dbus/dunstctl: add history command (#307)

[nikp123]

The current implementation tries to retrieve notifications stored in the
history buffer (which means they've been either discarded or hidden by
the user) and does so using the existing D-Bus system. However the
output of the dbus-send command is (in my opinion) super ugly and
unusable for any practical scripting application.

This is VERY much still a work in progress.

Any additional suggestions or ideas are welcome.

[fwsmit]

Maybe you could use busctl? After some playing around I came up with this command: busctl --user --json=pretty call org.freedesktop.Notifications /org/freedesktop/Notifications org.dunstproject.cmd0 NotificationListAll.
Btw, don't forget to add the method name to the introspection_xml and methods_fdn. And add the new command in the dunstctl --help output and man page.

[nikp123]

I like the idea. I've been also playing around the idea that instead of arrays i use a dbus dictionary instead

ie this: https://stackoverflow.com/questions/8846671/how-to-use-a-variant-dictionary-asv-in-dbus-send#38573254

[fwsmit]

I like the idea. I've been also playing around the idea that instead of arrays i use a dbus dictionary instead

ie this: https://stackoverflow.com/questions/8846671/how-to-use-a-variant-dictionary-asv-in-dbus-send#38573254

A dictionary seems good, since it would create a stable interface. You can add or remove items without breaking all scripts.

[nikp123]

I like the idea. I've been also playing around the idea that instead of arrays i use a dbus dictionary instead
ie this: https://stackoverflow.com/questions/8846671/how-to-use-a-variant-dictionary-asv-in-dbus-send#38573254

A dictionary seems good, since it would create a stable interface. You can add or remove items without breaking all scripts.

Implemented :) after an annoyingly long session of googling

[nikp123]

it's done

[fwsmit]

Thanks for working on this. If you want you can add the script you're gonna use this feature with to the contrib directory

[nikp123]

Thanks for working on this. If you want you can add the script you're gonna use this feature with to the contrib directory

sure

[codecov-commenter]

Codecov Report

Merging #970 (675993a) into master (e4ada4e) will decrease coverage by 0.08%.
The diff coverage is 55.08%.

@@            Coverage Diff             @@
##           master     #970      +/-   ##
==========================================
- Coverage   60.37%   60.29%   -0.09%     
==========================================
  Files          44       44              
  Lines        6655     6772     +117     
==========================================
+ Hits         4018     4083      +65     
- Misses       2637     2689      +52     

Flag	Coverage Δ
unittests	60.29% <55.08%> (-0.09%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
src/dbus.c	52.17% <0.00%> (-7.66%)	⬇️
src/queues.c	91.44% <88.88%> (-0.19%)	⬇️
test/queues.c	98.86% <100.00%> (+0.09%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update e4ada4e...675993a. Read the comment docs.

[fwsmit]

Could you add tests for queues_history_pop_by_id and dbus_cb_dunst_NotificationListHistory? We try to keep most of the code tested.

[nikp123]

sure, just not immediately

[fwsmit]

I have a few more comments on the code :)

[fwsmit]

sure, just not immediately

That's okay. Take your time 👍

[nikp123]

seems that queues_update breaks after 5 messages in waiting. is it supposed to do that?

[nikp123]

let me describe the breakage in more detail.

If i put into waiting more than 5 messages with skip_display on, ONLY the first five get to remain in the history after a queues_update(STATUS_NORMAL);

[nikp123]

oops seems that i have messed up the function prefixes when using sed, one sec

[nikp123]

here you go; added some tests

unfortunately i think adding the dbus tests would create way more code than the dbus functions themselves, at which point you would most likely debug the tests themselves than the actual code

[fwsmit]

There are still 2 small things that need to be done

[fwsmit]

Thanks for adding the tests! It's fine if you don't add the dbus tests.

[fwsmit]

It seems like a conflict was caused by me merging another PR

[nikp123]

here we go

[fwsmit]

Merged it!

[fwsmit]

I'm thinking that we don't need a separate command for history-pop-id. history-pop can be used with an argument instead. I'll go ahead and change that.

[nikp123]

thanks