Can CCS be made more stable inside of MPLAB..?

ChrisL · Joined: 16 Apr 2007 Posts: 26 Location: Arizona

Hi all,

Our environment:
CCS V4.025
CCS MPLAB Plug-in V6.20
XP SP2
MPLAB V7.60
PC HP-Pavallion

Just for reference, we don't see any of this when using MC C-18 compiler.

There appear to be a few stability issues when using CCS inside of MP-LAB. Here is what we see many, many times on a daily basis:

1 - MP-LAB will randomly abort on it's own with no error messages
2 - MP-LAB will randomly abort on it's own with error messages about un-readable memory locations
3 - Breakpoints stop functioning
4 - CCS, on compile, stops at random places in the overall compile process
5 - The hardest to deal with... Suddenly your code will stop working or start working in really strange ways. Something goes wrong behind the scenes but there is no indication of it having happened to the developer. If you save your work and quit MPLAB you will ALWAYS get a "cant read some location of memory" error. Our guys have spent hours debugging code that functions perfectly due to this error.

We really like CCS, but would desire a more stable environment inside of MPLAB.

We would consider using PCW, but all 6 of our ICD units are MC ICD-2s. It does not appear that they will work with PCW... IS that correct..?
_________________
Thank you,
Chris

PCM programmer · Joined: 06 Sep 2003 Posts: 21708

What PIC are you using ?

kevcon · Joined: 21 Feb 2007 Posts: 142 Location: Michigan, USA

Try up grading to MPLAB 7.62, it's a bit more stable than 7.60 but still crashes randomly.

There are several psts about this, do a search and there are instructions about downgrading.

ChrisL · Joined: 16 Apr 2007 Posts: 26 Location: Arizona

We are currently developing on:

PIC 16F883
PIC 16F887
PIC 18F6680
PIC 18F8660

All part of the same project.
_________________
Thank you,
Chris

ckielstra · Joined: 18 Mar 2004 Posts: 3680 Location: The Netherlands

Here another observation:
I'm using an older version of MPLAB and my system is very stable, none of the problems you mention.

My environment:
Our environment:
CCS V3.226
CCS MPLAB Plug-in V6.20
XP SP2
MPLAB V7.41
PC Dell desktop

CCS compiler version 4.025 is known to have many problems. The new v4.0xx compiler became more or less usable around version number 4.030. Upgrading your compiler will not solve all your problems but now you are working with at least one broken tool.

I wouldn't count on the MPLAB plugin version numbering to be correct. I downloaded my plugin 2 Sept 2005 which has the same version number as the file available for download today, but the new version is 655kb while my old version is 293kb. Most likely CCS did not update the version number.

With both the plug in and MPLAB having changed it is hard to tell which product is at error. Is it Microchip or CCS who caused the problems? The correct way to handle these problems is to send both Microchip and CCS support an error report.

The changes from MPLAB version 7.41 to 7.60 are minimal. Is it possible for you to use the older version? I could send you the old plugin I'm using.
Note: the PIC18F8660 is not supported in MPLAB 7.40 but maybe this can be solved by using the v7.60 files.

ChrisL · Joined: 16 Apr 2007 Posts: 26 Location: Arizona

OK,

We also have, but have not installed:

CCS 4.031.30.16

Would you recommend upgrading to that version from 4.025
_________________
Thank you,
Chris

ckielstra · Joined: 18 Mar 2004 Posts: 3680 Location: The Netherlands

Let's put it like this: v4.025 is an alpha quality release. v4.031 becomes close to resemble beta quality; not production quality but most of the old functionality is working and the new features don't crash the compiler every time. I would certainly upgrade the compiler but don't expect too much. The most recent version is v4.057, 26 releases with mostly bug fixes since your 4.031. I haven't seen large complaints on the latest compiler versions in this forum so I expect it is reaching a stable state.

CCS always provide two compiler versions for download at their website. A stable production quality version and a development version for the projects that require support for the newest processors and features.
V3.249 is the last known stable release and this still being present for download should give you a clue as to how CCS see things themselves.

ChrisL · Joined: 16 Apr 2007 Posts: 26 Location: Arizona

OK, will do..

From our other posts you know that we are developing CAN for J1939 requiring support for the INTs on the 18F parts. We will watch for the next PRODUCTION release from CCS and purchase the upgrade (happily). We are quite happy with the CCS product in general. The fact that we were able to convert the entire J1939 application from the MC C-18 application says a lot for CCS. Our only challange was to understand the code, as written, and use the philosophy to write the == in CCS.

The only answer that I need, although we have it working with CCS int handler, is how to call a C function from within: