CCS C 5.xxx

[pickit2]

Hi, I cannot see the attachemnt.. could you please post the link?

#94 has link for 5.037
#90 has link for KG
There has been no posts for version above 5.037 

[duende_azul]

Version 5.040..  new bugs and more fun  

[CHJAY]

A new release of the CCS C compiler for PICmicro(r) microcontrollers
has been released. The new version is: 5.041

Recent changes include:
5.041  An access violation for some users starting the PIC24 Wizard is fixed
5.041  A problem with %w where if the digit after the . was 0 it did not output on the PIC24 is fixed
5.041  Syntax errors that should have appeared for poorly formed bit_test() calls now appear
5.041  A bug is fixed that occurred with packed structures inside unpacked structures accessed with ->
5.041  Some issues with bit arrays inside non-RAM structures is fixed
5.041  An error for two RS232 ports without stream names is removed


Updated:  PIC12F1612
Updated:  PIC16F1613
New:      PIC16F1614
New:      PIC16F1615
New:      PIC16F1618
New:      PIC16F1619

[voltas5]

I would like to add to that list:

5.042  An error 103 on Win7/8 computers with full access control enabled is fixed

New Version currently: 5.042

[CHJAY]

Wow, CCS released a new version in a few days...

http://www.ccsinfo.com/devices.php?page=versioninfo

[nagkiller]

lol

The version I was using is 5.012 !!!

Many updates!!!   

[Old_but_Alive]

I guess the only positive is that all software has bugs, and at least CCS are quick to issues updates.
I see lots of complaints on the Mikro site criticising  no updates for new PICs etc etc.

CCS should do more testing though

[gan_canny]

Rapid versioning often is the best of both worlds. The user can save off a trusted version at any time but if the user suspects a defect in the trusted version there is always  a newer version available that might fix the issue. Some compilers are rarely updated so if there is an issue months may go by before it is resolved.
Anyway some compiler writers sharpen their tools every day, others twice yearly, so the tool is sometimes blunt and won't cut. A sharp tool does cut but care is needed to avoid a bleeding edge.

[nagkiller]

Although I am using an older version, it's really great to see that has updates.

Nothing is 100% sure!!! However, they work for it!!!

And this is important!!!

 

[danifox]

It's good to see that they update CCS frequently, but I think that they should test their software more deeply before releasing new versions. As a student, I have developed my project under an older version and I can't complain about it because I didn't pay for it. However, once I get a job and I have to make the decision  of buying a compiler, I have doubts about whether CCS is a trustable one or not.

[FTL]

It is good that they seem to create versions that support newer chips quickly. But . . . .

The quality of their code base must truly be terrible. It is probably written by one programmer that has created a mess and does not know how to fix it properly.

The types of bugs that keep getting fixed mean that the code is certainly poorly tested. I have managed programmers in the past that just had no idea of how to create a quality product. One guy would consider the code complete if it worked for one test case. Once he flagged a few hundred lines of code as complete, I could always find at least 10 bugs by simply providing test data that was not well-formed. i.e. Try to add something that was already there, or maybe add to the beginning or end of the data, or add some data with an invalid field. He would never code or test for those sort of edge conditions. Before I came along, he would ship that sort of code out to customers. After I left, he continued doing that sort of thing and the company is now pretty much bankrupt, just supporting one or two remaining customers, but not making any money at it.

I will give CCS credit though in that they have a difficult product to test. You can't just create a set of test cases and do automated testing. To properly test that code, you need to load it into a PIC and test it. But that often needs specialized hardware to test each function. The other thing they have to deal with is the huge number of PICs that Microchip produces. It would be pretty much impossible to have a hardware test setup for every feature of every PIC.

Some companies solve the rapid release issues by creating beta and production releases. in theory, the production releases come out less often and have been more fully tested, while the beta releases have only passed unit testing, and not full integration testing. the problem here is that CCS does not even do a good job of unit testing. There have been several bugs fixed in CCS 5 over the last year of so where the bug was reported as something like, "fixed an issue introduced with V5.xxx", where xxx was just released a couple of days ago.

Overall, I need to really trust a compiler. When I have a hard to find bug, I need to now that it is my code, not the compiler that had an issue. Many moons ago (in the previous century), when I wrote tens of thousands of lines of code under a Borland C++ compiler running under OS2, I did find a bug in the compiler generated code. I spent over a week looking for it in my code (it was a large complex system and the bug intermittently caused a crash in seemingly random places). Working with what you know is a buggy product makes that process much more difficult. You need to be able to trust your tools - especially with a compiler.

[mario2000]

   The CCS compiler, from version 5.037 onwards, the PcdWiz.exe for PIC24 not work,
   giving error when running.  

[duende_azul]

CCS 5.042

Link fixed .. forgive my mistake.

[danifox]

CCS 5.042

https://mega.co.nz/#!RUxHHCRL!vS92zzrPZbyImK52bciGAors6ghdvf7mDhqAJdIivYg

Thank you very much but I think you uploaded CCS 5.040 again, instead of 5.042.

[mario2000]

CCS 5.042

https://mega.co.nz/#!RUxHHCRL!vS92zzrPZbyImK52bciGAors6ghdvf7mDhqAJdIivYg

The error continues

[duende_azul]

Thank you very much but I think you uploaded CCS 5.040 again, instead of 5.042.

You're right .. it's fixed.

[luiz_a2software]

link dead

[praksk]

link dead

Try this link...

[mario2000]

link dead

Link mega pcwhdupd+PcdWiz+CCSUniversalKeygen 

[towlerg]

link dead

Copy and paste the whole link, not just the underscored bit.

[Interprog]

Posted by: duende_azul
Insert Quote
CCS 5.042

https://mega.co.nz/#!ZZxlwADK!oi9cAjsRze5h4YgJJfk70Q0CmwF3asefnUprq5FvbaE

Link fixed .. forgive my mistake.

Link and Software are OK 

[voltas5]

A new release of the CCS C compiler for PICmicro(r) microcontrollers
has been released. The new version is: 5.043

Recent changes include:
5.043  Forward referencing a 'rom' data item with & in an initializer is now accepted
5.043  The default fuse settings have changed back to the way they were in older compilers
5.043  Fixed a type conversion problem between float and int in some relational expressions on a dsPIC
5.043  The address used for & on large "rom" structures when passed to a  built in function on PIC24 is fixed
5.043  setup_wdt() in newer 12 bit parts no longer glitches the pull-ups
5.043  A number of updates have been made for the new EV chips


Updated:  PIC16F1784
Updated:  PIC16F1786
Updated:  PIC16F1787
Updated:  PIC16F1788
Updated:  PIC16F1789

[duende_azul]

Last version:

CCS 5.043

[voltas5]

A new release of the CCS C compiler for PICmicro(r) microcontrollers
has been released. The new version is: 5.044

Recent changes include:
5.044  Fixed a bug in PCD in subtraction of 32 bit numbers in complex expresions involving the -> operator
5.044  A warning in string.h is now gone
5.044  Corrected a formatting issue in %W (%LU, %LD should not have been affected)
5.044  A file signing issue has been resolved


Updated:  PIC16F648A
Updated:  PIC18F65J94

[duende_azul]

Version 5.044

New version in five minutes...