Re: xwisp2 1.7.01 released

Rob Hamerling <[email protected]> Tue, 20 Sep 2005 09:17:52 +0200
Newsgroups gmane.comp.hardware.microcontrollers.gnupic
Message-ID <[email protected]>
Hi Ciaofan,


Xiaofan Chen wrote:

> Still the verification caused false alarm.

To my opinion this is not a false alarm!

With 'verify' the contents of target memory and hex file are compared 
and differences should be reported. In your case there is a difference 
in the config word, and xwisp2 signals this rightly!
Now you have to find out which of both values is wrong. According to the 
datasheet some bits of the config word are not used and read back as 0, 
but your hex file specifies these bits as 1. So obviously your hex file 
is in error.

I have introduced fusefixedzero and fusefixedone masks to adjust the 
config word bits with the 'write' operation to *suppress* the verify 
alarm with the write-verify operation (implied in the most popular 
command: 'go'). This is done for the convenience of the 'simple' user (I 
myself use practically only the 'go' command!)

The choice for a change in xwisp2 is:
  1. disable suppress of the error report with write-verify
  2. enable suppress of the error report with verify
To my opinion: when chosing for the second option 'verify' is crippled: 
you will not find errors in the config word specification of your 
program (either caused by yourself or by the compiler you are using).

Regards, Rob.


-- 
Rob Hamerling, Vianen, NL phone +31-347-322822
homepage: http://www.robh.nl/