General Category > LisaList2
LisaEm 1.2.7-RC4 support bug reports
rayarachelian:
Released RC3:
* https://lisaem.sunder.net/downloads/LisaEm_1.2.7-RC3_2020.08.21-macosx.dmgSkinny packages:
* https://lisaem.sunder.net/downloads/lisaem-1.2.7-RC3-2020.08.21-macos-10.5-i386.pkg
* https://lisaem.sunder.net/downloads/lisaem-1.2.7-RC3-2020.08.21-macos-10.5-ppc.pkg
* https://lisaem.sunder.net/downloads/lisaem-1.2.7-RC3-2020.08.21-macos-10.7-x86_64.pkg
* https://lisaem.sunder.net/downloads/lisaem-1.2.7-RC3-2020.08.21-macos-10.8-x86_64.pkg
* https://lisaem.sunder.net/downloads/lisaem-1.2.7-RC3-2020.08.21-macos-10.10-x86_64.pkg
* https://lisaem.sunder.net/downloads/lisaem-1.2.7-RC3-2020.08.21-macos-10.11-x86_64.pkg
* https://lisaem.sunder.net/downloads/lisaem-1.2.7-RC3-2020.08.21-macos-10.12-x86_64.pkg
* https://lisaem.sunder.net/downloads/lisaem-1.2.7-RC3-2020.08.21-macos-10.15-x86_64.pkgSource:
* https://lisaem.sunder.net/downloads/lisaem-1.2.7-RC3-2020.08.21-src.tar.xz
* and github as usual: https://github.com/rayarachelian/lisaem/[Rather than only posting at the end of the thread for release notices, I'll update the top message as well to make it easy to find the latest release. Bug reports/replies will go to the end of the thread as usual, this way I'll avoid polluting the rest of LisaList2 with bug reports/release reports.]
D.Finni:
Here's another crash for you. This happens on 10.12.6.
ROMless Boot -> Choose ProFile Hard Drive -> click OK
Hard Drive Size? -> Click Cancel -> crash
This sequence however does not cause a crash when using a Lisa ROM file.
Text file attached below.
rayarachelian:
--- Quote from: D.Finni on August 23, 2020, 05:14:43 pm ---Here's another crash for you. This happens on 10.12.6.
ROMless Boot -> Choose ProFile Hard Drive -> click OK
Hard Drive Size? -> Click Cancel -> crash
This sequence however does not cause a crash when using a Lisa ROM file.
Text file attached below.
--- End quote ---
Great, thanks for finding this, thank you!
rayarachelian:
Released RC3a:
https://lisaem.sunder.net/downloads/LisaEm_1.2.7-RC3a_2020.08.24-macosx.dmg
Skinny packages:
https://lisaem.sunder.net/downloads/lisaem-1.2.7-RC3a-2020.08.24-macos-10.10-x86_64.pkg
https://lisaem.sunder.net/downloads/lisaem-1.2.7-RC3a-2020.08.24-macos-10.11-x86_64.pkg
https://lisaem.sunder.net/downloads/lisaem-1.2.7-RC3a-2020.08.24-macos-10.12-x86_64.pkg
https://lisaem.sunder.net/downloads/lisaem-1.2.7-RC3a-2020.08.24-macos-10.15-x86_64.pkg
https://lisaem.sunder.net/downloads/lisaem-1.2.7-RC3a-2020.08.24-macos-10.5-i386.pkg
https://lisaem.sunder.net/downloads/lisaem-1.2.7-RC3a-2020.08.24-macos-10.5-ppc.pkg
https://lisaem.sunder.net/downloads/lisaem-1.2.7-RC3a-2020.08.24-macos-10.7-x86_64.pkg
https://lisaem.sunder.net/downloads/lisaem-1.2.7-RC3a-2020.08.24-macos-10.8-x86_64.pkg
Windows 10:
https://lisaem.sunder.net/downloads/LisaEm-1.2.7-RC3a-2020.08.24-win64.zip
Source:
https://lisaem.sunder.net/downloads/lisaem-1.2.7-RC3a-2020.08.24-src.tar.xz
and github as usual: https://github.com/rayarachelian/lisaem/
rayarachelian:
So been having fun implementing pty support for the serial ports. Some weird things:
0. Serial port A hangs LOS. I haven't been able to figure out exactly why, but regression testing shows this has been the case since at least LOS 1.0 beta. I've also found a bug in the fliflo (fifo+lifo library) that broke some serial access which I've fixed.
1. LisaTerm (or perhaps the OS itself) will chop off output after X characters of input, that is it will accept the first ~64 chars of input, and then discard the rest until it stops coming in and then accept more later. I haven't worked out the exact timing of these, so working on a bit of code, where if LOS is the active OS (I don't yet have a way to detect just LisaTerminal), it will slow the input down. I don't know how Xenix/UniPlus will behave, will revisit that once those OS's boot.
I tested this by running seq 100 200 - the output stops at "112" right before outputting the "2" character. Why 100-200 instead of 1 1000? Because this way, the output is exactly 5 bytes per line. The 3 number chars + cr+ new line. So this adds up to roughly 60 chars.
On another attempt I tried to output a string of text that was around 80 chars via echo and it stopped exactly at the 64th one. So that seems to be the size of the buffer. Changing the baud rate up doesn't affect this at all.
But wait! It gets worse!
2. ANSI color codes confuse it. Modern linux/macos implementations insist on sending color, so setting TERM=vt100 is not enough. I can limit colors for things like ls and grep upto a point by setting environment variables before the execve bash, but not for everything.
As an example, ls -l itself will bold the output of some filenames right after the date/size, and LisaTerm discards the output immediately before the file name. Ditto for colorized grep.
3. Forget about using vi, it's horribly broken, the vt100 implementation in LisaTerm doesn't allow it to work properly. It can sort of work, but not quite. ed, or jed, haven't tried emacs, are your best bet.
4. I'm considering writing a small ANSI/xterm/vt100 escape sequence filter command line tool to remove color sequences and potentially any other things that break LisaTerminal by causing it to abort outputting text. It would have to do it's own fork/exec pty allocation. This way it can sit outside of LisaEm and be invoked by the user via preferences when pty is used.
I can build this into to LisaEm, but perhaps it might be more useful as a standalone unix cli filter tool.
I suppose this is analogous to ancient web browsers not understanding SSL, or modern HTML 1.x and requiring a filtering proxy in the form of a Raspberry Pi that connects to the outside world over wifi, and then presents TCP/IP traffic over SLIP/PPP for things like the MacPlus/SE or even MacWorks. :)
5. Using the default cooked output causes the pty to freeze, so the fix is to switch to raw before doing the execve, but then, the terminal is all messed up and I have to run stty sane to get echoing and other things to behave.
I'm adding back some of the options one by one of cooked/sane mode to the code to figure out which tcsetattr option is the culprit, but there's dozens of them.
I also seemed to have broken a couple of things in this effort:
1. BOOT ROM now throws an I/O error, so some of the z8530 behavior got messed up, and I'll need to fix that.
2. Mouse behavior in LOS, but not the BOOT ROM behaves as if everything is dragging, it's as if the mouse up event isn't being received even though it's sent and accepted by LOS, not sure what's going on there.
3. Still have XproFile/BLU/dc42 work to complete before RC4's release.
I think after I'm done with LisaEm 2.0, I'll work on a replacement LisaTerm program... one with XModem as well.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version