: BorCon 2004: A New Hope for C++Builder?

"Andrue Cope [TeamB]" < XXXX@XXXXX.COM >wrote in message news:4157ca92$ XXXX@XXXXX.COM ...
Quote
The only time we've known it be wrong is occasionally with strcpy().
Sometimes for no obvious reason it starts giving out spurious errors.
You can tell they are wrong because the range information is blatantly
valid. Thankfully it doesn't do it often and we usually find that a
complete ground-up (deleting OBJs /and/ EXE/DLL) build fixes it.
Haven't seen that one much but I do often get errors like
std::numeric_limits::max() returns std::numeric_limits::min()
when codeguard is enabled. I'm not sure that it's CG itself
though as I have to enable dynamic rtl (to prevent resource
from a different rtl errors) and "use debug
libraries" (to catch leaks using VCL code) for CG.
These are not normally enabled in
my projects. I found memory sleuth less of a problem
though CG does bounds checking much better.
 

Re:Re: BorCon 2004: A New Hope for C++Builder?

Greg Comeau < XXXX@XXXXX.COM >wrote:
Quote
In article < XXXX@XXXXX.COM >,
Randall Parker < XXXX@XXXXX.COM >wrote:
>If you use Comeau with one of those back-ends then can their
>de{*word*81}s debug code which had a first compile pass thru Comeau?

See bullet #8 of
www.comeaucomputing.com/4.3.0/minor/win95+/como433.html
"Resolution of template id length for use with Borland,
which has a 250 character id limit"
???
Schobi
--
XXXX@XXXXX.COM is never read
I'm Schobi at suespammers dot org
"Sometimes compilers are so much more reasonable than people."
Scott Meyers
 

Re:Re: BorCon 2004: A New Hope for C++Builder?

"Hendrik Schober" < XXXX@XXXXX.COM >writes:
Quote
Greg Comeau < XXXX@XXXXX.COM >wrote:
>In article < XXXX@XXXXX.COM >,
>Randall Parker < XXXX@XXXXX.COM >wrote:
>>If you use Comeau with one of those back-ends then can their
>>de{*word*81}s debug code which had a first compile pass thru Comeau?
>
>See bullet #8 of
>www.comeaucomputing.com/4.3.0/minor/win95+/como433.html


"Resolution of template id length for use with Borland,
which has a 250 character id limit"

???
As Comeau C++ generates C code, and its templates can get deep, the
internal (mangled) names generated can get quite long. So long that
it {*word*88}s up Borland's compiler, which has relatively small limits.
Previously, Comeau would generate long names that would get
truncated. The result was that multiple different variables, whose
names differed only at the "tail" of their name, appeared the same to
the C compiler because the part that made the names unique were
truncated.
I think Greg means by this bullet that he changed the way he generates
function names such that the're not so long as to reach this
limitation of the compiler. Most likely, it's some kind of hashing
algorithm, so that mangled names are probably just ugly encodings or
numbers or something like that.
I've been able to step through code with executables produced by como,
however it doesn't fit in too well for setting watches or inspecting
variables, because the de{*word*81} is seeing the mangled names, but you
don't, and the IDE/de{*word*81} interface doesn't know how to convert
between the names in your code and the mangled names in the C code
that it's really debugging.
--
Chris (TeamB);
 

Re:Re: BorCon 2004: A New Hope for C++Builder?

Hendrik Schober wrote:
Quote
Greg Comeau < XXXX@XXXXX.COM >wrote:

>In article < XXXX@XXXXX.COM >,
>Randall Parker < XXXX@XXXXX.COM >wrote:
>
>>If you use Comeau with one of those back-ends then can their
>>de{*word*81}s debug code which had a first compile pass thru Comeau?
>
>See bullet #8 of
>www.comeaucomputing.com/4.3.0/minor/win95+/como433.html



"Resolution of template id length for use with Borland,
which has a 250 character id limit"

???

Schobi

He meant to say the item with the #8 in front - down further. The eighth
bullet is not correct.
Jim Dodd
Onset Computer Corp.
 

Re:Re: BorCon 2004: A New Hope for C++Builder?

Randall Parker < XXXX@XXXXX.COM >wrote:
Quote
If one is running on processor X and talks to processor Y where a de{*word*81} back-end
is running and the de{*word*81} back-end sends bytes over you in response to an examine
request then on the client side the app has to know what the byte sex is for those bytes.

Also, if that other processor does not use the same floating point format or integer
format (e.g. a DSP that has 48 bit variables) then, again, you have to know on the
client side what data representations the target side is using.

Keep in mind that CBX has its own GUI front-end into back-end de{*word*81}s. A fair
amount of the formatting of the data and the logic for how the data is represented is
coded for on the front-end side. The de{*word*81} back-end expects the front-end to know
a fair amount about the target architecture.
Thanks for the explanation, that makes sense.