antiX-23.1 released

Forum Forums News Announcements antiX-23.1 released

  • This topic has 218 replies, 25 voices, and was last updated Mar 13-5:33 pm by Stéphane Ascoët.
Viewing 15 posts - 196 through 210 (of 219 total)
  • Author
  • #135704
    Brian Masinick

      A good example of a “program” that can influence the appearance and behavior of the system is the program
      /usr/bin/galternatives. It HAS to be run as a privileged program in order to adjust various features.

      I use it to get the cursor color I prefer and a few other things; but that cursor color doesn’t always “STAY PUT” unless I utilize this tool, so I find it to be an extremely useful tool indeed!

      Brian Masinick


        Dear @anticapitalista: in order to better harmonize Fluxbox configuration and features with the ones in IceWM and Fluxbox, I made a commit:
        That commit fixes the “show desktop” button (thanks to the user that showed me the light!), and adds buttons for skippy-xd, antiX-updater, Package Installer, File Manager, Web Browser, and even USB Eject and “Exit” (this last 2 are next to the clock)

        This important think is fixing the “show desktop”, you can accept my other suggestions, and not include them as active in “”. That way you keep the classic “no buttons” fluxbox look and make it easier for users that do want to have those generic buttons (they just have to edit that line, adding the buttons they want).


        • This reply was modified 2 months, 3 weeks ago by PPC.

          You know how to speedy make .mo files for testing the layout from the .po file you are working on, so you can easily check after each next change, whether this change bestowes the desired result on you already or not.

          Yesterday I ran several tests in the Control Center and they were all unsuccessful. I removed all the slashes and kept in the translation only where the slash was in the original text.
          I can’t lose the quality of the translation for the sake of aesthetics. Aesthetics are important, but translation is even more important. Furthermore, why are the contents of each of the options positioned to the right? Is there some code that forces the text and images to the right? Would there be some code that allows you to position it to the left? HTML, which is not even a programming language, has right or left positioning options. If the content were still positioned to the left, I think there would be that useless space on the right.
          At least I managed to make some improvements to the “pt_BR” texts and they are in Transifex, when an update to the program occurs and it is necessary to collect all the translations again, the most recent text in “pt_BR” will certainly be used and will benefit all native speakers of my language.

          – – – – –

          You know how to speedy make .mo files for testing the layout from the .po file you are working on, so you can easily check after each next change, whether this change bestowes the desired result on you already or not.

          Ontem eu fiz vários testes no Centro de Controle e todos foram sem sucesso. Eu removi todas as barras e mantive na tradução apenas onde tinha a barra no texto original.
          Eu não posso perder a qualidade da tradução em pró da estética. A estética é importante, mas a tradução é mais importante ainda. Além do mais, por que o conteúdo de cada uma das opções estão posicionadas para à direita? Será que tem algum código que força o texto e as imagens para à direita? Teria algum código que permite posicionar para à esquerda? O HTML que nem é uma linguagem de programação possui as opções de posicionamento para à direita ou para à esquerda. Se o conteúdo fosse posicionado para à esquerda ainda assim, eu acho que ficaria aquele espaço inútil à direita.
          Ao menos eu consegui fazer algumas melhorias nos textos “pt_BR” e estão no Transifex, quando ocorrer alguma atualização do programa e for necessário colher todas as traduções novamente, o texto mais recente em “pt_BR” certamente será utilizado e beneficiará todos os nativos do meu idioma.


            I removed all the slashes and kept in the translation only where the slash was in the original text.

            Marcelo, you could have saved that time if you had carefully read my answer about this issue. Precisely this was what I had tested already, you’ll find the detailed result some postings above.

            Is there some code that forces the text and images to the right?

            I don’t think so. It’s just the length of the expressions you insist to use in the tab entries which enforces the additional space. I believe there is a bug in gtkdialog, causing the width of the tab headers to have an influence on the field width, even when the tabs are not positioned above but at left side, so this doesn’t make any sense. For testing, try cutting all your tab entries by the half.

            As said, there are probably means to address this programmatically, but this will take hours and hours to search the gtkdialog manuals and try what you find therein. From my recollection when I designed antiXscreenshot2 tabs and aCSTV, both using gtkdialog, I can state this is pretty tricky, and adding a layout instruction in one position breaks the layout in another place the same time… Maybe a true programmer could handle this easily, but I can’t within a reasonable amount of time, since I am no programmer, so won’t touch it.

            Windows is like a submarine. Open a window and serious problems will start.

            Forum Admin

              @Robin, @marcelocripe;
              I have not tried this but some thoughts after a quick look.
              a) This may effect the translation function. However if I am not mistaken, the variables set in for each tab name could be changed from double quotes ” to single quotes ‘ and in doing so it should allow for line breaks to be used in the translation files via adding \n where you would like a new line. If there are troubles with gettext working only with double quotes then later on near the end of the script where the tab_labels are set, change line to use the echo command like tab_labels=$(echo -e “*|*|*|*”) instead of tab_labels=”*|*|*|*”

              b) for the truncating and right align due to large text; Add space-expand=”True” to the
              <notebook * labels=*>
              line to make it like
              <notebook * space-expand="True" labels=*>.
              This should allow for automatic expansion of the tab label both vertically and horizontally when the text overflows default sizes.

              • This reply was modified 2 months, 3 weeks ago by Dave.
              • This reply was modified 2 months, 3 weeks ago by Dave.

              Computers are like air conditioners. They work fine until you start opening Windows. ~Author Unknown


                Hello anticapitalista
                fluxbox button Show Desktop
                trial that code starting with conky off.

                Did Set up a system to trial fluxbox 1.35stable.
                The system did not have wmctrl, so testing started with
                plugging-in an Activate pattern command and it worked.
                Application windows normal responded to Activate
                onboard, magnus, gkrellm, and featherpad, file manager, synaptic, etc.
                Application windows tint2, lxpanel responded to a Deiconify command one for each panel. All of it remains visible by choice with ShowDesktop.
                conky pattern sometimes works, conky config and layer desktop a lot of variables.
                Can configure to suit most situation. Such as ft10 tint2 is one DeIconify has it stay on screen.

                Turn to the antiX fluxbox 1.37+antix23May 5 2023 with the same commands
                the DeIconify for tint2 works maybe because it does not use names to map.
                Activate not working. shared in its place the wmctrl ,,,
                Turn to mx fluxbox old alpha has fluxbox 1.37 and Activate semi-works with most except conky.
                Turn to diversity download three debs. 1.35+b not work. 1.35-2.1 semi-works. 1.37 cannot install. 1.37+available in antiX not working with same methods. need to find the new ways for these conditions.
                Clean lean mean is ShowDesktop.


                  Just wanted to delete some files I had stored on my disk shortly before, pressing the del key in zzzFM. (Meant originally to delete them, but they had merely a few bytes of size so I didn’t care for the del key would send them to trash instead.)

                  To my surprise the files won’t go away. So I repeated it, pressed del another time, and another and another. Nothing happened, besides: The PC turned into an air conditioner all of a sudden, vent went up, and a short look to the cpu-meter revealed it was groaning with full system load, processing 100% on all 4 cores, CPU frequency auto-raised to top speed. Screen, mouse and keyboard mostly irresponsive.

                  It took 10 minutes to get the system responsive again, by repeatedly pressing ctrl+alt+f2 in long intervals first, to enter console mode (since screen was frozen and irresponsive after a short time), then spotting the culprit by ps and finally killing the four ragging trash processes by sending them an unapologetic kill command from within terminal directly, one by one. The very moment the CPU load went back to normal and not long after this the air conditioner turned back into a silent PC.

                  caught from htop:

                  CPU%  Time+    Command
                  99%   0:02:09  /usr/bin/python3 /usr/bin/trash /path/to/filename...
                  100%  0:02:20  /usr/bin/python3 /usr/bin/trash /path/to/filename...
                  101%  3:10:40  /usr/bin/python3 /usr/bin/trash /path/to/filename...
                  100%  9h04:11  /usr/bin/python3 /usr/bin/trash /path/to/filename...

                  You can see /usr/bin/trash runs amok, here I saw 4 times ~100% system load (on an 2(4) core i3), causing system almost to BSOD.
                  When pressing the del button some more times, more of these 100% trash processes will start, and your system will stall completely I reckon.

                  After this event I did some testing to avoid this kind of trouble in future. What caused it?

                  – Filesystem was fine, no errors found.
                  – Deleting the files (shift+del) instead of sending them to trash (by del merely) worked flawlessly, from within zzzFM as well as by console command.

                  Only moving them to trash by using the del key caused trouble, and this is reproducible, with some more files.

                  First I was suspecting the blanks or special characters in some of them could have caused it, but no, when creating testfiles containing the very characters or blanks, they went to trash flawlessly. So next I turned my attention to the character set itself, the untrashable files happened to have filenames in Cyrillic characters. So I created some test files with a couple of Cyrillic characters. Again, these went to trash flawlessly.

                  Then I checked the filename length. And that actually was the troublemaker. Even when I have sent flawlessly files to trash with even longer file names, these files caused trouble now. I created test files of different length using arbitrary Cyrillic characters, and it turned out, they fail already when merely having half of the size of English file names.

                  Check out yourself, create test files with these names (make sure to copy the full strings):


                  And then highlight them and press the del key in zzzFM a couple of times on them, see what happens.

                  This must be fixed.

                  Sorry, have no time to look further into it, have never studied the send-to-trash script. Somebody else will have to take care of this issue.

                  Until this is fixed: Warning to the users. Don’t press the del key another time if file won’t move to trash on first attempt. Instead, immediately open your process management and kill the out-of-control 100% /usr/bin/trash process. Then delete the file instead of sending it to trash (or shorten the filename by the half before trashing it). This goes for very long filenames in English language and only half as long filenames already in other character sets.

                  Windows is like a submarine. Open a window and serious problems will start.


                    My hat is off to anticapitalista, because he keeps on going, even when antiX is a boat that seems to be trying to sink at the same time we try to plug the holes and renovate it… I’m losing my will power…

                    To the point:
                    I was the one responsable to adding the Trash feature to antiX, in antiX 23. My initial versions of that feature merely copied to files to be discarded to the Trash folder, but I then found out about trash-cli, a terminal app that did that, also creating an index of the place or origin of the file/folder, an including a “restore” option, that users could use to… well, restore stuff sent to trash. Both I and anticapitalista thought it would be better to use a dedicated app for that end.
                    It seems that what you are reporting is a bug in trash-cli. I can’t solve it, but I can, after trying to deal with other problems, redo the script that “sends file to the trash can” and use a simple “move” command. Doing that is easy. The problematic problem is creating the index, used to “restore” the files…

                    @Robin and Marcelo seem to have a nose to sniff out very improvable situations that no other user comes upon and report them 🙁 In a way, that is good: We know exactly what problems we have. The bad part: we know EXACTLY how many problems we have to solve.

                    I just tested, and I can confirm I replicated this bug in antiX 23.1 64 runit Full, in Live mode.

                    Work-around/ partial solution:

                    Right click a file > Right click “Send to Trash” > left click “Edit Command” and replace the current command with:
                    folder=%d; if [[ ${folder} != *".local/share/Trash/files"* ]]; then mv %F .local/share/Trash/files; fi

                    This works, IF the folder ~/.local/share/Trash/files already exists, so, it’s probably better to use this command:

                    mkdir -p ~/.local/share/Trash/files ; folder=%d; if [[ ${folder} != *".local/share/Trash/files"* ]]; then mv %F .local/share/Trash/files; fi

                    This will always safely move files to the trash folder, but once in the Trash, the “restore” button does not work.
                    The way out to stick with this simple solution and not break the “restore” function would be replacing the “restore” command with one that opens a GUI, asking where should the file(s) by restored to. In some ways this would be very practical, since I have restored files and forgot what their original place was, and had to end up searching for the restored files, at least twice…


                    • This reply was modified 2 months, 2 weeks ago by PPC.
                    • This reply was modified 2 months, 2 weeks ago by PPC.
                    • This reply was modified 2 months, 2 weeks ago by PPC.

                      Howdy forum,

                      I finally installed antiX-23.1 full (SysVinit, 64-bit) on my Samsung R60 plus Notebook from 2008.
                      I hesitated, since it does not like kernels 5.x and above and – ran into problems with
                      the SATA controller with both provided kernels. But they worked good enough to install
                      4.9.0-326 and it runs decently.
                      But now I get a lot of “invalid argument” messages on boot – is there a way to fix this?


                        is a bug in trash-cli.

                        @PPC Then probably best solution would be to file a bug report at the package maintainers or the program author. Looked up trash-cli is packaged by debian, so I could just use their bug-reporting system, simply by sending a form via Email to the debian bug-list tracker. But in this case I’d ask you to do the reporting, since I have no insight how your script makes use of trash-cli internally, so I can’t answer detailed questions or append adequate pieces of information. I guess they will care for this speedy, since it causes systems to BSOD, so it is a severe bug in this application. I reckon it is caused by the way the trashed files are managed, by appending something to the filename rather than replacing the end of the filename without making it longer than before (which could be done, since the proper filename is stored somewhere else for restore anyway, if I’m not mistaken from a short glance at it.

                        even when antiX is a boat that seems to be trying to sink at the same time we try to plug the holes and renovate it

                        Oh, c’mon, antiX is on a fine path. Nothing what humans have created was ever done without bugs or errors. So that’s just business as usual. Nothing to be worried about. We’ll fix it, sooner or later 🙂

                        Windows is like a submarine. Open a window and serious problems will start.

                        Stéphane Ascoët

                          Do you mean like the old Windows “character map” program? Show all characters in a GUI, and the user can click anyone and it gets “typed”?

                          Yes, it would have helped us at this event. I don’t remember how we unlocked the account finally, I guess we tried several keyboards maps, looked what character was written with each of them to find the one that may be in the password. But I’ve recalled the trick our leader told us after that, I wonder if Qwerty user are concerned: it’s the ability to put some character in exposant after having typed “^”. For example, I made “²” this way just now. Of course, it doesn’t work with “a”: I get â.

                          EDIT: if you want something like that, you can install the package gnome-characters using the Package Installer “Repos” tab, Synaptic, or apt, via the terminal.

                          I don’t like Gnome but thanks! Don’t worry for package management, I’m an expert 😉


                            Hello Stéphane Ascoët
                            character map

                            libre office toolbar File Menu bar >Insert >Special Character
                            is a character map with search.

                            kcharselect is a character map with search.

                            gucharmap is a character map no search.

                            Brian Masinick

                              (man page for dumpkeys; man 5 keymaps may also be useful).

                              DUMPKEYS(1) General Commands Manual DUMPKEYS(1)

                              dumpkeys – dump keyboard translation tables

                              dumpkeys [OPTIONS]

                              dumpkeys writes, to the standard output, the current contents of the
                              keyboard driver’s translation tables, in the format specified by

                              Using the various options, the format of the output can be controlled
                              and also other information from the kernel and the programs dump‐
                              keys(1) and loadkeys(1) can be obtained.

                              -h –help
                              Prints the program’s version number and a short usage message
                              to the program’s standard error output and exits.

                              -i –short-info
                              Prints some characteristics of the kernel’s keyboard driver.
                              The items shown are:

                              Keycode range supported by the kernel

                              This tells what values can be used after the keycode
                              keyword in keytable files. See keymaps(5) for more in‐
                              formation and the syntax of these files.

                              Number of actions bindable to a key

                              This tells how many different actions a single key can
                              output using various modifier keys. If the value is 16
                              for example, you can define up to 16 different actions
                              to a key combined with modifiers. When the value is 16,
                              the kernel probably knows about four modifier keys,
                              which you can press in different combinations with the
                              key to access all the bound actions.

                              Ranges of action codes supported by the kernel

                              This item contains a list of action code ranges in hexa‐
                              decimal notation. These are the values that can be used
                              in the right hand side of a key definition, ie. the vv’s
                              in a line

                              keycode xx = vv vv vv vv

                              (see keymaps(5) for more information about the format of
                              key definition lines). dumpkeys(1) and loadkeys(1) sup‐
                              port a symbolic notation, which is preferable to the nu‐
                              meric one, as the action codes may vary from kernel to
                              kernel while the symbolic names usually remain the same.
                              However, the list of action code ranges can be used to
                              determine, if the kernel actually supports all the sym‐
                              bols loadkeys(1) knows, or are there maybe some actions
                              supported by the kernel that have no symbolic name in
                              your loadkeys(1) program. To see this, you compare the
                              range list with the action symbol list, see option
                              –long-info below.

                              Number of function keys supported by kernel

                              This tells the number of action codes that can be used
                              to output strings of characters. These action codes are
                              traditionally bound to the various function and editing
                              keys of the keyboard and are defined to send standard
                              escape sequences. However, you can redefine these to
                              send common command lines, email addresses or whatever
                              you like. Especially if the number of this item is
                              greater than the number of function and editing keys in
                              your keyboard, you may have some “spare” action codes
                              that you can bind to AltGr-letter combinations, for ex‐
                              ample, to send some useful strings. See loadkeys(1) for
                              more details.

                              Function strings

                              You can see you current function key definitions with
                              the command

                              dumpkeys –funcs-only

                              -l -s –long-info
                              This option instructs dumpkeys to print a long information
                              listing. The output is the same as with the –short-info ap‐
                              pended with the list of action symbols supported by loadkeys(1)
                              and dumpkeys(1), along with the symbols’ numeric values.

                              -n –numeric
                              This option causes dumpkeys to by-pass the conversion of action
                              code values to symbolic notation and to print the in hexadeci‐
                              mal format instead.

                              -f –full-table
                              This makes dumpkeys skip all the short-hand heuristics (see
                              keymaps(5)) and output the key bindings in the canonical form.
                              First a keymaps line describing the currently defined modifier
                              combinations is printed. Then for each key a row with a column
                              for each modifier combination is printed. For example, if the
                              current keymap in use uses seven modifiers, every row will have
                              seven action code columns. This format can be useful for exam‐
                              ple to programs that post-process the output of dumpkeys.

                              -Sshape –shape=shape
                              Available shapes:

                              2 default output.

                              4 one line for each keycode.

                              8 one line for each (modifier,keycode) pair.

                              16 one line for each keycode until 1st hole.

                              -1 –separate-lines
                              This forces dumpkeys to write one line per (modifier,keycode)
                              pair. It prefixes the word plain for plain keycodes.

                              -t –funcs-only
                              When this option is given, dumpkeys prints only the function
                              key string definitions. Normally dumpkeys prints both the key
                              bindings and the string definitions.

                              -k –keys-only
                              When this option is given, dumpkeys prints only the key bind‐
                              ings. Normally dumpkeys prints both the key bindings and the
                              string definitions.

                              -d –compose-only
                              When this option is given, dumpkeys prints only the compose key
                              combinations. This option is available only if your kernel has
                              compose key support.

                              -ccharset –charset=charset
                              This instructs dumpkeys to interpret character code values ac‐
                              cording to the specified character set. This affects only the
                              translation of character code values to symbolic names. Valid
                              values for charset currently are iso-8859-X, Where X is a digit
                              in 1-9. If no charset is specified, iso-8859-1 is used as a
                              default. This option produces an output line `charset
                              “iso-8859-X”‘, telling loadkeys how to interpret the keymap.
                              (For example, “division” is 0xf7 in iso-8859-1 but 0xba in

                              -Cdev –console=dev
                              The affected console device can be specified using the -C (or
                              –console ) option. This option supports exactly one device

                              -v –verbose
                              Turn on verbose output.

                              -V –version
                              Prints version number and exits.

                              The recommended directory for keytable files.

                              SEE ALSO
                              loadkeys(1), keymaps(5)

                              kbd 1 Sep 1993 DUMPKEYS(1)

                              Brian Masinick

                              Brian Masinick

                                I’ve used xmodmap in the past; that may be even more useful if you don’t use libre office toolbar keymap.

                                Brian Masinick


                                  This will always safely move files to the trash folder, but once in the Trash, the “restore” button does not work.

                                  Moving all trashed files to ~/.local/share/Trash/files is not a great solution since when you trash files from another drive/partition, they’ll have to get copied to the partition with $HOME on it. Kind of a pain for large files and it’s not really the way trash bins ought to work, I don’t think. Every partition should have its own trash folder so that no data needs to get moved when trashing a file. trash-cli handles that automatically.

                                  I just did some testing and it seems like the issue Robin found (in addition to the fact that Cyrillic characters are two bytes each) is that filenames between 246 and 255 bytes long will be over 255 bytes once the “.trashinfo” extension gets added, which is too long for Linux filesystems. So it could be resolved by just not adding that extension, though that would create files in the Trash/info directory with incorrect extensions, I guess.

                                  Since this is such an edge case (filenames between 246 and 255 bytes), why not just continue to use trash-cli and work around it? Adding a check to the zzzfm command is easy enough. Something like this:

                                  if [[ ${folder} != *".local/share/Trash/files"* ]]; then
                                    for file in %N; do
                                      if [ $(echo -n "$file" | wc -c) -gt 245 ]; then
                                        zzzfm -g --title "ERROR: Send to Trash" --label "Filename is too long to send to trash.\n\n$file" --button Sorry
                                        exit 1
                                    trash %F

                                  So if trashing multiple files, the whole operation will fail if any of the selected filenames is too long. A dialog gives you the offending filename, though there could be others since it exits after the first catch. I suppose you could work through them all and even use an ‘else’ line to still trash the other files that pass the test and only error on the ones that are too long, but I don’t know, seems kind of complicated for such a rare event.

                                  EDIT: Heck, I guess it’s not actually that complicated, is it?

                                  if [[ ${folder} != *".local/share/Trash/files"* ]]; then
                                    for file in %N; do
                                      if [ $(echo -n "$file" | wc -c) -gt 245 ]; then
                                        zzzfm -g --title "ERROR: Send to Trash" --label "Filename is too long to send to trash.\n\n$file" --button Sorry
                                        trash "$folder"/"$file"
                                  • This reply was modified 2 months, 2 weeks ago by anti-apXos.
                                Viewing 15 posts - 196 through 210 (of 219 total)
                                • You must be logged in to reply to this topic.