Re: [Amforth] successful compilation with avra
Perhaps a howto on the site not as an email thread would be in order. This > may not be a big deal for windows users, but people that want to use > free-libre software will be drawn to use amforth if the entire build > toolchain is FLOSS. > That is a very good suggestion. The whole thing is ridiculously simple but having the steps set down, or better yet, cooked into the amForth repo itself. All I did was to clone the repo we were talking about then run 'make' from the top of it. Since I am extraordinarily lazy this summer I didn't even bother to install it somewhere into my PATH. I just made a link and copied that into my project directory so I could change my makefile to './avra_amforth' because I'm a lazy old man and Debian. Then everything else just works like it always did with avrasm2. But yes, the proper way would be to have a version of avra that works maybe in the tools directory that could be built once then used from there. But I am at best an old hack not a proper programmer. Personally, my thought was to fork into a github repo from the git stuff that is already done by Erich and use that. I'm honestly very pleasantly surprised to see a number of you all talking about this. I feared I was just yelling into the void (or shaking my fists at clouds, choose your metaphor) when I put this up here. There is so much great stuff on the Sourceforge site that is really helpful to get spun up from scratch. Adding a 'How to use avra' would most certainly be beneficial. But, even I, a dedicated lover of email no matter how much the kids mock, find the mailing list clunky to carry on conversations. I miss the ease of issue pages. I do understand though that some people would be turned off by github. So anyhow, that is my 40 degree C afternoon rant. For some reason the gods have decided that we need more heat in Greece again. I think the moving forward part has to be decided by Erich if they want to add the avra stuff into the repo. I have a few days until the end of the week that I will be time rich so maybe I'll try to get a clean copy of the git work that has been done and the avra stuff. I wish there was a real avra repo that we could just mash into the source tree properly so it could be updated correctly, but it seems more like a one off at this point. I know my debian one is way way way old and I don't see that changing. As always, all the best. I wish I would have discovered this project long before I did. Mark > Brian-in-ohio > > On Mon, Aug 28, 2023 at 7:47 AM Jan Kromhout via Amforth-devel < > amforth-devel@lists.sourceforge.net> wrote: > > > Hello, > > > > Seems simple to use the avra compiler. > > Is it posible to show example how the build is doing? > > > > Cheers, > > > > Jan > > > > Verstuurd vanaf mijn iPad > > > > > Op 28 aug. 2023 om 09:02 heeft Mark Roth het > > volgende geschreven: > > > > > > You are using the same repo I was for avra Tristan. I just cloned the > > > issue-54 branch and built it here. Apart from a couple of 'zero byte > in a > > > .DB' warnings things do seem to be working fine here as well. > > > > > >> On Mon, Aug 28, 2023 at 9:54 AM tristan wrote: > > >> > > >> Hello, > > >> > > >> I flashed the hex files created by avra to an uno and, to the extent > > >> that getting a serial prompt, defining a word and executing it > > >> constitutes a test, it worked perfectly. > > >> > > >>> All that being said, I would be very interested to see the > > >>> changes, maybe, just maybe we can fix the amForth source tree > > >>> enough to make avra happy. > > >> > > >> No changes to the source tree were needed to create the uno hex files. > > >> The only change made was to edit the Makefile to use avra. > > >> > > >> Best wishes, > > >> Tristan > > >> > > >> > > >>> On 2023-08-27 06:29, Tristan Williams wrote: > > >>> Hello Mark, Brian, Erich, George > > >>> > > >>> Thank you! A very welcome set of messages on a bank holiday > > >>> weekend. For non-windows users having avra as the assembler in the > > >>> build chain would go along way in making AmForth more approachable > and > > >>> maintainable. > > >>> > > >>> I think this is the repo for avra that does not have the macro/ > > >>> parenthesis issue you mention[1] > > >>> > > >>> https://github.com/srtlg/avra/tree/development > > >>> > > >>> I downloaded it and built it on macOS (requiring only typing 'make') > > >>> and updated the AmForth Makefile to run arva. The updated makefile > > >>> built the hex files for an uno with AmForth 6.9. I did not experience > > >>> any issues with d0< but I recall there were some changes in that > > >>> area between 6.8 and 6.9. I've not flashed it yet as I have to dig > out > > >>> an uno from storage but the hex files are the same size and with zero > > >>> diffs when compared with my previous wine/avrasm32 builds. > > >>> > > >>> -rw-r--r-- 1 tw staff 29346 26 Aug 17:53 uno.hex > > >>> -rw-r--r-- 1 tw staff 29346 26 Aug 16:29 save.hex > > >>> -rw-r--r-- 1 tw staff
Re: [Amforth] successful compilation with avra
Perhaps a howto on the site not as an email thread would be in order. This may not be a big deal for windows users, but people that want to use free-libre software will be drawn to use amforth if the entire build toolchain is FLOSS. Brian-in-ohio On Mon, Aug 28, 2023 at 7:47 AM Jan Kromhout via Amforth-devel < amforth-devel@lists.sourceforge.net> wrote: > Hello, > > Seems simple to use the avra compiler. > Is it posible to show example how the build is doing? > > Cheers, > > Jan > > Verstuurd vanaf mijn iPad > > > Op 28 aug. 2023 om 09:02 heeft Mark Roth het > volgende geschreven: > > > > You are using the same repo I was for avra Tristan. I just cloned the > > issue-54 branch and built it here. Apart from a couple of 'zero byte in a > > .DB' warnings things do seem to be working fine here as well. > > > >> On Mon, Aug 28, 2023 at 9:54 AM tristan wrote: > >> > >> Hello, > >> > >> I flashed the hex files created by avra to an uno and, to the extent > >> that getting a serial prompt, defining a word and executing it > >> constitutes a test, it worked perfectly. > >> > >>> All that being said, I would be very interested to see the > >>> changes, maybe, just maybe we can fix the amForth source tree > >>> enough to make avra happy. > >> > >> No changes to the source tree were needed to create the uno hex files. > >> The only change made was to edit the Makefile to use avra. > >> > >> Best wishes, > >> Tristan > >> > >> > >>> On 2023-08-27 06:29, Tristan Williams wrote: > >>> Hello Mark, Brian, Erich, George > >>> > >>> Thank you! A very welcome set of messages on a bank holiday > >>> weekend. For non-windows users having avra as the assembler in the > >>> build chain would go along way in making AmForth more approachable and > >>> maintainable. > >>> > >>> I think this is the repo for avra that does not have the macro/ > >>> parenthesis issue you mention[1] > >>> > >>> https://github.com/srtlg/avra/tree/development > >>> > >>> I downloaded it and built it on macOS (requiring only typing 'make') > >>> and updated the AmForth Makefile to run arva. The updated makefile > >>> built the hex files for an uno with AmForth 6.9. I did not experience > >>> any issues with d0< but I recall there were some changes in that > >>> area between 6.8 and 6.9. I've not flashed it yet as I have to dig out > >>> an uno from storage but the hex files are the same size and with zero > >>> diffs when compared with my previous wine/avrasm32 builds. > >>> > >>> -rw-r--r-- 1 tw staff 29346 26 Aug 17:53 uno.hex > >>> -rw-r--r-- 1 tw staff 29346 26 Aug 16:29 save.hex > >>> -rw-r--r-- 1 tw staff239 26 Aug 17:53 uno.eep.hex > >>> -rw-r--r-- 1 tw staff239 26 Aug 16:29 save.eep.hex > >>> > >>> > >>> Best wishes, > >>> Tristan > >>> > >>> [1] https://github.com/Ro5bert/avra/issues/54 > >>> > >>> > >>> On 25Aug23 17:12, George Herzog wrote: > Thanks for your efforts. > > People don't often appreciate how much knowledge and effort goes into > successful compilation of code. > > > > On Fri, Aug 25, 2023, 3:15 PM Erich Wälde > wrote: > > > Hello Brian and Mark, > > > > very nice to see emails on this list :) > > > > Compiling amforth with avra? > > > > I have made numerous experiments a long time ago and again more > > recently. If memory serves me well: > > - Amforth had been good with avra, at least in the 4.x range. > > - However, avrasm2.exe could do more clever tricks, and Matthias > > started using those. > > - I did make a fork of amForth from Version 5.0, this can be > > assembled with avra, see: > > https://git.sr.ht/~ew/hbv3_am50forth > > - avra received a bit of attention not so long ago (same repo > > you found): > > https://github.com/Ro5bert/avra > >> $ avra --version > >> AVRA: advanced AVR macro assembler (version 1.4.2) > > which among other changes now includes my favourite atmega644p. > > > > So, I am currently dabbling with my fork again in the hope to > > eventually catch that problem of long term stability. There is > > absolutely no reason, why I have to reprogram one or two of my > > controllers a few times per year, because they do not start up > > after a power cycle, which in turn is done, because the > > communication with that controller ceases to work. I went back > > to amforth 5.0 for simplicity reasons. > > > > > > All that being said, I would be very interested to see the > > changes, maybe, just maybe we can fix the amForth source tree > > enough to make avra happy. > > > > > > Cheers, > > Erich > > > > Brian K Navarette writes: > > > >> That is awesome news! > >> Brian-in-ohio > >> > >> > >> On Thu, Aug 24, 2023 at 2:59 PM Mark Roth > >> wrote: > >> > >>> Hello AmForth. It has been some time and quite weird things since > >> last > > I've > >>> been
Re: [Amforth] successful compilation with avra
Hello, Seems simple to use the avra compiler. Is it posible to show example how the build is doing? Cheers, Jan Verstuurd vanaf mijn iPad > Op 28 aug. 2023 om 09:02 heeft Mark Roth het volgende > geschreven: > > You are using the same repo I was for avra Tristan. I just cloned the > issue-54 branch and built it here. Apart from a couple of 'zero byte in a > .DB' warnings things do seem to be working fine here as well. > >> On Mon, Aug 28, 2023 at 9:54 AM tristan wrote: >> >> Hello, >> >> I flashed the hex files created by avra to an uno and, to the extent >> that getting a serial prompt, defining a word and executing it >> constitutes a test, it worked perfectly. >> >>> All that being said, I would be very interested to see the >>> changes, maybe, just maybe we can fix the amForth source tree >>> enough to make avra happy. >> >> No changes to the source tree were needed to create the uno hex files. >> The only change made was to edit the Makefile to use avra. >> >> Best wishes, >> Tristan >> >> >>> On 2023-08-27 06:29, Tristan Williams wrote: >>> Hello Mark, Brian, Erich, George >>> >>> Thank you! A very welcome set of messages on a bank holiday >>> weekend. For non-windows users having avra as the assembler in the >>> build chain would go along way in making AmForth more approachable and >>> maintainable. >>> >>> I think this is the repo for avra that does not have the macro/ >>> parenthesis issue you mention[1] >>> >>> https://github.com/srtlg/avra/tree/development >>> >>> I downloaded it and built it on macOS (requiring only typing 'make') >>> and updated the AmForth Makefile to run arva. The updated makefile >>> built the hex files for an uno with AmForth 6.9. I did not experience >>> any issues with d0< but I recall there were some changes in that >>> area between 6.8 and 6.9. I've not flashed it yet as I have to dig out >>> an uno from storage but the hex files are the same size and with zero >>> diffs when compared with my previous wine/avrasm32 builds. >>> >>> -rw-r--r-- 1 tw staff 29346 26 Aug 17:53 uno.hex >>> -rw-r--r-- 1 tw staff 29346 26 Aug 16:29 save.hex >>> -rw-r--r-- 1 tw staff239 26 Aug 17:53 uno.eep.hex >>> -rw-r--r-- 1 tw staff239 26 Aug 16:29 save.eep.hex >>> >>> >>> Best wishes, >>> Tristan >>> >>> [1] https://github.com/Ro5bert/avra/issues/54 >>> >>> >>> On 25Aug23 17:12, George Herzog wrote: Thanks for your efforts. People don't often appreciate how much knowledge and effort goes into successful compilation of code. On Fri, Aug 25, 2023, 3:15 PM Erich Wälde wrote: > Hello Brian and Mark, > > very nice to see emails on this list :) > > Compiling amforth with avra? > > I have made numerous experiments a long time ago and again more > recently. If memory serves me well: > - Amforth had been good with avra, at least in the 4.x range. > - However, avrasm2.exe could do more clever tricks, and Matthias > started using those. > - I did make a fork of amForth from Version 5.0, this can be > assembled with avra, see: > https://git.sr.ht/~ew/hbv3_am50forth > - avra received a bit of attention not so long ago (same repo > you found): > https://github.com/Ro5bert/avra >> $ avra --version >> AVRA: advanced AVR macro assembler (version 1.4.2) > which among other changes now includes my favourite atmega644p. > > So, I am currently dabbling with my fork again in the hope to > eventually catch that problem of long term stability. There is > absolutely no reason, why I have to reprogram one or two of my > controllers a few times per year, because they do not start up > after a power cycle, which in turn is done, because the > communication with that controller ceases to work. I went back > to amforth 5.0 for simplicity reasons. > > > All that being said, I would be very interested to see the > changes, maybe, just maybe we can fix the amForth source tree > enough to make avra happy. > > > Cheers, > Erich > > Brian K Navarette writes: > >> That is awesome news! >> Brian-in-ohio >> >> >> On Thu, Aug 24, 2023 at 2:59 PM Mark Roth >> wrote: >> >>> Hello AmForth. It has been some time and quite weird things since >> last > I've >>> been here. I am still using AmForth with my trusty atmega1284p and > learning >>> the language as time permits. I remember having heard talk that >> avra had >>> gotten (almost) to the point of being able to compile the source >> tree > here. >>> First I tried with 1.3 I think and it failed miserably. Then I >> found a > repo >>> on github (Ro5bert/avra) that seemed to almost but not quite do >> it. I > was >>> getting a pile of errors for macro calls. So looking into the >> issues I > saw >>> that someone had forked that repo and
Re: [Amforth] successful compilation with avra
You are using the same repo I was for avra Tristan. I just cloned the issue-54 branch and built it here. Apart from a couple of 'zero byte in a .DB' warnings things do seem to be working fine here as well. On Mon, Aug 28, 2023 at 9:54 AM tristan wrote: > Hello, > > I flashed the hex files created by avra to an uno and, to the extent > that getting a serial prompt, defining a word and executing it > constitutes a test, it worked perfectly. > > > All that being said, I would be very interested to see the > > changes, maybe, just maybe we can fix the amForth source tree > > enough to make avra happy. > > No changes to the source tree were needed to create the uno hex files. > The only change made was to edit the Makefile to use avra. > > Best wishes, > Tristan > > > On 2023-08-27 06:29, Tristan Williams wrote: > > Hello Mark, Brian, Erich, George > > > > Thank you! A very welcome set of messages on a bank holiday > > weekend. For non-windows users having avra as the assembler in the > > build chain would go along way in making AmForth more approachable and > > maintainable. > > > > I think this is the repo for avra that does not have the macro/ > > parenthesis issue you mention[1] > > > > https://github.com/srtlg/avra/tree/development > > > > I downloaded it and built it on macOS (requiring only typing 'make') > > and updated the AmForth Makefile to run arva. The updated makefile > > built the hex files for an uno with AmForth 6.9. I did not experience > > any issues with d0< but I recall there were some changes in that > > area between 6.8 and 6.9. I've not flashed it yet as I have to dig out > > an uno from storage but the hex files are the same size and with zero > > diffs when compared with my previous wine/avrasm32 builds. > > > > -rw-r--r-- 1 tw staff 29346 26 Aug 17:53 uno.hex > > -rw-r--r-- 1 tw staff 29346 26 Aug 16:29 save.hex > > -rw-r--r-- 1 tw staff239 26 Aug 17:53 uno.eep.hex > > -rw-r--r-- 1 tw staff239 26 Aug 16:29 save.eep.hex > > > > > > Best wishes, > > Tristan > > > > [1] https://github.com/Ro5bert/avra/issues/54 > > > > > > On 25Aug23 17:12, George Herzog wrote: > >> Thanks for your efforts. > >> > >> People don't often appreciate how much knowledge and effort goes into > >> successful compilation of code. > >> > >> > >> > >> On Fri, Aug 25, 2023, 3:15 PM Erich Wälde wrote: > >> > >> > Hello Brian and Mark, > >> > > >> > very nice to see emails on this list :) > >> > > >> > Compiling amforth with avra? > >> > > >> > I have made numerous experiments a long time ago and again more > >> > recently. If memory serves me well: > >> > - Amforth had been good with avra, at least in the 4.x range. > >> > - However, avrasm2.exe could do more clever tricks, and Matthias > >> > started using those. > >> > - I did make a fork of amForth from Version 5.0, this can be > >> > assembled with avra, see: > >> > https://git.sr.ht/~ew/hbv3_am50forth > >> > - avra received a bit of attention not so long ago (same repo > >> > you found): > >> > https://github.com/Ro5bert/avra > >> > > $ avra --version > >> > > AVRA: advanced AVR macro assembler (version 1.4.2) > >> > which among other changes now includes my favourite atmega644p. > >> > > >> > So, I am currently dabbling with my fork again in the hope to > >> > eventually catch that problem of long term stability. There is > >> > absolutely no reason, why I have to reprogram one or two of my > >> > controllers a few times per year, because they do not start up > >> > after a power cycle, which in turn is done, because the > >> > communication with that controller ceases to work. I went back > >> > to amforth 5.0 for simplicity reasons. > >> > > >> > > >> > All that being said, I would be very interested to see the > >> > changes, maybe, just maybe we can fix the amForth source tree > >> > enough to make avra happy. > >> > > >> > > >> > Cheers, > >> > Erich > >> > > >> > Brian K Navarette writes: > >> > > >> > > That is awesome news! > >> > > Brian-in-ohio > >> > > > >> > > > >> > > On Thu, Aug 24, 2023 at 2:59 PM Mark Roth > wrote: > >> > > > >> > >> Hello AmForth. It has been some time and quite weird things since > last > >> > I've > >> > >> been here. I am still using AmForth with my trusty atmega1284p and > >> > learning > >> > >> the language as time permits. I remember having heard talk that > avra had > >> > >> gotten (almost) to the point of being able to compile the source > tree > >> > here. > >> > >> First I tried with 1.3 I think and it failed miserably. Then I > found a > >> > repo > >> > >> on github (Ro5bert/avra) that seemed to almost but not quite do > it. I > >> > was > >> > >> getting a pile of errors for macro calls. So looking into the > issues I > >> > saw > >> > >> that someone had forked that repo and fixed the issue. Something > to do > >> > with > >> > >> not having a space between the opening parenthesis and the macro > name. > >> > So I > >> > >> tracked down the fix branch (srtlg/avra -b development-issue54
Re: [Amforth] successful compilation with avra
Hello, I flashed the hex files created by avra to an uno and, to the extent that getting a serial prompt, defining a word and executing it constitutes a test, it worked perfectly. All that being said, I would be very interested to see the changes, maybe, just maybe we can fix the amForth source tree enough to make avra happy. No changes to the source tree were needed to create the uno hex files. The only change made was to edit the Makefile to use avra. Best wishes, Tristan On 2023-08-27 06:29, Tristan Williams wrote: Hello Mark, Brian, Erich, George Thank you! A very welcome set of messages on a bank holiday weekend. For non-windows users having avra as the assembler in the build chain would go along way in making AmForth more approachable and maintainable. I think this is the repo for avra that does not have the macro/ parenthesis issue you mention[1] https://github.com/srtlg/avra/tree/development I downloaded it and built it on macOS (requiring only typing 'make') and updated the AmForth Makefile to run arva. The updated makefile built the hex files for an uno with AmForth 6.9. I did not experience any issues with d0< but I recall there were some changes in that area between 6.8 and 6.9. I've not flashed it yet as I have to dig out an uno from storage but the hex files are the same size and with zero diffs when compared with my previous wine/avrasm32 builds. -rw-r--r-- 1 tw staff 29346 26 Aug 17:53 uno.hex -rw-r--r-- 1 tw staff 29346 26 Aug 16:29 save.hex -rw-r--r-- 1 tw staff239 26 Aug 17:53 uno.eep.hex -rw-r--r-- 1 tw staff239 26 Aug 16:29 save.eep.hex Best wishes, Tristan [1] https://github.com/Ro5bert/avra/issues/54 On 25Aug23 17:12, George Herzog wrote: Thanks for your efforts. People don't often appreciate how much knowledge and effort goes into successful compilation of code. On Fri, Aug 25, 2023, 3:15 PM Erich Wälde wrote: > Hello Brian and Mark, > > very nice to see emails on this list :) > > Compiling amforth with avra? > > I have made numerous experiments a long time ago and again more > recently. If memory serves me well: > - Amforth had been good with avra, at least in the 4.x range. > - However, avrasm2.exe could do more clever tricks, and Matthias > started using those. > - I did make a fork of amForth from Version 5.0, this can be > assembled with avra, see: > https://git.sr.ht/~ew/hbv3_am50forth > - avra received a bit of attention not so long ago (same repo > you found): > https://github.com/Ro5bert/avra > > $ avra --version > > AVRA: advanced AVR macro assembler (version 1.4.2) > which among other changes now includes my favourite atmega644p. > > So, I am currently dabbling with my fork again in the hope to > eventually catch that problem of long term stability. There is > absolutely no reason, why I have to reprogram one or two of my > controllers a few times per year, because they do not start up > after a power cycle, which in turn is done, because the > communication with that controller ceases to work. I went back > to amforth 5.0 for simplicity reasons. > > > All that being said, I would be very interested to see the > changes, maybe, just maybe we can fix the amForth source tree > enough to make avra happy. > > > Cheers, > Erich > > Brian K Navarette writes: > > > That is awesome news! > > Brian-in-ohio > > > > > > On Thu, Aug 24, 2023 at 2:59 PM Mark Roth wrote: > > > >> Hello AmForth. It has been some time and quite weird things since last > I've > >> been here. I am still using AmForth with my trusty atmega1284p and > learning > >> the language as time permits. I remember having heard talk that avra had > >> gotten (almost) to the point of being able to compile the source tree > here. > >> First I tried with 1.3 I think and it failed miserably. Then I found a > repo > >> on github (Ro5bert/avra) that seemed to almost but not quite do it. I > was > >> getting a pile of errors for macro calls. So looking into the issues I > saw > >> that someone had forked that repo and fixed the issue. Something to do > with > >> not having a space between the opening parenthesis and the macro name. > So I > >> tracked down the fix branch (srtlg/avra -b development-issue54 if I > >> remember correctly) and built that locally. Then substituted that avra > >> version with the wine one I had been using to build. > >> It still didn't build. Very almost, but not quite. > >> However, the issue was with errors in /avr8/words/d-lesszero.asm about > the > >> Y register not being declared and/or able to be used for the adiw call. > >> Looking into the source tree I found other usages of y in those calls > but > >> they were all in lower case. > >> Yeah, that did fix it. I'm not sure that I can flash my controller here > >> since I'm on summer break but it does seem promising. Or maybe I did > pack > >> my programmer and can give it a go. The file sizes are the same or > similar > >> but there are differences. Granted, I've made changes that may not
Re: [Amforth] successful compilation with avra
Hello Mark, Brian, Erich, George Thank you! A very welcome set of messages on a bank holiday weekend. For non-windows users having avra as the assembler in the build chain would go along way in making AmForth more approachable and maintainable. I think this is the repo for avra that does not have the macro/ parenthesis issue you mention[1] https://github.com/srtlg/avra/tree/development I downloaded it and built it on macOS (requiring only typing 'make') and updated the AmForth Makefile to run arva. The updated makefile built the hex files for an uno with AmForth 6.9. I did not experience any issues with d0< but I recall there were some changes in that area between 6.8 and 6.9. I've not flashed it yet as I have to dig out an uno from storage but the hex files are the same size and with zero diffs when compared with my previous wine/avrasm32 builds. -rw-r--r-- 1 tw staff 29346 26 Aug 17:53 uno.hex -rw-r--r-- 1 tw staff 29346 26 Aug 16:29 save.hex -rw-r--r-- 1 tw staff239 26 Aug 17:53 uno.eep.hex -rw-r--r-- 1 tw staff239 26 Aug 16:29 save.eep.hex Best wishes, Tristan [1] https://github.com/Ro5bert/avra/issues/54 On 25Aug23 17:12, George Herzog wrote: > Thanks for your efforts. > > People don't often appreciate how much knowledge and effort goes into > successful compilation of code. > > > > On Fri, Aug 25, 2023, 3:15 PM Erich Wälde wrote: > > > Hello Brian and Mark, > > > > very nice to see emails on this list :) > > > > Compiling amforth with avra? > > > > I have made numerous experiments a long time ago and again more > > recently. If memory serves me well: > > - Amforth had been good with avra, at least in the 4.x range. > > - However, avrasm2.exe could do more clever tricks, and Matthias > > started using those. > > - I did make a fork of amForth from Version 5.0, this can be > > assembled with avra, see: > > https://git.sr.ht/~ew/hbv3_am50forth > > - avra received a bit of attention not so long ago (same repo > > you found): > > https://github.com/Ro5bert/avra > > > $ avra --version > > > AVRA: advanced AVR macro assembler (version 1.4.2) > > which among other changes now includes my favourite atmega644p. > > > > So, I am currently dabbling with my fork again in the hope to > > eventually catch that problem of long term stability. There is > > absolutely no reason, why I have to reprogram one or two of my > > controllers a few times per year, because they do not start up > > after a power cycle, which in turn is done, because the > > communication with that controller ceases to work. I went back > > to amforth 5.0 for simplicity reasons. > > > > > > All that being said, I would be very interested to see the > > changes, maybe, just maybe we can fix the amForth source tree > > enough to make avra happy. > > > > > > Cheers, > > Erich > > > > Brian K Navarette writes: > > > > > That is awesome news! > > > Brian-in-ohio > > > > > > > > > On Thu, Aug 24, 2023 at 2:59 PM Mark Roth wrote: > > > > > >> Hello AmForth. It has been some time and quite weird things since last > > I've > > >> been here. I am still using AmForth with my trusty atmega1284p and > > learning > > >> the language as time permits. I remember having heard talk that avra had > > >> gotten (almost) to the point of being able to compile the source tree > > here. > > >> First I tried with 1.3 I think and it failed miserably. Then I found a > > repo > > >> on github (Ro5bert/avra) that seemed to almost but not quite do it. I > > was > > >> getting a pile of errors for macro calls. So looking into the issues I > > saw > > >> that someone had forked that repo and fixed the issue. Something to do > > with > > >> not having a space between the opening parenthesis and the macro name. > > So I > > >> tracked down the fix branch (srtlg/avra -b development-issue54 if I > > >> remember correctly) and built that locally. Then substituted that avra > > >> version with the wine one I had been using to build. > > >> It still didn't build. Very almost, but not quite. > > >> However, the issue was with errors in /avr8/words/d-lesszero.asm about > > the > > >> Y register not being declared and/or able to be used for the adiw call. > > >> Looking into the source tree I found other usages of y in those calls > > but > > >> they were all in lower case. > > >> Yeah, that did fix it. I'm not sure that I can flash my controller here > > >> since I'm on summer break but it does seem promising. Or maybe I did > > pack > > >> my programmer and can give it a go. The file sizes are the same or > > similar > > >> but there are differences. Granted, I've made changes that may not be > > >> represented in my working project and it may just be that. > > >> Time will tell but it would be great to get rid of the need to use wine > > to > > >> build AmForth here. > > >> Well well well. It appears to have worked. I make install'd the whole > > thing > > >> (since for some reason I did pack my usbasp and could try it out. I'm > > sure > > >>
Re: [Amforth] successful compilation with avra
Thanks for your efforts. People don't often appreciate how much knowledge and effort goes into successful compilation of code. On Fri, Aug 25, 2023, 3:15 PM Erich Wälde wrote: > Hello Brian and Mark, > > very nice to see emails on this list :) > > Compiling amforth with avra? > > I have made numerous experiments a long time ago and again more > recently. If memory serves me well: > - Amforth had been good with avra, at least in the 4.x range. > - However, avrasm2.exe could do more clever tricks, and Matthias > started using those. > - I did make a fork of amForth from Version 5.0, this can be > assembled with avra, see: > https://git.sr.ht/~ew/hbv3_am50forth > - avra received a bit of attention not so long ago (same repo > you found): > https://github.com/Ro5bert/avra > > $ avra --version > > AVRA: advanced AVR macro assembler (version 1.4.2) > which among other changes now includes my favourite atmega644p. > > So, I am currently dabbling with my fork again in the hope to > eventually catch that problem of long term stability. There is > absolutely no reason, why I have to reprogram one or two of my > controllers a few times per year, because they do not start up > after a power cycle, which in turn is done, because the > communication with that controller ceases to work. I went back > to amforth 5.0 for simplicity reasons. > > > All that being said, I would be very interested to see the > changes, maybe, just maybe we can fix the amForth source tree > enough to make avra happy. > > > Cheers, > Erich > > Brian K Navarette writes: > > > That is awesome news! > > Brian-in-ohio > > > > > > On Thu, Aug 24, 2023 at 2:59 PM Mark Roth wrote: > > > >> Hello AmForth. It has been some time and quite weird things since last > I've > >> been here. I am still using AmForth with my trusty atmega1284p and > learning > >> the language as time permits. I remember having heard talk that avra had > >> gotten (almost) to the point of being able to compile the source tree > here. > >> First I tried with 1.3 I think and it failed miserably. Then I found a > repo > >> on github (Ro5bert/avra) that seemed to almost but not quite do it. I > was > >> getting a pile of errors for macro calls. So looking into the issues I > saw > >> that someone had forked that repo and fixed the issue. Something to do > with > >> not having a space between the opening parenthesis and the macro name. > So I > >> tracked down the fix branch (srtlg/avra -b development-issue54 if I > >> remember correctly) and built that locally. Then substituted that avra > >> version with the wine one I had been using to build. > >> It still didn't build. Very almost, but not quite. > >> However, the issue was with errors in /avr8/words/d-lesszero.asm about > the > >> Y register not being declared and/or able to be used for the adiw call. > >> Looking into the source tree I found other usages of y in those calls > but > >> they were all in lower case. > >> Yeah, that did fix it. I'm not sure that I can flash my controller here > >> since I'm on summer break but it does seem promising. Or maybe I did > pack > >> my programmer and can give it a go. The file sizes are the same or > similar > >> but there are differences. Granted, I've made changes that may not be > >> represented in my working project and it may just be that. > >> Time will tell but it would be great to get rid of the need to use wine > to > >> build AmForth here. > >> Well well well. It appears to have worked. I make install'd the whole > thing > >> (since for some reason I did pack my usbasp and could try it out. I'm > sure > >> more testing is needed but this really is pretty cool. > >> > >> ___ > >> Amforth-devel mailing list for http://amforth.sf.net/ > >> Amforth-devel@lists.sourceforge.net > >> https://lists.sourceforge.net/lists/listinfo/amforth-devel > >> > > > > ___ > > Amforth-devel mailing list for http://amforth.sf.net/ > > Amforth-devel@lists.sourceforge.net > > https://lists.sourceforge.net/lists/listinfo/amforth-devel > > > -- > May the Forth be with you ... > > > ___ > Amforth-devel mailing list for http://amforth.sf.net/ > Amforth-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/amforth-devel > ___ Amforth-devel mailing list for http://amforth.sf.net/ Amforth-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amforth-devel
Re: [Amforth] successful compilation with avra
very nice to see emails on this list :) I couldn't agree with you more. :) > Compiling amforth with avra? > > I have made numerous experiments a long time ago and again more > recently. If memory serves me well: > - Amforth had been good with avra, at least in the 4.x range. > What I am talking about is a fork of the Ro3ert/avra that fixes issues with the way the macros are called. It seems it was the fault of avra, not Amforth that was causing the problem. The only problem locally was the 'Y' vs the 'y' it needed to be. The fork I refer to can be found by looking at the issues page of Ro3ert's repo and then getting the fix branch. It's merely a cherry-picked pr. My local Amforth is 6.8 but I think I hand applied all the stuff that would make it effectively 6.9. Pretty sure about that. > - However, avrasm2.exe could do more clever tricks, and Matthias > started using those. > This is interesting. I've not dug deep into the outputted hex files apart from flashing my controller with them. I do need to look at the addresses of both hex files to see if there are issues. Part of the thing is that with avrasm2.exe I was packing my nwwr section to within 5 words of overflow. I want to see (they should, it is all .asm files) if they live at the same locations. I should try to use your git repos for that. I have cloned them locally since I prefer using git than svn. All that being said, I would be very interested to see the > changes, maybe, just maybe we can fix the amForth source tree > enough to make avra happy. > It seems that the 'Y' to 'y' fix and the forked branch does do just that. All the best to the list (and a good finish to the summer) Mark > Brian K Navarette writes: > > > That is awesome news! > > Brian-in-ohio > > > > > > On Thu, Aug 24, 2023 at 2:59 PM Mark Roth wrote: > > > >> Hello AmForth. It has been some time and quite weird things since last > I've > >> been here. I am still using AmForth with my trusty atmega1284p and > learning > >> the language as time permits. I remember having heard talk that avra had > >> gotten (almost) to the point of being able to compile the source tree > here. > >> First I tried with 1.3 I think and it failed miserably. Then I found a > repo > >> on github (Ro5bert/avra) that seemed to almost but not quite do it. I > was > >> getting a pile of errors for macro calls. So looking into the issues I > saw > >> that someone had forked that repo and fixed the issue. Something to do > with > >> not having a space between the opening parenthesis and the macro name. > So I > >> tracked down the fix branch (srtlg/avra -b development-issue54 if I > >> remember correctly) and built that locally. Then substituted that avra > >> version with the wine one I had been using to build. > >> It still didn't build. Very almost, but not quite. > >> However, the issue was with errors in /avr8/words/d-lesszero.asm about > the > >> Y register not being declared and/or able to be used for the adiw call. > >> Looking into the source tree I found other usages of y in those calls > but > >> they were all in lower case. > >> Yeah, that did fix it. I'm not sure that I can flash my controller here > >> since I'm on summer break but it does seem promising. Or maybe I did > pack > >> my programmer and can give it a go. The file sizes are the same or > similar > >> but there are differences. Granted, I've made changes that may not be > >> represented in my working project and it may just be that. > >> Time will tell but it would be great to get rid of the need to use wine > to > >> build AmForth here. > >> Well well well. It appears to have worked. I make install'd the whole > thing > >> (since for some reason I did pack my usbasp and could try it out. I'm > sure > >> more testing is needed but this really is pretty cool. > >> > >> ___ > >> Amforth-devel mailing list for http://amforth.sf.net/ > >> Amforth-devel@lists.sourceforge.net > >> https://lists.sourceforge.net/lists/listinfo/amforth-devel > >> > > > > ___ > > Amforth-devel mailing list for http://amforth.sf.net/ > > Amforth-devel@lists.sourceforge.net > > https://lists.sourceforge.net/lists/listinfo/amforth-devel > > > -- > May the Forth be with you ... > > > ___ > Amforth-devel mailing list for http://amforth.sf.net/ > Amforth-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/amforth-devel > ___ Amforth-devel mailing list for http://amforth.sf.net/ Amforth-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amforth-devel
Re: [Amforth] successful compilation with avra
Hello Brian and Mark, very nice to see emails on this list :) Compiling amforth with avra? I have made numerous experiments a long time ago and again more recently. If memory serves me well: - Amforth had been good with avra, at least in the 4.x range. - However, avrasm2.exe could do more clever tricks, and Matthias started using those. - I did make a fork of amForth from Version 5.0, this can be assembled with avra, see: https://git.sr.ht/~ew/hbv3_am50forth - avra received a bit of attention not so long ago (same repo you found): https://github.com/Ro5bert/avra > $ avra --version > AVRA: advanced AVR macro assembler (version 1.4.2) which among other changes now includes my favourite atmega644p. So, I am currently dabbling with my fork again in the hope to eventually catch that problem of long term stability. There is absolutely no reason, why I have to reprogram one or two of my controllers a few times per year, because they do not start up after a power cycle, which in turn is done, because the communication with that controller ceases to work. I went back to amforth 5.0 for simplicity reasons. All that being said, I would be very interested to see the changes, maybe, just maybe we can fix the amForth source tree enough to make avra happy. Cheers, Erich Brian K Navarette writes: > That is awesome news! > Brian-in-ohio > > > On Thu, Aug 24, 2023 at 2:59 PM Mark Roth wrote: > >> Hello AmForth. It has been some time and quite weird things since last I've >> been here. I am still using AmForth with my trusty atmega1284p and learning >> the language as time permits. I remember having heard talk that avra had >> gotten (almost) to the point of being able to compile the source tree here. >> First I tried with 1.3 I think and it failed miserably. Then I found a repo >> on github (Ro5bert/avra) that seemed to almost but not quite do it. I was >> getting a pile of errors for macro calls. So looking into the issues I saw >> that someone had forked that repo and fixed the issue. Something to do with >> not having a space between the opening parenthesis and the macro name. So I >> tracked down the fix branch (srtlg/avra -b development-issue54 if I >> remember correctly) and built that locally. Then substituted that avra >> version with the wine one I had been using to build. >> It still didn't build. Very almost, but not quite. >> However, the issue was with errors in /avr8/words/d-lesszero.asm about the >> Y register not being declared and/or able to be used for the adiw call. >> Looking into the source tree I found other usages of y in those calls but >> they were all in lower case. >> Yeah, that did fix it. I'm not sure that I can flash my controller here >> since I'm on summer break but it does seem promising. Or maybe I did pack >> my programmer and can give it a go. The file sizes are the same or similar >> but there are differences. Granted, I've made changes that may not be >> represented in my working project and it may just be that. >> Time will tell but it would be great to get rid of the need to use wine to >> build AmForth here. >> Well well well. It appears to have worked. I make install'd the whole thing >> (since for some reason I did pack my usbasp and could try it out. I'm sure >> more testing is needed but this really is pretty cool. >> >> ___ >> Amforth-devel mailing list for http://amforth.sf.net/ >> Amforth-devel@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/amforth-devel >> > > ___ > Amforth-devel mailing list for http://amforth.sf.net/ > Amforth-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/amforth-devel -- May the Forth be with you ... ___ Amforth-devel mailing list for http://amforth.sf.net/ Amforth-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amforth-devel
Re: [Amforth] successful compilation with avra
That is awesome news! Brian-in-ohio On Thu, Aug 24, 2023 at 2:59 PM Mark Roth wrote: > Hello AmForth. It has been some time and quite weird things since last I've > been here. I am still using AmForth with my trusty atmega1284p and learning > the language as time permits. I remember having heard talk that avra had > gotten (almost) to the point of being able to compile the source tree here. > First I tried with 1.3 I think and it failed miserably. Then I found a repo > on github (Ro5bert/avra) that seemed to almost but not quite do it. I was > getting a pile of errors for macro calls. So looking into the issues I saw > that someone had forked that repo and fixed the issue. Something to do with > not having a space between the opening parenthesis and the macro name. So I > tracked down the fix branch (srtlg/avra -b development-issue54 if I > remember correctly) and built that locally. Then substituted that avra > version with the wine one I had been using to build. > It still didn't build. Very almost, but not quite. > However, the issue was with errors in /avr8/words/d-lesszero.asm about the > Y register not being declared and/or able to be used for the adiw call. > Looking into the source tree I found other usages of y in those calls but > they were all in lower case. > Yeah, that did fix it. I'm not sure that I can flash my controller here > since I'm on summer break but it does seem promising. Or maybe I did pack > my programmer and can give it a go. The file sizes are the same or similar > but there are differences. Granted, I've made changes that may not be > represented in my working project and it may just be that. > Time will tell but it would be great to get rid of the need to use wine to > build AmForth here. > Well well well. It appears to have worked. I make install'd the whole thing > (since for some reason I did pack my usbasp and could try it out. I'm sure > more testing is needed but this really is pretty cool. > > ___ > Amforth-devel mailing list for http://amforth.sf.net/ > Amforth-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/amforth-devel > ___ Amforth-devel mailing list for http://amforth.sf.net/ Amforth-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amforth-devel
[Amforth] successful compilation with avra
Hello AmForth. It has been some time and quite weird things since last I've been here. I am still using AmForth with my trusty atmega1284p and learning the language as time permits. I remember having heard talk that avra had gotten (almost) to the point of being able to compile the source tree here. First I tried with 1.3 I think and it failed miserably. Then I found a repo on github (Ro5bert/avra) that seemed to almost but not quite do it. I was getting a pile of errors for macro calls. So looking into the issues I saw that someone had forked that repo and fixed the issue. Something to do with not having a space between the opening parenthesis and the macro name. So I tracked down the fix branch (srtlg/avra -b development-issue54 if I remember correctly) and built that locally. Then substituted that avra version with the wine one I had been using to build. It still didn't build. Very almost, but not quite. However, the issue was with errors in /avr8/words/d-lesszero.asm about the Y register not being declared and/or able to be used for the adiw call. Looking into the source tree I found other usages of y in those calls but they were all in lower case. Yeah, that did fix it. I'm not sure that I can flash my controller here since I'm on summer break but it does seem promising. Or maybe I did pack my programmer and can give it a go. The file sizes are the same or similar but there are differences. Granted, I've made changes that may not be represented in my working project and it may just be that. Time will tell but it would be great to get rid of the need to use wine to build AmForth here. Well well well. It appears to have worked. I make install'd the whole thing (since for some reason I did pack my usbasp and could try it out. I'm sure more testing is needed but this really is pretty cool. ___ Amforth-devel mailing list for http://amforth.sf.net/ Amforth-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/amforth-devel