Hi Rich,
I'm still testing: here are my results so far:
Text Machines
A) Samsung Netbook, Intel Atom, Windows 7 starter
B) Sony VGC-LV1S, Intel Core 2 Duo, Windows Vista pro
C) Dell Latitude E7240, Core i5, Windows 8.1
LPCXpresso Versions
7.8.0_426 (latest version)
7.6.2_326
Tests:
2. Run pre-existing make file in Device folder
Results
On machine A with v7.8.0 both tests built successfully and created hex files around 307K in size. However, when deployed to the pixy I got the invalid address range error and tried quite a few things to try to resolve. I figured it might be a linking issue. I have done a lot of dev on that machine and have versions of Cygwin and mingw and generally all sorts of stuff in the PATH..
I hadn't seen the other forum comment about versions but discovered you were on v7.6.2 from doing a GIT diff
On machine B with v7.6.2 I got a working hex file.
Replacing the version on on machine A with v7.6.2 and I got a working hex file from both tests. Great!
I did some comparison between the 2 versions. The noticeable difference is that the 7.8.0 hex file lines start :10800 rather than :10000. I.e.
:020000041001E9
:108000000000000000000000000000000000000070
:108010000000000000000000000000000000000060
:108020000000000000000000000000000000000050
rather than
:020000041400E6
:1000000000800020CD010014910100149301001420
:1000100095010014970100149901001400000000DC
:100020000000000000000000000000009B01001420
The build steps seemed to produce different results in the main_m0 project - the rest looked the same.
The cool thing is that it is working for me now on 7.6.2. So it's worth updating the instructions here:
to specify 7.6.2 or add a GitHub readme. I'm going to put in a couple more hours testing before cracking on with my project. Happy to contribute directly to instructions or have a go at resolving on v7.8.0. Whatever you think adds most value?
Thanks for all the effort you've put into this.
Rich