bat_now.status is nil #451

bitwombat · 2020-04-11T12:28:42Z

output of awesome -v and lua -v

awesome v4.2 (Human after all)
lua not on path

expected behavior and actual behavior
Expecting bat_now.status to have a value as defined in the wiki

Actual: bat_now.status is nil
without bat_now.status, the below config works.

steps to reproduce the problem

local theme                                     = {}
theme.font                                      = "Terminus 10.5"

local lain  = require("lain")
local markup     = lain.util.markup
local separators = lain.util.separators
local gray       = "#9E9C9A"

local bat = lain.widget.bat({
    settings = function()
        bat_header = " Bat "
        bat_p      = bat_now.perc .. "% "
        bat_t      = bat_now.time .. " "
        bat_s      = bat_now.status .. " "
        widget:set_markup(markup.font(theme.font, markup(gray, bat_header) .. bat_p .. bat_t .. bat_s))
    end
})

X error log

Apr 11 22:25:43 thinkpad /usr/lib/gdm3/gdm-x-session[1402]: stack traceback:
Apr 11 22:25:43 thinkpad /usr/lib/gdm3/gdm-x-session[1402]: #011/home/gbell2/.config/awesome/rc.lua:154: in upvalue 'settings'
Apr 11 22:25:43 thinkpad /usr/lib/gdm3/gdm-x-session[1402]: #011/home/gbell2/.config/awesome/lain/widget/bat.lua:185: in local 'func'
Apr 11 22:25:43 thinkpad /usr/lib/gdm3/gdm-x-session[1402]: #011/usr/share/awesome/lib/gears/object.lua:143: in function 'gears.object.emit_signal'
Apr 11 22:25:43 thinkpad /usr/lib/gdm3/gdm-x-session[1402]: #011/home/gbell2/.config/awesome/lain/helpers.lua:97: in function 'lain.helpers.newtimer'
Apr 11 22:25:43 thinkpad /usr/lib/gdm3/gdm-x-session[1402]: #011/home/gbell2/.config/awesome/lain/widget/bat.lua:212: in function 'lain.widget.bat'
Apr 11 22:25:43 thinkpad /usr/lib/gdm3/gdm-x-session[1402]: #011/home/gbell2/.config/awesome/rc.lua:149: in main chunk
Apr 11 22:25:43 thinkpad /usr/lib/gdm3/gdm-x-session[1402]: error: /home/gbell2/.config/awesome/rc.lua:154: attempt to concatenate a nil value (field 'status')
Apr 11 22:25:49 thinkpad /usr/lib/gdm3/gdm-x-session[1402]: 2020-04-11 22:25:49 E: Error opening directory '/usr/share/awesome/applications': No such file or directory
Apr 11 22:25:49 thinkpad /usr/lib/gdm3/gdm-x-session[1402]: 2020-04-11 22:25:49 E: Error opening directory '/usr/local/share/applications': No such file or directory
Apr 11 22:25:49 thinkpad /usr/lib/gdm3/gdm-x-session[1402]: 2020-04-11 22:25:49 E: Error opening directory '/home/gbell2/.nix-profile/share/applications': No such file or directory
Apr 11 22:25:49 thinkpad /usr/lib/gdm3/gdm-x-session[1402]: 2020-04-11 22:25:49 E: Error opening directory '/usr/share/awesome/applications': No such file or directory
Apr 11 22:25:49 thinkpad /usr/lib/gdm3/gdm-x-session[1402]: 2020-04-11 22:25:49 E: Error opening directory '/usr/local/share/applications': No such file or directory
Apr 11 22:25:49 thinkpad /usr/lib/gdm3/gdm-x-session[1402]: 2020-04-11 22:25:49 E: Error opening directory '/home/gbell2/.nix-profile/share/applications': No such file or directory
Apr 11 22:25:50 thinkpad /usr/lib/gdm3/gdm-x-session[1402]: 2020-04-11 22:25:50 W: awesome: a_glib_poll:432: Last main loop iteration took 0.619277 seconds! Increasing limit for this warning to that value.

The text was updated successfully, but these errors were encountered:

bitwombat · 2020-04-11T12:37:46Z

Looking at bat.lua, I see that it sets bat_now.status to bat_now.n_status[1], which is null because "battery" isn't set.

This wasn't autodetected. It's BAT0 on my system. Setting arg.battery solves the problem, but I'm wondering if this should be considered a bug?

lcpz · 2020-04-11T21:33:36Z

lua not on path

What does this mean?

Do you have the directory /sys/class/power_supply/?

Try the following in a Lua shell (spawn it in a terminal by typing lua):

io.popen("ls -1 /sys/class/power_supply/"):read("*all"):match("BAT%w+")

Tell me if it outputs "BAT0".

Also, if you have a ThinkPad, consider trying this widget instead.

bitwombat · 2020-04-12T06:01:48Z

What does this mean?

The lua binary is not on my shell's path, as defined by the PATH environment variable...

It doesn't seem to be installed anywhere. Awesome seems to use liblua5.3-0, and not require a lua CLI?

Anyway...

Do you have the directory /sys/class/power_supply/?

Yep, which is why everything works when I define arg.battery.

Tell me if it outputs "BAT0".

Yep, which is why everything works when I define arg.battery.

lcpz · 2020-04-14T16:43:36Z

Yep, which is why everything works when I define arg.battery.
Yep, which is why everything works when I define arg.battery.

I know that you have the file. What I asked you is to run a command in the Lua command-line interpreter that does exactly what the widget does in this function, to see what is the output and fix it.

Both Awesome and Lain depend on Lua.

I understand that liblua5.3.-0 installs it, even though you say you do not have it in your PATH variable.

Hence, from what you say I do not understand what you have done.

Can I assume that you added Lua to your user path and ran the command as I asked?

Otherwise, do it.

And if you want my further attention, avoid pointless and childish copy-pasting.

bitwombat · 2020-04-15T13:25:35Z

Sorry, I didn't realize copy-pasting could be construed as childish. I was simply answering two different questions with the same answer.

There is no lua binary on this system, yet Awesome works fine. The package liblua5.3.-0 only has some .so files in it. The lua binary seems to be in lua5.2, which is not installed on this system.

Just some confusion, that's all. I've installed lua5.2, just to run your test. Outputs BAT0, no problem.

The problem is in here somewhere.

lain/widget/bat.lua

Lines 39 to 48 in 33c0e0c

    
           function bat.get_batteries() 
        
               helpers.line_callback("ls -1 " .. pspath, function(line) 
        
                   local bstr =  string.match(line, "BAT%w+") 
        
                   if bstr then 
        
                       batteries[#batteries + 1] = bstr 
        
                   else 
        
                       ac = string.match(line, "A%w+") or "AC0" 
        
                   end 
        
               end) 
        
           end

It's not pspath - that's correct. Hmm... I'll let you know what I figure out.

lcpz · 2020-04-15T20:31:23Z

I was simply answering two different questions with the same answer.

Just be nice and not redundant. It's my honest advice.

I'll let you know what I figure out.F

You say that the string matching works. Let's debug the function integrity.

Change this:

lain/widget/bat.lua

Line 50 in 33c0e0c

if #batteries == 0 then bat.get_batteries() end

To this:

bat.get_batteries()

If it doesn't work, put after here the following line:

naughty.notify{ timeout = 0, text = bstr .. " " .. tostring(batteries[1]) .. " " .. ac }

bitwombat · 2020-04-15T22:09:32Z

You say that the string matching works. Let's debug the function integrity.

I'm a few steps ahead of that. Believe it or not, it's actually something to do with Gio.UnixInputStream. Easiest way to demonstrate it is that

helpers.line_callback("ls -1 /" ...

Works, but

helpers.line_callback("ls -1 " .. pspath, ...

doesn't work, even though pspath is correct. It either has something to do with the nature of /sys, or with the fact that BAT0 is a symlink in there.

lcpz · 2020-04-15T23:08:56Z

Is pspath somehow nil in that scope?

Did you try constructing the argument before?

if pspath == nil then naughty.notify { text = "nil", timeout = 0 } end
local cmd_str = string.format("ls -1 %s", pspath)
helpers.line_callback(cmd_str, function(line)
    -- [...]
end)

bitwombat · 2020-04-16T00:34:35Z

Is pspath somehow nil in that scope?

No, believe it or not! Which is why I'm down in the pits looking at Gio.UnixInputStream

I should have been clearer and said:

helpers.line_callback("ls -1 /" ...

gives results, while

helpers.line_callback("ls -1 /sys/class/power_supply" ...

returns nothing, even though from the shell ls -1 /sys/class/power_supply returns two lines. So it's something in spawn or with Gio.UnixInputStream

lcpz · 2020-04-16T19:49:15Z

If this is not a bug of mine, I can only speculate:

Maybe GLib2 does not have access permission to /sys/class/power_supply.
Maybe you need Glib2 > 2.53 for some reason.

EDIT: The net widget also uses helpers.line_callback. Do you have issues with that as well?

bitwombat · 2020-04-17T12:10:43Z

Is line_callback non-blocking?

Because I can't make sense of the trace.

I sprinkled some notifies in. It is finding BAT0.

    function bat.get_batteries()
        naughty.notify { timeout = 0, text = 'Autodetecting'}
        helpers.line_callback("ls -1 " .. pspath, function(line)
            local bstr =  string.match(line, "BAT%w+")
            naughty.notify { timeout = 0, text = bstr }
            if bstr then
                batteries[#batteries + 1] = bstr
                naughty.notify { timeout = 0, text = 'is now' .. #batteries }
            else
                ac = string.match(line, "A%w+") or "AC0"
            end
        naughty.notify { timeout = 0, text = 'at the inner end is' .. #batteries }
        end)
        naughty.notify { timeout = 0, text = 'at the end is' .. #batteries }
    end

    if #batteries == 0 then bat.get_batteries() end
    naughty.notify { timeout = 0, text = 'later is' .. #batteries }

What I get is

Autodetecting
at the end is 0
later is 0
at the inner end is 0
BAT0
is now 1
at the inner end is 1

Besides the weird order ("later is" where it is), it looks like the callback's getting called twice?

lcpz · 2020-04-17T20:31:24Z

line_callback is asynchronous, hence non-blocking.

Your log makes perfect sense:

The If condition is true; get_batteries is spawned.
'Autodetecting' is printed; line_callback is spawned, it will be executed on the nex Awesome cycle.
'at the end is' is printed; get_batteries terminates.
The line after the If is executed: 'later is' is printed.
Next Awesome cycle: line_callback starts; the command cmd = "ls -1 ..." is spawned; the callback function(line) ... end is executed on each line returned by cmd.
First line does not match "BAT%w+"; #batteries = 0, hence "at the inner end is 0" is printed.
Second line matches; #batteries = 1, hence printing of bstr, "is now 1" and "at the inner end is 1".
There are no more lines: line_callback terminates.

As a result, the widget is perfectly working, the problem is that you are trying to get the information before this is retrieved.

The solution is to use defensive programming:

local bat = lain.widget.bat {
    settings = function()
        bat_header = " Bat "
        bat_p      = (bat_now.perc or "N/A") .. "% "
        bat_t      = (bat_now.time or "N/A") .. " "
        bat_s      = (bat_now.status or "N/A/") .. " "
        widget:set_markup(markup.font(theme.font, markup(gray, bat_header) .. bat_p .. bat_t .. bat_s))
    end
}

The reason behind is that my widgets are asynchronous. However, this style of programming is also useful for detecting system changes (e.g., battery removed or added) or failures (e.g., battery failing!).

Sorry, I should have detected this immediately from your code. Anyway, I believe this thread is a contribution to new users. I will add a link to the main wiki page.

In the future, check how I configure widgets in my themes for reference. You have different examples there.

Also, as I wrote before, you can check tp_smapi if you're using a ThinkPad, or also my upower recipe.

bitwombat · 2020-04-17T23:53:02Z

Because of operator precedence, the fix actually needs to be:

local bat = lain.widget.bat {
    settings = function()
        bat_header = " Bat "
        bat_p      = bat_now.perc .. "% "
        bat_t      = bat_now.time .. " "
        bat_s      = (bat_now.status or "N/A/ ") .. " "
        widget:set_markup(markup.font(theme.font, markup(gray, bat_header) .. bat_p .. bat_t .. bat_s))
    end
}

Also, I've added the guard to just bat_s because only bat_now.status comes back nil.

So, you could probably fix it for all users at:
https://github.com/lcpz/lain/blob/master/widget/bat.lua#L139

If I'm not mistaken, the widget updates every 30 seconds, not asynchronously?

lcpz · 2020-04-18T01:22:18Z

I will add the fix, thanks.

If I'm not mistaken, the widget updates every 30 seconds, not asynchronously?

That's the update frequency, but the update itself is asynchronous, as you have seen.

bitwombat · 2020-04-18T07:49:39Z

Cool, after a rough start, I knew we'd be friends Luca.
Thanks for talking through that with me, and for the fix.

lcpz closed this as completed in af125b3 Apr 18, 2020

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

bat_now.status is nil #451

bat_now.status is nil #451

bitwombat commented Apr 11, 2020

bitwombat commented Apr 11, 2020

lcpz commented Apr 11, 2020

bitwombat commented Apr 12, 2020

lcpz commented Apr 14, 2020

bitwombat commented Apr 15, 2020

lcpz commented Apr 15, 2020

bitwombat commented Apr 15, 2020

lcpz commented Apr 15, 2020

bitwombat commented Apr 16, 2020 •

edited

Loading

lcpz commented Apr 16, 2020 •

edited

Loading

bitwombat commented Apr 17, 2020

lcpz commented Apr 17, 2020 •

edited

Loading

bitwombat commented Apr 17, 2020 •

edited

Loading

lcpz commented Apr 18, 2020

bitwombat commented Apr 18, 2020

bat_now.status is nil #451

bat_now.status is nil #451

Comments

bitwombat commented Apr 11, 2020

bitwombat commented Apr 11, 2020

lcpz commented Apr 11, 2020

bitwombat commented Apr 12, 2020

lcpz commented Apr 14, 2020

bitwombat commented Apr 15, 2020

lcpz commented Apr 15, 2020

bitwombat commented Apr 15, 2020

lcpz commented Apr 15, 2020

bitwombat commented Apr 16, 2020 • edited Loading

lcpz commented Apr 16, 2020 • edited Loading

bitwombat commented Apr 17, 2020

lcpz commented Apr 17, 2020 • edited Loading

bitwombat commented Apr 17, 2020 • edited Loading

lcpz commented Apr 18, 2020

bitwombat commented Apr 18, 2020

bitwombat commented Apr 16, 2020 •

edited

Loading

lcpz commented Apr 16, 2020 •

edited

Loading

lcpz commented Apr 17, 2020 •

edited

Loading

bitwombat commented Apr 17, 2020 •

edited

Loading