SOLVED - Issues to compile TAQOZ with flexprop v 5.9.8

jurop

As per subject, is there anybody achieving to compile TAQOZ using flexprop v 5.9.8 ? The same file used to (and still) compile(s) with flexprop v 5.3.0 so I suspect some sort of regression happened at a certain point. Just collecting opinions and hints before bothering Eric/totalspectrum

ersmith

I suspect the Spin2 language has changed in ways that have broken Peter's code, e.g. that some of the new keywords like DEBUG are interfering with existing symbols. But if you could post the exact version of TAQOZ that you're trying, and the error messages, that would be helpful.

jurop

Hi Eric,

thanks for chiming in! Attached the TAQOZ version that does not compile, AFAIR it should be the last one released. Flexprop 5.3.0 compiles it with a couple of warnings, whereas Flexprop V 5.9.8 spits 2 warnings (RES has 0 space) and 5 errors. 2x Only addresses of identifiers allowed and 3x Unknown symbol (l3, l2, L0).

ersmith

Some of the problems are indeed Spin language changes, and you may run into them with PNut as well -- specifically, there are new operators like +. and -. for floating point, and so an expression like badthen-.L0 is now being parsed as "badthen" "-." "L0" rather than "badthen" "-" ".L0". The solution is fairly simple, just insert spaces around the "-" in the expressions causing problems (the ones involving .L0, .L1, and so on).

The remaining problem I see is an unknown symbol _BAUD, which actually is a puzzle because I don't see any definition for _BAUD, but there is one for _BAUDD with an extra D. Could a typo have crept into the file? flexspin 5.3.0 gives the same error for me, so I don't know how that file compiiled for you.

jurop

@ersmith said:
Some of the problems are indeed Spin language changes, and you may run into them with PNut as well -- specifically, there are new operators like +. and -. for floating point, and so an expression like badthen-.L0 is now being parsed as "badthen" "-." "L0" rather than "badthen" "-" ".L0". The solution is fairly simple, just insert spaces around the "-" in the expressions causing problems (the ones involving .L0, .L1, and so on).

That was it. I added all the needed spaces and now it compiles again, many thanks!

@ersmith said:
The remaining problem I see is an unknown symbol _BAUD, which actually is a puzzle because I don't see any definition for _BAUD, but there is one for _BAUDD with an extra D. Could a typo have crept into the file? flexspin 5.3.0 gives the same error for me, so I don't know how that file compiiled for you.

Sorry, that was my fault - I replaced the _BAUD symbol with _BAUDD a long ago in order to compile Taqoz and very likely it crept in the file I sent you via copy&paste from two different sources. On my copy the remaining BAUD is in a comment so it works.

The attached TAQOZ file does compile with Flexprop 5.9.8.
@bob_g4bby, do you know how to let @"Peter Jakacki" know about these needed changes?

bob_g4bby

Hi @jurop

Not sure how to contact Peter reliably. I get the occasional email from him, but he doesn't answer mine to him. He admits he's not good with emails and is also suffering from 'burnout' on P2.

I think he looks in on this forum, so anything new in the Forth section would eventually be noticed.

There's nothing wrong with you publishing what you've done as a new thread in the Forth section as long as it zips down small enough. I maintain a number of pdfs like that, and just replace them at the top of the thread when I upissue. Give what you do an issue number, so that people know when it has been changed (Anything you write on the forum, you can re-edit as many times as you like.

Cheers, Bob

Maciek

I think, chances are he'll notice if you post to his sourceforge Tachyon page. He may even have notifications set up, who knows. Anyway, lots of good TAQOZ/TACHYON stuff on his sourceforge wiki.

Cluso99

I know Peter checks occasionally.
If you need to get a quicker response you can PM me and I'll ask him. I video with him most weeks, some times much more