the setting operation was not allowed, but the default matlab (R2019b) started up fine (no mysterious error box), and then I could not provoke it to appear. I then launched R2020a matlab in the same way, and it also was "cured."
I killed the shell, started a new one, and then started both matlab versions without the preceding statement; still cured.
What is really confusing is how this could've gotten screwed up in real time by running a simple matlab script.
Anyway thanks much. I owe you a beer.
]]>Let's see what we get for "en_US ISO-8859-1" - uncomment the locale, don't forget to run "locale-gen" and try "LC_ALL=en_US.iso88591 matlab"
I suspect that matlab ships a bunch of scripts to interpret the input and at least one of them has a bogus™ encoding.
Does it happen to write a more verbose log somewhere?
]]>localectl
System Locale: LANG=en_US.UTF-8
VC Keymap: n/a
X11 Layout: us
X11 Model: pc105
locale
LANG=en_US.UTF-8
LC_CTYPE="en_US.UTF-8"
LC_NUMERIC=en_US.UTF-8
LC_TIME=en_US.UTF-8
LC_COLLATE="en_US.UTF-8"
LC_MONETARY=en_US.UTF-8
LC_MESSAGES="en_US.UTF-8"
LC_PAPER=en_US.UTF-8
LC_NAME=en_US.UTF-8
LC_ADDRESS=en_US.UTF-8
LC_TELEPHONE=en_US.UTF-8
LC_MEASUREMENT=en_US.UTF-8
LC_IDENTIFICATION=en_US.UTF-8
LC_ALL=
locale -a
C
en_US.utf8
POSIX
Since we figured that the locale has an impact on this and matlab doesn't like the C locale - what's the output of the previsouly requested
localectl
locale
locale -a
(ie. your normal locale settings/condition)
]]>You would seem to be onto something though; if I might impose, please keep the suggestions coming...
]]>localectl
locale
locale -a
LC_ALL=C matlab # this makes matlab run in a 7bit locale
if I do something which causes
So it doesn't happen right with the startup, but requires a trigger?
What triggers it? Any pattern notable?
Here is a summary of the output of pacman -Qkk:
[djg-lt ~]# pacman -Qkk > check_output.txt
[djg-lt ~]# cat check_output.txt | wc -l
2454
[djg-lt ~]# cat check_output.txt | grep "0 altered" | wc -l
2362
[djg-lt ~]# cat check_output.txt | grep mismatch | wc -l
78
[djg-lt ~]# cat check_output.txt | grep "1 altered" | wc -l
8
[djg-lt ~]# cat check_output.txt | grep "1 altered"
audit: 155 total files, 1 altered file
gdm: 498 total files, 1 altered file
ghc-libs: 1412 total files, 1 altered file
lighttpd: 105 total files, 1 altered file
nfs-utils: 106 total files, 1 altered file
systemd: 1818 total files, 1 altered file
update-grub: 3 total files, 1 altered file
vlc: 1057 total files, 1 altered file
[djg-lt ~]# cat check_output.txt | grep "2 altered"
java-runtime-common: 21 total files, 2 altered files
shadow: 558 total files, 2 altered files
trousers: 138 total files, 2 altered files
[djg-lt ~]# cat check_output.txt | grep "3 altered"
cups: 874 total files, 3 altered files
[djg-lt ~]# cat check_output.txt | grep "4 altered"
openca-base: 1600 total files, 4 altered files
[djg-lt ~]# cat check_output.txt | grep "5 altered"
[djg-lt ~]#
pacman -Qkk
tests the integrity of the installation, but from the earlier link, it really seems as if matlab just doesn't like some kinds of file/directory names when it triggers that warning.
]]>>% strace matlab
yields thousands of lines of output. There are many -1 ENOENT type entries. Tens or hundreds of libraries named.
But as I said, I also did a fresh install of R2020a and immediately observed the same behavior.
man strace shows it is fairly versatile. Do you have any suggestions as to how best to use it?
Is it even possible under current arch that, somehow, while using matlab 2019b, some lib files, or portions thereof, got erased or overwritten or somehow messed up? These would have to be files that do not come from matlab, and were also not "fixed" during the recent pacman -Syu, nor worked around by booting into lts.
]]>I'd strace it to see whether it crosses any non-7bit paths or failes ENOENT on some file you accidentally might have removed.
]]>The libraries that come with matlab have a host of problems with undefined symbols, but all of those are fixed by comprehensively relinking to the system lib files.
even after I remove all of those problems, and fix the hardware opengl, and implement every other fix that seems to be out there on the wiki and general internet, I still cannot stop the pop-up error message "An unexpected error has occurred" from appearing, and as long as it appears, some functionality within the program is broken.
I got to the point where matlab launches with no errors echoed to the terminal, executes commands with no errors echoed to the terminal, but still, the pop-up error message "An unexpected error has occurred" keeps popping up, and if it does, some functionality is broken.
I think I have to give up for the moment, since the program is proprietary and is giving no further clues as to what's wrong under the hood.
]]>