Beta compiler problems

[doggy]

I had been posting on the main thread and wondered why no response and no-one else commenting, so thanks to Janni for pointing out this specific thread.

I have been running the Beta compiler with 18F26K42 and also with PIC18F1330.

a. 18F26K42. I had problems relating to two 256byte arrays crossing pages and also the library EEPROM. After page aligning the two arrays and using my own EEPROM routines, the code (which is a Battery Management System module attached to a Lithium cell) has been running for a couple of weeks. So now this code, which uses NO M.E library functions is okay.

Modules used are:

1. USART
2. Interrupts (legacy).
3. Timers 2 and 0.
4. PWM output
5. Two A/D channels.
6. High precision Texas external voltage reference.
7. Several input and output bits.
8. EEPROM.

One minor problem is that some of the bit definitions are missing. The following were missing for me, but I know from a quick look that others are also missing:

T2CON.ON, CCP1CON.EN, T0CON0.EN

EEPROM: WHY DID YOU STILL NOT FIX THE EEPROM Editor problem????? WHEN will it be fixed????

b. 18F1330. This is a project driving a three phase buck converter. Uses EEPROM, A/D, UART, PWM*6, general I/O, Soft_SPI, two timers. There is a major issue using the new Beta compiler, but it seems subtle. This project only uses the M.E. EEPROM routines and the Soft_SPI. Now, I had been particularly pleased with the Soft_SPI which is interfaced to another project which is 18F26K42 using MPLAB and hardware SPI.
But using code from the new compiler, the SoftSPI routine randomly stops working. When I say stops working, the Slave Select line still toggles but there are no clocks and no SDO. I put in a workaround to detect it stopping and reset the code. But logging shows that it is happening 60 to 100 times a day. I reinstalled the 7.2 compiler on another machine and recompiled the identical code. It has been running for 24hrs without a single error!

The SoftSPI is sending 30 3byte messages a second to the other processor and receiving the same number of messages back. Interrupts not enabled (but maybe the M.E EEPROM library routines are screwing this up). So, I have now sent/received 4.8million messages with no SoftSPI failure. Nor were there problems previously with MPLAB.

I will see if I can rerun the test with the 7.3 compiler and my own (not M.E's library routines) and see if SoftSPI is now okay with 7.3. But because these have now become production environments and something going wrong can cause a massive electrical "BANG" I am a little reluctant. (10,000amps and 28volts for the electrically savvy).

Regds,
David

[filip]

Hi,

One minor problem is that some of the bit definitions are missing. The following were missing for me, but I know from a quick look that others are also missing:

T2CON.ON, CCP1CON.EN, T0CON0.EN

EEPROM: WHY DID YOU STILL NOT FIX THE EEPROM Editor problem????? WHEN will it be fixed????

I apologize for this, we will fix this in the next official release of the compiler.

As for the Soft SPI issues, could you please tell us more about it ?
We have tried simple example but had no problems.

Regards,
Filip.