phroger_rabbit wrote:
> --- In rabbit-semi@rabb..., Scott Henion
wrote:
>
>> Once I saw it generate wrong code for char math, I uninstalled it.
>> Supposedly it is fixed in 9.41.
>>
>> Were you told that there would be a v9.41? My "impression" (I won't
>> quote yet) is that there won't be.
>>
I was told that the char promotion problem is fixed in 9.41. When/If
they actually release it was not mentioned.
------
| Scott G. Henion| shenion@shen... |
| Consultant | Stone Mountain, GA |
| SHDesigns | PGP Key 0xE98DDC48 |
| http://www.shdesigns.org |
------
today's fortune
There are new messages.
Dynamic C 9.40 bugs?
Started by ●March 13, 2006
Reply by ●March 13, 20062006-03-13
Reply by ●March 13, 20062006-03-13
Reply by ●March 13, 20062006-03-13
They are probably pushing to get DC v10 ready for the R4000 release
and skipping any revisions to 9.40. That will mean we will have to
pay to go up to v10.
Scott Henion wrote:
>
> Gary Newsted wrote:
> >> When ZW starts actually changing the compiler, it scares me. I
> >> don't think they have the resources to write and understand
> >> a compiler.
> >>
> >
> > OK, these are scary observations! Does anyone from ZW actually
pay
> > attention to the problems reported in this forum?
> >
> No, they dont. That was confirmed from a call from them recently.
>
> ------
> | Scott G. Henion| shenion@... |
> | Consultant | Stone Mountain, GA |
> | SHDesigns | PGP Key 0xE98DDC48 |
> | http://www.shdesigns.org |
> ------
> today's fortune
> There are new messages.
>
and skipping any revisions to 9.40. That will mean we will have to
pay to go up to v10.
Scott Henion wrote:
>
> Gary Newsted wrote:
> >> When ZW starts actually changing the compiler, it scares me. I
> >> don't think they have the resources to write and understand
> >> a compiler.
> >>
> >
> > OK, these are scary observations! Does anyone from ZW actually
pay
> > attention to the problems reported in this forum?
> >
> No, they dont. That was confirmed from a call from them recently.
>
> ------
> | Scott G. Henion| shenion@... |
> | Consultant | Stone Mountain, GA |
> | SHDesigns | PGP Key 0xE98DDC48 |
> | http://www.shdesigns.org |
> ------
> today's fortune
> There are new messages.
>
Reply by ●March 13, 20062006-03-13
Move to the Softools compiler! I've heard nothing bad about it. The
last ZW compiler I used was 8.61.
I think ZW should get out of the compiler business and just make chips &
boards.
-Pete
last ZW compiler I used was 8.61.
I think ZW should get out of the compiler business and just make chips &
boards.
-Pete
Reply by ●March 13, 20062006-03-13
Mike van Meeteren wrote:
> Interesting you mention 8.01 since that's the compiler I use. What
> specifically does it not compile well that I should look out for? Since
> I'm not spending any money to upgrade this compiler, or spending a ton of
> time porting a rather large group of products to Softools, any benefit to
> switching to 8.61? I noticed the object code it generates is significantly
> larger than 8.01 which is why I haven't switched.
>
The only major one I have seen is the stack issue with bytes near 256 from the top of the stack.
func(int param)
{
char buff[250];
unsigned long,x,y,z;
}
the compiler may try accessing y,x using (IX+nn); nn is an 8-bit #. The offset from IX is greater than 255, so the wrong addresses get read/written to. Structures that span this 256-byte offset may have similar problems.
I believe this was fixed in 8.51.
I was able to work around the problem by reordering auto variables. That improves code speed a bit as it can use 8-bit offsets most of the time. Move arrays last as they need address+index calculations anyway.
I have dealt with compiler issues by copying the compiler and dll's over older versions. That gets the compiler fix without the increased bloat of the newer bios/libs. My 7.32 compiler has the ethernet stack from 8.61 backported (via a compare and edit). So I get the more reliable compiler, smaller bios, but the ethernet bug fixes.
------
| Scott G. Henion| shenion@shen... |
| Consultant | Stone Mountain, GA |
| SHDesigns | PGP Key 0xE98DDC48 |
| http://www.shdesigns.org |
------
today's fortune
Uncle Ed's Rule of Thumb:
Never use your thumb for a rule.
You'll either hit it with a hammer or get a splinter in it.
> Interesting you mention 8.01 since that's the compiler I use. What
> specifically does it not compile well that I should look out for? Since
> I'm not spending any money to upgrade this compiler, or spending a ton of
> time porting a rather large group of products to Softools, any benefit to
> switching to 8.61? I noticed the object code it generates is significantly
> larger than 8.01 which is why I haven't switched.
>
The only major one I have seen is the stack issue with bytes near 256 from the top of the stack.
func(int param)
{
char buff[250];
unsigned long,x,y,z;
}
the compiler may try accessing y,x using (IX+nn); nn is an 8-bit #. The offset from IX is greater than 255, so the wrong addresses get read/written to. Structures that span this 256-byte offset may have similar problems.
I believe this was fixed in 8.51.
I was able to work around the problem by reordering auto variables. That improves code speed a bit as it can use 8-bit offsets most of the time. Move arrays last as they need address+index calculations anyway.
I have dealt with compiler issues by copying the compiler and dll's over older versions. That gets the compiler fix without the increased bloat of the newer bios/libs. My 7.32 compiler has the ethernet stack from 8.61 backported (via a compare and edit). So I get the more reliable compiler, smaller bios, but the ethernet bug fixes.
------
| Scott G. Henion| shenion@shen... |
| Consultant | Stone Mountain, GA |
| SHDesigns | PGP Key 0xE98DDC48 |
| http://www.shdesigns.org |
------
today's fortune
Uncle Ed's Rule of Thumb:
Never use your thumb for a rule.
You'll either hit it with a hammer or get a splinter in it.
Reply by ●March 13, 20062006-03-13
Since November of 2004, I've been using v8.61 exclusively. Almost
every one of our products ships with a unique frimware build. I can
not point to any problem attributable to the compiler or the library.
(The one possible exception is the issue with some Realtek 8019 chips
not reading the MAC address properly after power-up. The library
should, but doesn't, account for the problem.)
--- In rabbit-semi@rabb..., Mike van Meeteren wrote:
>
> At 02:58 PM 3/13/2006 -0500, you wrote:
>
> Interesting you mention 8.01 since that's the compiler I use. What
> specifically does it not compile well that I should look out for?
Since
> I'm not spending any money to upgrade this compiler, or spending a
ton of
> time porting a rather large group of products to Softools, any
benefit to
> switching to 8.61? I noticed the object code it generates is
significantly
> larger than 8.01 which is why I haven't switched.
>
> -Mike
>
every one of our products ships with a unique frimware build. I can
not point to any problem attributable to the compiler or the library.
(The one possible exception is the issue with some Realtek 8019 chips
not reading the MAC address properly after power-up. The library
should, but doesn't, account for the problem.)
--- In rabbit-semi@rabb..., Mike van Meeteren wrote:
>
> At 02:58 PM 3/13/2006 -0500, you wrote:
>
> Interesting you mention 8.01 since that's the compiler I use. What
> specifically does it not compile well that I should look out for?
Since
> I'm not spending any money to upgrade this compiler, or spending a
ton of
> time porting a rather large group of products to Softools, any
benefit to
> switching to 8.61? I noticed the object code it generates is
significantly
> larger than 8.01 which is why I haven't switched.
>
> -Mike
>
Reply by ●March 13, 20062006-03-13
Reply by ●March 13, 20062006-03-13
"They" used to post answers to questions on this group. I can't remember
the last time I saw a response from someone at ZW or RabbitSemi.
Reply by ●March 13, 20062006-03-13
On Tue, 14 Mar 2006 00:34:34 -0000, phroger_rabbit wrote:
>Is it just me noticing it or has their support deteriorated?
>
>Is it because of the Digi buyout?
I've noticed its been heading that way for awhile in my opinion.
I have a hard time believing the DC compiler gets beta tested
at all, with all the bugs I've seen and read about. Not just v9.40
but all of them. Anyone care to admit to beta testing any version
of DC ?
Even the smallest thing like editor option "cursor though tabs" has
been broken or the checkbox is backwards since 8.00 and still is.
Some times I have a need to use the old Z180 DynamicC 32 and it
seems extremely fast at compiling compared to the rabbit compilers.
With programs about the same size, at least thats how it seems to me.
Matt
>Is it just me noticing it or has their support deteriorated?
>
>Is it because of the Digi buyout?
I've noticed its been heading that way for awhile in my opinion.
I have a hard time believing the DC compiler gets beta tested
at all, with all the bugs I've seen and read about. Not just v9.40
but all of them. Anyone care to admit to beta testing any version
of DC ?
Even the smallest thing like editor option "cursor though tabs" has
been broken or the checkbox is backwards since 8.00 and still is.
Some times I have a need to use the old Z180 DynamicC 32 and it
seems extremely fast at compiling compared to the rabbit compilers.
With programs about the same size, at least thats how it seems to me.
Matt
Reply by ●March 13, 20062006-03-13
Gripe, Gripe, Gripe!
Get over it!
This is a technical help site for serious users of the ZWorld/Rabbit Semi products.
If you do not like the products, leave.
I like the products, and I deal with the problems just like every body else does.
I have not had a problem with the technical support, they bend over backwards to assist me when it counts the most.
I have also read just about every thing there is to read about DC, and I have very few problems using it.
Gripe, Gripe, Gripe!
Get over it!
Lets all get back to technical issues.
JIMA
Get over it!
This is a technical help site for serious users of the ZWorld/Rabbit Semi products.
If you do not like the products, leave.
I like the products, and I deal with the problems just like every body else does.
I have not had a problem with the technical support, they bend over backwards to assist me when it counts the most.
I have also read just about every thing there is to read about DC, and I have very few problems using it.
Gripe, Gripe, Gripe!
Get over it!
Lets all get back to technical issues.
JIMA
