adding a method to patch linux idle check with required libraries.

[0x0OZ]

Fixes # Linux checking being idle, A fix for #1187

Description of the Change

added a method that does the check without really replacing it with the existing poor check to be reviewed and added by a C++ developer. Alternate Designs

Release Notes

Fixing linux idle check by listening for input events in /dev/input/*.

If this change is not user-facing or notable enough to be included in release notes you may use the string "N/A" here. --> To this run correctly, the user running the client has to have read and execute permissions on /dev/input/* files, this can be done using setfacl

[0x0OZ]

It seems I added the includes in the wrong place as android is still trying to include them

https://github.com/BOINC/boinc/pull/5354/files#diff-b710fcc072350b8bbb1bc13eb394da5cdbc0b0dfab019a93890ec10eadb41b86R180-R188

[AenBleidd]

@0x0OZ, Client is failed to build as well, you need to include 'libevdev' here to install it on CI: https://github.com/BOINC/boinc/blob/eac3e3a2d9fc74b95290273ebe2e9b0b9b133997/.github/workflows/linux.yml#L53

[0x0OZ]

@AenBleidd client still fails, I need to add -levdev flag to the MakeFile, but I am not sure how it's being built

[AenBleidd]

@AenBleidd client still fails,

I need to add -levdev flag to the MakeFile, but I am not sure how it's being built

You need to add this linker flags somewhere to this file: https://github.com/BOINC/boinc/blob/master/configure.ac

[0x0OZ]

@AenBleidd client still fails, I need to add -levdev flag to the MakeFile, but I am not sure how it's being built

You need to add this linker flags somewhere to this file: https://github.com/BOINC/boinc/blob/master/configure.ac

Like I understand a single word 😅, I'll try to.

[0x0OZ]

I made it worse with the last change to configure.ac... I will try to find out how this thing work.

[0x0OZ]

@AenBleidd Linux client now builds successfully, But other clients still fail because they try to import libevdev when it should be included only in Linux env.

[0x0OZ]

Hey @AenBleidd, I want to remind you of this pull, is there anything I need to do from my side?

[AenBleidd]

@0x0OZ, it would be nice to fix this build for all other OSs...

[0x0OZ]

Snap fails because libevdev is not being installed for it, it's only installed using apt, android fails because it tries to add the linker -levdev. I thought adding the weird condition I added should fix it but it didn't, for snap I couldn't yet figure out where the package installation should be added.

[0x0OZ]

@0x0OZ, it would be nice to fix this build for all other OSs...

That's done, everything builds successfully now.

[codecov[bot]]

Codecov Report

Merging #5354 (ffc5f0d) into master (eac3e3a) will decrease coverage by 0.01%. Report is 16 commits behind head on master. The diff coverage is n/a.

Additional details and impacted files

@@             Coverage Diff              @@
##             master    #5354      +/-   ##
============================================
- Coverage     10.87%   10.87%   -0.01%     
  Complexity     1068     1068              
============================================
  Files           279      279              
  Lines         36059    36061       +2     
  Branches       8335     8335              
============================================
  Hits           3920     3920              
- Misses        31745    31747       +2     
  Partials        394      394              

see 3 files with indirect coverage changes

[0x0OZ]

Snap failed, I thought it did not as at first it was skipped and didn't give an error.

[0x0OZ]

@AenBleidd All success but snap, and I know what's wrong but I don't know what could be done because I can't reproduce the error and it's done in a container... It seems_ that snap doesn't go over the libs path /usr/lib/x86_64-linux-gnu/, because apt installs the missing lib there. but I don't know what causes this. I installed an Ubuntu VM and tried Ubnutu container inside that VM, Both recognized libevdev.so.2.

[0x0OZ]

I now have more time to work on this, Apart from the dead end for snap that I couldn't reproduce the bug and find a solution but it seems a docker thing? I have seen some issues that the libs installed during init weren't being recognized... Also it doesn't detect ssh connection and it requires some privileges to access /dev/input/event* files. because of this I think it would be better to make a stand-alone tool for an idle timer for every linux system.

This would be a better approach because of the permissions issue.

[AenBleidd]

because of this I think it would be better to make a stand-alone tool for an idle timer for every linux system.

And that is why we have this idea: #5226.

Since you're already have some experience with that, maybe you could work on this tool? IT would be also nice, if before starting a coding you could do an investigation first and present your solution in the ticket I just mentioned and afterwards start doing coding?

[0x0OZ]

because of this I think it would be better to make a stand-alone tool for an idle timer for every linux system.

And that is why we have this idea: #5226.

Since you're already have some experience with that, maybe you could work on this tool? IT would be also nice, if before starting a coding you could do an investigation first and present your solution in the ticket I just mentioned and afterwards start doing coding?

Yes, I was thinking of doing that myself, I saw this ticket which gave me some insight. After doing enough investigation I will share what I will build there which should solve the problem in all mentioned cases and be easily adaptable.

[AenBleidd]

@0x0OZ, nice! Thank you very much! In case you have any questions - feel free to reach out to me here or in BOINC Discord: https://discord.gg/wPRafUq