Speech Sound driver
Didn't you read https://sfconservancy.org/GiveUpGitHub/ ? I don't think NitrOS-9 is a flag-ship free-software project, or has to follow FSF, but I do sympathize with many of the points :) Anyway, there are two things here. 1) Converting to git - I am all for that. But only after next NitrOS-9 release. I want to check and possibly do the conversion myself, I do care about history preservation and attribution. 2) Then where to host, possibly move from SF. It can also be multiple places, git is decentralized....
Tormod, It's time to move this project to GitHub. Here are my big two reasons: Visiblity - all of the momentum in open source projects is there. I think the NitrOS-9 project officially being hosted on GitHub would increase its prominence. Collaboration - GitHub has a great interface for managing for pull requests which are great for change management. It's easy to see what has changed through the web interface. Viewing diffs is elegant and clear, as is commenting on those changes. There's already...
It is a long-lived project, and it is not unusual that it is quiet some months (or years!) at a time. It can also be considered mature, and more maintained than further developed. Personally I have been hoping to get to make a new release for a while, and merge in some changes from the Ease of Use project, but I haven't had much time for retro-computing over the last years. Still hoping to get to this "soon". This would also include some release management coordination with ToolShed. Otherwise, if...
Is this project dead? I have not seen any activity in a LONG time.
https://nitros9.sourceforge.io/wiki/index.php/Building_NitrOS9 It says "prebuilt" but that means binaries. Unix/etc allow spaces in filenames, but many tools are not so much at ease with it. If you do shell scripting, adding quotation marks all over (except where you shouldn't) works. If you do Makefiles, it is well, complicated.
On Jul 3, 2022, at 4:20 PM, Tormod Volden wrote: anyone with a modern Window PC would have defaulted to OneDrive Good grief, I hope you are not right, and I think you aren't. I’ve reinstalled my Work DELL laptop so many times — it’s been back in the shop 6 or 7 times, and each time I get it back I do a Restart and fresh install to make sure nothing got “put” on my machine. And if you just go through all the defaults, it wants you to specify a Microsoft login, and it sets up OneDrive, etc. by default....
On Jul 3, 2022, at 4:20 PM, Tormod Volden wrote: anyone with a modern Window PC would have defaulted to OneDrive Good grief, I hope you are not right, and I think you aren't. I’ve reinstalled my Work DELL laptop so many times — it’s been back in the shop 6 or 7 times, and each time I get it back I do a Restart and fresh install to make sure nothing got “put” on my machine. And if you just go through all the defaults, it wants you to specify a Microsoft login, and it sets up OneDrive, etc. by default....
On Jul 3, 2022, at 4:20 PM, Tormod Volden tormod@users.sourceforge.net wrote: anyone with a modern Window PC would have defaulted to OneDrive Good grief, I hope you are not right, and I think you aren't. I’ve reinstalled my Work DELL laptop so many times — it’s been back in the shop 6 or 7 times, and each time I get it back I do a Restart and fresh install to make sure nothing got “put” on my machine. And if you just go through all the defaults, it wants you to specify a Microsoft login, and it sets...
anyone with a modern Window PC would have defaulted to OneDrive Good grief, I hope you are not right, and I think you aren't. One could say that anyone with a Windows PC usually doesn't build his own software, and that's why we are providing release and snapshot binaries :) And if it can be fixed simply by adding quotes around things in makefiles, why don't we just do that and solve it for everyone? I think you tried that 2 years ago, and I think it is not so easy. GNU Make doesn't deal well with...
Is this documented? As it stands today, anyone with a modern Window PC would have defaulted to OneDrive, and would be unable to build this. Likewise, anyone with a modern Mac, with iCloud on by default, would be unable to build this. At the very least, this needs to be documented. I consider myself a far more advanced user than most, and it's bitten me on both operating systems. And if it can be fixed simply by adding quotes around things in makefiles, why don't we just do that and solve it for everyone?...
Go to a folder with no spaces in it, i.e. /Users/allen then create a soft link there: cd /Users/allen ln -s /Volumes/WhatNot/With Spaces/etc/nitros-9 cd nitros-9 make
Go to a folder with no spaces in it, i.e. /Users/allen then create a soft link there: cd ln -s /Volumes/WhatNot/With Spaces/etc/nitros-9 cd nitros-9 make
This hit me again tonight when I started trying to get NitrOS9 going ;-) It's the first time I've tried to build from an iCloud drive on Mac, and it has the same issue any Windows system with OneDrive has -- space in the path name, darnit.
Oh, I just noticed that with your change, if the specified PORT does not exist, you don't really get an obvious "File not Found" error. It does print a bunch out like it is doing something (I guess just assmbling "NOSLIB"). You just get: make PORTS=fake_port [...] make[1]: Nothing to be done for `all'. Before, it would show: make PORTS=fake_port [...] make: *** fake_port: No such file or directory. Stop. make[1]: *** [all] Error 2 make: *** [all] Error 2 I am not sure that is a big deal, though,...
Pushed.
Dynamic PORT name for level 2 builds
makefile: Allow PORTS with level1/2 mix
Even better! Works great for me even when specifiying multiple PORTS and mixing levels. Thank you!
Even better! Works great for me specifiying multiple PORTS and mixing levels. Thank you!
Nice! That is a good idea. Taking this a bit further, what do you think of this? It allows mixing level1 and level2 ports. As a side-effect it bypasses level1/makefile and level2/makefile but that is fine. diff --git a/makefile b/makefile --- a/makefile +++ b/makefile @@ -9,18 +9,9 @@ include rules.mak dirs = $(NOSLIB) $(LEVEL1) $(LEVEL2) $(LEVEL3) $(3RDPARTY) # Allow the user to specify a selection of ports to build -# All selected ports must be of the same level ifdef PORTS -dirs = $(NOSLIB) -ifneq...
Dynamic PORT name for level 2 builds
level 2 pmap assembly error with 4K block DAT
Fixed in the repo. Thanks for the report!
pmap: Fix building for 4K blocks DAT
Yes, it will assemble after just that line is fixed.
Ah, the printed header string. I get it. Is it only line 104 that needs fixing?
Twice as many blocks to show need a longer heading/header string, I guess?
Twice as many blocks to show need a longer header, I guess?
The "force direct" change was probably added to save a byte here and there and is not strictly necessary. I guess the same binary code is generated for CoCo3 because the assembler silently optimizes it to direct mode (see the LWASM forwardrefmax pragma). Why is the header string longer for the 4K block DAT?
level 2 pmap assembly error with 4K block DAT
rules.mak: Make os9 command settable
level1 makefiles: Separate basic09 module lists
Moved the emudsk driver into level1/modules.
Add fixed emudsk driver from EJJaquay. Add emudsk operating system disk image.
Toromod Volen, Thank you for your response. What about basic structure and layout of the OS/APIs/System calls? Once again, I am not going to use any of the source code, but OS9's overall structure was/is elegant and compact. That's not to state that it couldn't stand to be modernized for current hardware, and perhaps security could be improved in some fashioned. Do you have any other suggestions about how to best protect such an Open-Source project? Or which traps to avoid? Thank you, --Brenda
The older question of whether the NitrOS-9 source code (or polished disassembly in many cases) is open-source (and under what kind of license) is difficult, but your question is easier to answer, I think. If you just build it on the structure as seen from the outside, and use the same API names, it will be fine. For instance is WINE implementing the Win32 API without legal issues, and is used by well-known companies. ReactOS is building an OS compatible with Windows 5+. You would be in the same situation,...
Hi, I have seen differing answers on this.... Is NitroOS9's clean enough for it to be a full-fledged free unencumbered open-source project? In other words, I would like to take NitroOS9's structure and system call names and try creating a newer operating system from it to run on other hardware, using original code, likely in Arduino-flavored C. Is this legal, or is it that if myself and others were successful, then could the legal Sward of Damocles come down should the project come into Microware's...
mismatched case
Fix makefile mismatched filenames
mismatched case
Adding Dragon Plus builds with native 80 column and VRAM disk support
CRC in modules is not verified
rules.mak: Whitespace fixes
display: Fixed bug with '0' characters in decimal conversion
Thank you.
Change level2/coco3/defs/Defsfile to use coco.d instead of coco3.d
Looks correct. It seems that 8 years ago in commit 2624:b8c7b7fbf3c9 the coco defs were put into a new "coco.d" (from "systype"), and the various level*/<port>/defsfile were updated. However, the level*/<port>/defs/Defsfile (that are copied to the disk images under DEFS) were apparently wrongly updated. I added fixups for the other Defsfile's too.
Correct coco.d filename in shipped Defsfile files
Change level2/coco3/defs/Defsfile to use coco.d instead of coco3.d
EOU
EOU
Large file deletion fails in exciting ways
Updated source L. Curtis Boyle used back during the day.
I think it has to quote the whole thing, like if it did: $(NITROS9DIR)/make That whole thing needs quoted: "$(NITROS9DIR)/make" I will do some tests. As to hard-coded path versus relative, going all relative would solve this as well, since as long as the make is ran from the initial directory, everything would be controlled below it and the upper install path (spaces or no) would not matter.
I don't mind doing the "trained monkey" work for massive cleanups like this. I tend to be the one that takes those on at companies I work for because no one else will do it and I know it pays off in the end ;-)
This would be a quick way to just add those quotes: sed -i 's/\$(NITROS9DIR)/"$(NITROS9DIR)"/g' $(find . -name makefile) but frankly I am unsure if that would be enough. "make" simply doesn't deal well with spaces. However we have a mix of using NITROS9DIR and using relative path e.g. ../../../rules.mak that it would be nice to clean up. I think relative paths are good, and they are already used in half of the places for finding rules.mak. Apart from rules.mak, NITROS9DIR is only used a few plac...
If the work is just quoting all the references, I could knock that out one evening. Assuming that works.
Oh, that will be some work :) Take back control! But actually there are workarounds too: On Linux and MacOS and even on Windows you can use a soft link to the repo in a place without spaces in its path.
makefiles fail is NITROS9DIR path has spaces in it.
DACIA Driver Missing
Update ChangeLog with features and fixes
level2/3: Generate motd in makefile so LEVEL is right
mc09l2: Don't try to include non-built CCB (CoCoBoot) kernel
level3/makefile: Avoid copying same files twice to disk image
level2: Set motd attributes correctly on disk images
defs: Add Bt.Sec for atari and corsham
d64: Ship bootlist suitable for Dragon DS 40-track floppy
Dragon ports: Replace mb scripts
d64: Add mb.ddisk script for making Dragon boot floppy
os9gen: Fix bitmap check/update where boot sector is non-zero
level1/d64: Include Dragon disk specific commands on disk image
format.asm: Correctly accept lower-case (Dragon disk format) option
scripts/os9_gdb: Show CRC in module directory list
Updated format.asm to include the option for the enchanced 20 sector per track sector layout.
Added information about what was changed in last commit.
I have only done minimal testing so far, but those updated utilities seem to be working on my syste, with the Mini MPI installed. Thanks!
Mark, I can't test sdrive as I don't have a MiniMPI. However, if you obtain http://www.colorcomputerarchive.com/coco/Disks/Utilities/Coco%20SDC%20Utilities%20%28Barry%20Nelson%29%20%28OS-9%29.zip you will find an updated source for sdrive and the program. More to the point, you will find the program SDC which works with a stock MPI and if it works with the MiniMPI should be much better for you. Robert Gault
sdrive fails with MPI
Hello Call for Participation for the Retrocomputing DevRoom at FOSDEM 2020 ends in 5 days. Hurry up if you want to talk at the largest open source event in Europe! ---------- Forwarded message --------- From: Pau Garcia Quiles pgquiles@elpauer.org Date: Sun, Oct 27, 2019 at 11:48 PM Subject: FOSDEM 2020 - Retrocomputing DevRoom Call for Participation To: retrocomputing-devroom@lists.fosdem.org Hello, FOSDEM is a free software event that offers open source communities a place to meet, share ideas...
Hello, FOSDEM is a free software event that offers open source communities a place to meet, share ideas and collaborate. It is renowned for being highly developer-oriented and brings together 8000+ participants from all over the world. It is held in the city of Brussels (Belgium). FOSDEM 2020 will take place during the weekend of February 1st-2nd 2020. More details about the event can be found at http://www.fosdem.org. CALL FOR PARTICIPATION After success in the past two years, the Retrocomputing...
Updated stack copy routines for level2 krn module for 6809 provided by L. Curtis Boyle.
Get gfx2 to build from cmds folders
NOTE: Edition 4 now includes very slight optimizations, and two more drawing
Build errors with pulu/pshu u errors
Neil, your repo must be out of date. This was fixed in commit 3080:db58cdca8a7f almost a year ago. Please try: hg pull hg update default
Build errors with pulu/pshu u errors
RAMMER Edition #5 Revision 3 updates:
ASM Edition 11. Adds -o option (automatically overwrites object file), and is 7-13% faster (both on 6809 and 6309).
Updated thexs.asm to correct for 1MB and 2MB systems so that Thexder will run correctly.
Disassembled Shanghai and verified the new binary is exactly the same.
cowprs: Add register descriptions from PBJ WORD-PAK manual
co80: Use HW port address from descriptor (via SCF and VTIO)
co80: Better description in source header
defs/cocovtio.d: Add ModCo80 COLoad bit mask for Co80 VDIO submodule
level3: nitro module: Use KrnBlk instead of hardcoded $3F
scripts/os9.gdb: Use module directory start/end vectors
level3: bootfiles: Move device descriptors out of RBF block
cowprs: Note address from PBJ WORD-PAK/WORDPAK II User's Manual