stable version

[cabel]

How many years we have to wait for a pcw stable version?
I'm using pcw since many years and every new version has newer stupid bugs.
I've purchased 4.093 recently and it's nasty to see many bugs.

[asmboy]

so would you post a LIST of the many bugs you have found - ?

It would be helpful if an expert such as your self would point out
the specifics of what you are talking about.

I guess i don't write programs with enough complexity - because - i have been very pleased with the lack of bugs .

SO i'd be curious to learn more about the problems you have discovered
so i can be sure they are not important features instead.

[dyeatman]

I would be interested to see the list as well. I just finished two 18F4550 industrial control applications with PCWH and encountered no problems at all.
_________________
Google and Forum Search are some of your best tools!!!!

[FvM]

[asmboy]

Merely having a background awareness of issues that might arise is helpful - hence my perpetual interest in what are reported as "bugs".

I am more inclined to be GRATEFUL to know in advance that there might be a problem with say Structure composition - or modal problems with specific hardware on a PIC family.
I don't hold some bugs against a tool maker who reveals the issues they are struggling with.
and HEY ! BECAUSE it is C for PICS - && the LST is ordinarily helpful and revealing && it is SO easy to introduce a snippet of my own inline ASM -
I have never had a beef with CCS over the various small deficiencies i know of.

My ONLY A** burn Issue is that bug FIXES only get mentioned in the most oblique and USELESSLY non-specific way when a new version makes a dash for freedom. The language that DOES BUG me is " some problem,
with some number formats were fixed." next to the release number.
not confidence inspiring at all.

Working AROUND or THROUGH an issue is pretty much the same to me these days, and time beefing on the forum is time wasted from finding
what the LST dump can tell me.

To Paraphrase Winston Churchill:

"CCS is the WORST compiler there is, except for ALL THE REST!"

[FvM]

Just a newly found (probably old) PCD V4.093 bug from the road:

In the below compare with a conditional expression, the involved numeric
variables are int16, is_f130 has int1 type. PCD is erroneously using byte
moves.

[Franck26]

Hi,
I don't think that CCS compiler is worse than others. Anyway, what I don't like is that CCS doesn't admit their issues and it doesn't look like they try to fix it...

Some examples:
- when I bought the compiler I was hoping using the multi-compilation unit... still didn't work on V4.072.
- I bought the CCS Load and Go 1 year ago, I should have read the comment on the forum before: I was not able to use it for 1 year. With the new update it starts working for some CPU, of course not for mine!!!
- I bought the IDE too, I spend my time clicking on the end button on my key board because when you try to add text at the end of a line, it add thousand of space before or write it in the middle of your line code...
- Ctrl+Z is nearly stable, but don't try Ctrl+Y, you would lost part of your code!!!
- The examples supplied with the compiler have a lot of bugs and they are full of warnings when you compile it (ex13b.c generate more than 30 warnings).

But as said before what is very annoying is that when you send them emails, it looks like those issues are not important to them, they prefer adding new features instead of correcting the old problems...

Anyway, I've never tried other compiler for pic, maybe it's the fact that pic are middle range CPU, so less used by professional (no offence).

C.U.
Franck.

[cabel]

I'm talking about the IDE bugs. For example:
- sometime appears the message "the software is unregistered"
- I assign a shortcut to program chips, but I lost it on the next launch of PCW
- sometime the compile window doesn't disappear
- in the tools->program chip menu there is the command "tìtle2=title2=title2=title2=title2=icd"

[asmboy]

Sorry - I thought you meant the compiler, which at the LOW and MIDDLE end of the spectrum is working - great for me.

But the IDE ? who uses that ?

I tried for a short time, it but it is too clumsy and buggy, and the way it "absorbs" existing code , for no useful , functional reason led me to abandon it - and never look back.

For code creation - I use a self-customized version of Multi Edit that has served me very well over time for managing multiple code windows - and has easy cut paste etc between screens.

[Loyal CCS user]

cabel, don't back down. You'll notice that most of the old timers aren't commenting. That's because they know CCS constantly has bugs and not only that, they break things that worked in previous versions. 80% of the coolest features CCS advertises don't fully work.

addressmod occasionally generates the wrong number of bytes
#import cuts the imported file in half if you specify the correct range
The keyword ROM mostly works but some pointers to rom always return 0xFF
write_configuration_bits recently had problems
PCD is just now starting to be usable and was just recently riddled with I2C problems
The list just goes on and on. Why don't I give more examples? Because this is the part of CCS that I am bored to death with.

Yesterday at about 11PM I discovered that the latest CCSLOAD erases firmware if the HEX includes values for the EEPROM.

I think CCS has/had the PIC compiler world by the tail. But I don't think they make the right decisions to keep the grip. I don't want to add insult to injury but Microchip bought Hi Tech instead of CCS. Why is that.

The unique CCS approach to code is a great one and I will remain a loyal user until they force me to do something else. I hope to be blown away by version 4.100

[PCM programmer]

OK, I'll comment. I avoid the IDE bugs by using MPLAB. I avoid other
problems by not using extra features such as addressmod.

[Franck26]

Hi PCM,

Just a small question:
Does the CCS ICD-U40 debugger works well with the microchip IDE?
With the same functionalities?

Thanks,
Franck.

[PCM programmer]

I use MPLAB with the Microchip ICD2. I don't use ICD-U40.

[Loyal CCS user]

And I respect that.

But why should we have to avoid an extremly powerful concept like addressmod just because CCS won't ring it out. All I ask is that they finish it and leave it alone. After all, they got the hard part right. They just didn't always recognize the correct number of bytes in a variable.

I worked around it and reported it but like everyone is saying; who knows if it got fixed. If I really wanted to know I'd have to drudge up that project again and recompile. Just don't have the time for it. LOL Too busy beta testing for them.