Post 1286716 by mattskala@mstdn.io
(DIR) More posts by mattskala@mstdn.io
(DIR) Post #1259711 by mattskala@mstdn.io
2018-11-16T22:52:38Z
0 likes, 0 repeats
This is it - the smallest test case I can find that demonstrates the bug in the PIC24 assembler, by computer search followed by a little bit of human editing. Assemble this with listing off or the -am option and the symbol x will be at offset 2, correctly; assemble with listing on but not the -am option, and the symbol x will be at offset 1, which should be impossible..macro m.byte 1x:.endm.byte 1m
(DIR) Post #1263817 by woolie@mastodon.social
2018-11-17T01:51:23Z
0 likes, 0 repeats
@mattskala you do a lot of PIC24?
(DIR) Post #1263820 by mattskala@mstdn.io
2018-11-17T01:51:45Z
0 likes, 0 repeats
@woolie No, I only just started.
(DIR) Post #1263833 by mattskala@mstdn.io
2018-11-17T01:52:48Z
0 likes, 0 repeats
@woolie Probably going to use it in a future project, but at the moment I'm experimenting trying to learn. Kind of disheartening to trip over an assembler bug so early in that effort.
(DIR) Post #1286692 by woolie@mastodon.social
2018-11-18T00:22:22Z
0 likes, 0 repeats
@mattskala I did a whole lot of PIC16 this year. Despite the name it is 8 bit. I was looking at PIC24 and PIC32 for some more muscle, but I think I will just go to Cortex M series instead. Still evaluating options.
(DIR) Post #1286716 by mattskala@mstdn.io
2018-11-18T00:23:45Z
0 likes, 0 repeats
@woolie Yeah, I remember you said. PIC24 is mostly looking good for my current project - which isn't DSP, but a USB-to-CV interface. But finding bugs in the assembler like this is kinda worrying for me. I have a ticket open with Microchip support on it.
(DIR) Post #1292109 by woolie@mastodon.social
2018-11-18T06:09:54Z
0 likes, 0 repeats
@mattskala please do share any response you get. TIA