antiX-17.rc1 full versions available for public testing - Page 2

BitJam · posted: 2017-10-08 02:56

Summary of issues reported above

And of issue reported in the antiX-17-rc1 thread started by Shay. Issues I'm addressing are in dark red and my replies are in dark green. I hope other devs can handle the issues that I didn't color. I don't cover skidoo's very recent snapshot and live-usb-maker bug reports. If I missed an issue you reported, please report it again. All of this feedback has been extremely helpful.

@skidoo

========= SCRAPER REMOVED AN EMBEDDED LINK HERE ===========
url was"antix-17-rc1_x64-full-heather-heyer-6-october-2017-t7223.html#p53573"
linktext was:"antix-17-rc1_x64-full-heather-heyer-6-o ... tml#p53573"
====================================

issues with workspaces,
jwm theme selection problem
clipit overshadowed by [some other windows?]
broken links in jwm desktop menu Help --> JWM -->"WM Puppy Linux"
broken links in fluxbox desktop menu:
h t t p ://fluxbox-wiki.org/index.php/Faqs
h t t p ://fluxbox-wiki.org/index.php/Fluxbox-wiki

human-readable, non-COLORized version of some live logfiles are absent
What program is this a problem in? IMO we should add LESS="-R" to the default .bashrc which fixes this in less.

v---- is this info still accurate ?
No. I'm trying to maintain a more recent version at /etc/live/README. It is plain text. Perhaps I should use mark-down. I'm not going to keep all of the documentation up to date this go-round.

o preclude maintaining an accurate list within the docs, could just state list of services that are disabled is in the file /etc/live/config/disabled
But that is not exactly true. Those are the services that were disabled. There are other services we will disable if they are installed. The most accurate lists are in the file /live/etc/init.d/live-disable-services

firejail is pre-installed, but...
no menu entry b/c the front-end package (127kb)"firetools" is absent

longstanding repaint bug rxvt.

Well, adding xorg=x didn't break anything (is that good? bad? I haven't noticed that it has any effect
A /etc/X11/xorg.conf file should show up. If this works for most systems then in the next version of antiX we may try to make this the default.

During live boot, subsequent to having chosen via kernel VGA menu...
no, not prompted to save (an observation, in response to your question; I'm not reporting this as a bug).
There are a number of things that will prevent us from asking if you want to save:
Did you already select"save" or"ask+save"?
Did you use vga=ask"?
Did a non-standard vga code cause you to get to the VGA menu?
Is this not a live-usb?

Same as with previous ax17 betas, 3 seconds or so into init... as soon as KMS kicks in, the console huge fontface is shown.
I really want to fix this but I will need your help and cooperation. Can we do this via MX PM or email?

choice made via console-font-select is not sticky ~~ not across boots, not even across consoles.
Correct. The choice made via console-width-select should span consoles and should be sticky in installed systems (if the Debian sysv gods ae not upset with us) but not live. I think you can still make console fonts stick on the live system if you edit the file /etc/default/console-setup and follow the instructions they give you for Braille. It's been tricky to get it to work right automatically and also work with Debian's existing sysv init code, some of which got broken right before they switched to systemd.

ControlCenter --} Live --}"Edit Exclude files"
D_O: Controlcenter->Live->edit boot loader menu also opens blank.
I will try to fix these. Another others?

[live-remaster]: same quirks as before...
Problems with the cancel button
Are there other problems in addition to the cancel button and the version identification (below)?

remaster version identification
It show up in the file /etc/live/version/linuxfs.ver in the remastered system. This file keeps a history of your remasters.

@dolphin_oracle
Spacefm desktop on live does not have an installer icon on desktop. It is present on the menu.

Remaster options selection screen runs off the bottom of my 1366x768 monitor (see screenshot).
Thanks! Someone else reported a similar problem with persist-save. I think something may have changed with an upgrade to yad. I will look into this.

@eugen-b
@BitJam, to properly quote here on the new forum, you need to switch to"View Source" mode in the edit window (three dots or Ctrl+Shift+s).
Thanks! Very useful.

automount option at live boot menu gives about ten lines with the same error message that it failed to mount sdb1 which is an SD card, but the card is mounted when I log in.
Maybe it is because it is formatted exfat? I'll upload the error messages...
The error messages you posted start with"Filtered devices /dev/sdb1 /dev/sdb" so those seem to be from the early boot process when we are trying to find the live-usb/cd device. I exposed those error messages to debug the ntfs frugal stuff. I left them exposed but I see that was a mistake so I will hide them again. Thanks.

You might also be getting error messages from the mount=<all|usb> cheat but those will occur later and will be in the (colorized) log file /var/log/live/live-initrd.log. You can view the colors correctly with"less -R /var/log/live/live-init.log". You can also pause the boot process before all those messages go away with the cheat"bp=b9" which will give you a Bash shell when the initrd is almost done. I will get rid of the error messages you posted. If there are others, later on, then I want to see them.

BitJam · posted: 2017-10-08 03:03

skidoo wrote: snapshot fails on a machine which has (only!) 2Gb RAM.
Tested on fresh rc1 system (no additional packages installed).
Except choice"personal" (default is"general"), no settings were changed away from the defaults within isosnaphot.

No warnings are issued ahead of time. It churns the CPUs for 20min... then pukes with the screencapped error message dialog.

This is not my code but I think the problem is that it can be very time consuming to estimate the size of the resulting linuxfs file and iso file on an installed system. The time is usually reasonable on the live system which is why I can perform the size check in live-remaster. We might be able to add in size checks if we detect you are on a live system or maybe we could add it as an option. You are going to need roughly twice the size of the iso file we end up creating because first we have to make the linuxfs file and then we have to make the iso file which contains the linuxfs file.

I don't have a great solution.

BitJam · posted: 2017-10-08 03:23

skidoo wrote: live-usb-maker
successfully tested: copy running system to pendrive (chose first numbered item in the menu, aka"full-featured")

quirk:
Upon first boot of the created pendrive, F5 root_persist (along with other details of the source system) is already pre-selected.
IIRC, even the second ("custom") boot menu line is pre-selected ~~ can just press Enter to boot.

If you clone a running live system then we copy the bootloader settings from the running live system. I agree, this may not be ideal. There will be a similar situation on the UEFI side.

The system recognizes that persistence has been requested and that rootfs does not yet exist, so prompts to create the rootfs...
...but does not provide the expected opportunity to choose semi-automatic save mode.

Interesting. Maybe the persistence config file was not excluded during either a remaster or an iso-snapshot. That would account for not being asked what save mode you want.

Were you prompted to change passwords? That should only happen if you are using the default passwords"demo" and"root".

BitJam

I think I fixed the problem with edit excludes in the control centre. I couldn't fix the problem with editing the bootloader config. Maybe someone has an idea on how to get that to work. Details and the latest antixcc-03.sh are available here:
========= SCRAPER REMOVED AN EMBEDDED LINK HERE ===========
url was:"(viewtopi c.p hp)?f=16&t=7225&p=53620#p53620"
linktext was:"(viewtopi c.p hp)?f=16&t=7225&p=53620#p53620"
====================================

SamK · posted: 2017-10-08 07:56

eugen-b wrote: A strange behaviour when you change the WM from the main menu: Evince gets closed, IceWM gets"Alpha" and"Beta" as workspace labels.

This was also reported when testing b1.

========= SCRAPER REMOVED AN EMBEDDED LINK HERE ===========
url was:"(viewtopi c.p hp)?f=15&t=7049&p=52195&hilit=workspace#p52195"
linktext was:"(viewtopi c.p hp)?f=15&t=7049&p=52195&hilit ... ace#p52195"
====================================

After the switch to tapatalk, I can't find a way to directly show the relevant post, so refer to post number #96 in the link.

anticapitalista · posted: 2017-10-08 11:47

@skidoo
antix-17-rc1_x64-full-heather-heyer-6-o ... tml#p53573
issues with workspaces, FLUXBOX WORKSPACES NOW 2 - FIXED
jwm theme selection problem WON'T FIX since a fix I know spams .xsession file
clipit overshadowed by [some other windows?] DON'T KNOW A FIX
broken links in jwm desktop menu Help --> JWM -->"WM Puppy Linux" FIXED
broken links in fluxbox desktop menu: FIXED
h t t p ://fluxbox-wiki.org/index.php/Faqs
h t t p ://fluxbox-wiki.org/index.php/Fluxbox-wiki

firejail is pre-installed, but...
no menu entry b/c the front-end package (127kb)"firetools" is absent FIXED

@dolphin_oracle

Spacefm desktop on live does not have an installer icon on desktop. It is present on the menu. FEATURE!

Expect a load of updated debs to hit the repos soon. (BTW Upstream Debian also upgraded dbus, samba which broke our noystemd set up. I spent yesterday recompiling them for antiX)

dolphin_oracle · posted: 2017-10-08 12:03

BitJam wrote: I think I fixed the problem with edit excludes in the control centre. I couldn't fix the problem with editing the bootloader config. Maybe someone has an idea on how to get that to work. Details and the latest antixcc-03.sh are available here:
========= SCRAPER REMOVED AN EMBEDDED LINK HERE ===========
url was:"(viewtopi c.p hp)?f=16&t=7225&p=53620#p53620"
linktext was:"(viewtopi c.p hp)?f=16&t=7225&p=53620#p53620"
====================================

odd, because you can use"gksu geany ........" from the terminal just fine.

even odder, specifying"geany" instead of $EDITOR in the code also works, so to me that implies something screwy with the text expansion rather than the path problem. I don't know if that's easier or harder to fix.

**edit*** yep, try this kind of entry.

Code: Select all

"gksu '$EDITOR /live/boot-dev/boot/syslinux/syslinux.cfg' &" \

why the other way works for the excludes list files and not the bootloader I have no idea. perhaps some bug in gtk-dialog that is skipped because the paths are expanded in the excludes entry but not in the bootloader entry. or maybe do to the path permissions as bitjam as postulated. hard to tell.

dolphin_oracle · posted: 2017-10-08 15:26

SamK wrote:
eugen-b wrote: A strange behaviour when you change the WM from the main menu: Evince gets closed, IceWM gets"Alpha" and"Beta" as workspace labels.
This was also reported when testing b1.

========= SCRAPER REMOVED AN EMBEDDED LINK HERE ===========
url was:"(viewtopi c.p hp)?f=15&t=7049&p=52195&hilit=workspace#p52195"
linktext was:"(viewtopi c.p hp)?f=15&t=7049&p=52195&hilit ... ace#p52195"
====================================

After the switch to tapatalk, I can't find a way to directly show the relevant post, so refer to post number #96 in the link.

the Naming of workspaces in icewm appears to only actually function from a login from slim. Just restarting icewm from the"Restart Icewm" menu entry does not rename the workspaces per the preferences file, so you get stuck with what whatever the workspaces were named by the other WM's (alpha and beta for Fluxbox, 1 and 2 for jwm).

BitJam · posted: 2017-10-08 16:58

dolphin_oracle wrote:even odder, specifying"geany" instead of $EDITOR in the code also works, so to me that implies something screwy with the text expansion rather than the path problem. I don't know if that's easier or harder to fix.

**edit*** yep, try this kind of entry.

Brilliant! Thank you!

BTW: starting with version 03 you can use this to check the expansions:

Code: Select all

antixcc.sh --debug > some-file

SamK · posted: 2017-10-08 17:41

dolphin_oracle wrote: the Naming of workspaces in icewm appears to only actually function from a login from slim. Just restarting icewm from the"Restart Icewm" menu entry does not rename the workspaces per the preferences file, so you get stuck with what whatever the workspaces were named by the other WM's (alpha and beta for Fluxbox, 1 and 2 for jwm).

Just retested this in antiX-16, installed and live environments. Both work as expected i.e. IceWM->Fluxbox->IceWM, the workspace names show as 1&2, Alpha&Beta, 1&2. respectively, so it seems limited to 17.

dolphin_oracle · posted: 2017-10-08 17:48

SamK wrote:
dolphin_oracle wrote: the Naming of workspaces in icewm appears to only actually function from a login from slim. Just restarting icewm from the"Restart Icewm" menu entry does not rename the workspaces per the preferences file, so you get stuck with what whatever the workspaces were named by the other WM's (alpha and beta for Fluxbox, 1 and 2 for jwm).
Just retested this in antiX-16, installed and live environments. Both work as expected i.e. IceWM->Fluxbox->IceWM, the workspace names show as 1&2, Alpha&Beta, 1&2. respectively, so it seems limited to 17.

I think I agree with you...there is a bug in icewm which as far as I can tell, when it restarts either by the"restart icewm" command and the desktop-switcher command, some of the configuration files are not processed, which seems to happen fine when logging in from slim...with the possible exception of the dpi selection. flux and jwm do not share this problem AFAICT.

dolphin_oracle · posted: 2017-10-08 17:51

I'm attaching a log file for cli-aptiX. The initial request for an apt-get update goes into an endless loop unless you just get tired and say"N" at some point. Looking at the log file, it looks to me like maybe the counts for the list files aren't going as expected.

I'm sure this worked before...I really like cli-aptiX. The update from the regular menu entry seems to work without that error (which I think doesn't do the check maybe).

skidoo · posted: 2017-10-08 17:55

I reached the same conclusion (problem related to shell text expansion).
Also, $EDITOR is preferable to"desktop-defaults-run -te"; each item, or grouping, receives a new editor instance.
Tested all launchers are working & drafted the following notes-to-self along the way

Code: Select all

#   noted: Disks --}"Install antiX Linux"    leads straight to minstall  (antixsources.sh not called)

#   noted: closing gparted window while it is scanning... leads to runaway processes: gpartedbin and ntfsresize

#   grub boot image   ??? MUST be: jpg _AND_ must be 8-bit (256 color), non-indexed, RGB ???

skidoo · posted: 2017-10-08 18:16

clipit hotkeys overshadowed by [window manager global keybinds] DON'T KNOW A FIX

Can be fixed by editing /etc/skel/.config/clipit/clipitrc to emove the preconfigured hotkey assignments
(possible that future package upgrade operations will overwrite this file though)

Code: Select all

history_key=
actions_key=
menu_key=
search_key=
offline_key=

dolphin_oracle · posted: 2017-10-08 18:40

back to cli-aptiX

the check fails when the google-chrome repo in google-chrome.list is used. I think the repo definition line may be a case the heuristic hasn't taken into account.

Code: Select all

demo@antix1:~
$ cat /etc/apt/sources.list.d/google-chrome.list 
### THIS FILE IS AUTOMATICALLY CONFIGURED ###
# You may comment out this entry, but any other modifications may be lost.
deb [arch=amd64] http://dl.google.com/linux/chrome/deb/ stable main

I expect its just the

Code: Select all

while read type url repo other

in check_list_files() that fails, since it doesn't account for the [arch=amd64] in the deb line. that line is required for chrome.