#ASM problem with 5.108

Ttelmah · Joined: 11 Mar 2010 Posts: 19966

Just had an odd problem with some old code using #ASM/#ENDASM
refusing to compile on 5.108. Gives an error 53.
Compiling for PIC18.
Same code compiles on 5.107.
Their examples (or at least ex_glint.c), give the same problem.

CCS have confirmed it is as a problem and are meant to be doing a new
release in the next couple of days to fix it.

So 'beware' to anyone using assembler with 5.108... Sad

hmmpic · Joined: 09 Mar 2010 Posts: 314 Location: Denmark

Sorry for that, and thanks for the info.

For me it have always looked like this:
All new release is beta build, and will be followed by a newer release, there is a bug fix for the latest release:-)

Therefore i have suggested CCS many times to release beta build. And only do the release when they have tested it.
The release log is for nothing too it only show so little of what is changed.
CCS never reply me on my suggestion...

Ttelmah · Joined: 11 Mar 2010 Posts: 19966

Yes, readme.txt at one time was being updated, now it gives almost
nothing. I think also the testing they carry out seems to have dropped
off in recent releases, making the new releases closer to alpha than
beta...

temtronic · Posted: Fri May 06, 2022 5:22 am

Let's face it guys, there's just tooooo many versions of PICs being made these days ! How anyone can check out every feature of a compiler on literally 100s of different but similar PICs is kinda mind blowing.

We should go back to 16C84s ! Life was a lot simpler and easier , yeah, when us dinosaurs roamed..... Laughing

hmmpic · Joined: 09 Mar 2010 Posts: 314 Location: Denmark

@temtronic ...Therefore releasing a beta will be so much more the case!

Ttelmah · Joined: 11 Mar 2010 Posts: 19966

True.
However for most chips 95% of the routines and handling will be the same
as their brethren. This will also be the same or similar to their ancestors.
The recent problems with parts of the newer compilers are in a few cases
incorrect handling of a new feature, or miss-entering feature details in
the database. However far more seem to be attempts by CCS to (potentially)
change/improve things. These need to be properly checked when added to
the release compilers. In some cases I suspect they have lost people who
originally wrote things, and then a new programmer tries to change things
without realising all the implications of what they are doing. Sad

PrinceNai · Joined: 31 Oct 2016 Posts: 554 Location: Montenegro

PrinceNai · Joined: 31 Oct 2016 Posts: 554 Location: Montenegro

BTW, you can't buy any. They all go to Philips toothbrushes and other 10 million series products.

PrinceNai · Joined: 31 Oct 2016 Posts: 554 Location: Montenegro

On the other hand, I also sympathize with the developers at the Microchip. They are the developers, paid to develop. So they do develop. What new they are developing is anybody's guess. Features unknown to CCS, beyond anything C or even Z can handle?

Ttelmah · Joined: 11 Mar 2010 Posts: 19966

Just been fixed in 5.109.

Now to see what else has gone wrong...