LXQt Memory Usage On Par With LXDE, Loiwer Than Xfce

[mir3x]

Hello, I'm the author of the "joke".
Would you like to share some ideas to help future improvement?
Which part specifically are missing and what's your suggestions?
Thanks!

Hi. Great work.
I would move filter QLineEdit to right of location QLineEdit, bc it takes too much space on bottom.
And I think hitting Esc on filter should clean filter.

Also it could be nice to specify background image or color to default and places view.
And option to preserve settings to any modified folder would be cool too.

[PLum]

For webwrowsing checkout palemoon its based on old firefox 24+... but it works nice and dont eat much memory (on my old T61 it works perfect I can use about 70tabs without any issue )

[sverris]

Hello, I'm the author of the "joke".
Would you like to share some ideas to help future improvement?
Which part specifically are missing and what's your suggestions?
Thanks!

I use the 'joke' for years - and I am really happy with it...

[starshipeleven]

1) No, we don't need realistic tests because every person in the world has a different workflow and there's no way in the universe you can measure everything for everyone.

So, you are telling me that people don't power up the PC and don't just open a FM?

And that if you give them totally synthetic data with caches wiped (what is there to wipe just after boot, but anyway) it's better yes, as people can integrate that in their excel sheet to number-crunch it to get the data for their workflow.

2) Do you understand that certain boot up services/application/whatever else may pollute cache and never be actually used for anything involving a running DE session? We are measuring a DE session, now whatever fuq you may run in background. Also see point 1.

They did a default install dumbass, that's the default stuff in there, what most people use.

That's why you test with compositing off.

He tested with compositing off in LXQt and xfce as by default it's off there.
Cinnamon and GNOME3 are supposed to work with compositing, you cannot claim the test is worth shit if you test Cinnamnon in "fallback mode" or GNOME3 in the same situation.

Also, KWin is known to leak memory,

None of the tested DEs were using Kwin, nice trolling on KDE on a thread that has nothing to do with it.

and as of yesterday I discovered a huge memory leak

Bug report or it didn't happen.

[caligula]

Hello, I'm the author of the "joke".
Would you like to share some ideas to help future improvement?
Which part specifically are missing and what's your suggestions?
Thanks!

Just a suggestion, but for me the spacing in pcmanfm is too tight. I prefer the one in thunar. For instance the sidebar would look a lot better with more spacing around the category titles and with bold text. Maybe it's just me who likes it that way. Also the range of icon sizes is too limiting for large high dpi displays. Why 20x20 - 96x96. My screen has > 117 dpi. So basically the smallest icon is 0.1 x 0.1 inches and the largest is 0.8 x 0.8

[dungeon]

The way he "measured" the memory consumption is wrong.

This is how it must have been done:

1) Fire up VirtualBox with your favorite DE - that's OK
2) Then open an xterm window and run echo 3 | sudo tee /proc/sys/vm/drop_caches
3) Run `free -mh` as usual and calculate Memory allocated to your VM minus Free RAM

This value will represent the actual memory consumption.

Michael, could you actually compare different DEs in several modern distros of your choice using the outlined method? Say Ubuntu 16.XX/Fedora 24/Linux Mint/etc.

It is not only minus free there , most Linux GUIs measures it like this:

Total - (Free + Buffers + Cached)

And that is it... but some like to measure it probably in some Windows-like i guess so cut Slab too, so:

Total - (Free + Buffers + Cached + Slab)

[birdie]

So, you are telling me that people don't power up the PC and don't just open a FM?

And that if you give them totally synthetic data with caches wiped (what is there to wipe just after boot, but anyway) it's better yes, as people can integrate that in their excel sheet to number-crunch it to get the data for their workflow.

They did a default install dumbass, that's the default stuff in there, what most people use.

He tested with compositing off in LXQt and xfce as by default it's off there.
Cinnamon and GNOME3 are supposed to work with compositing, you cannot claim the test is worth shit if you test Cinnamnon in "fallback mode" or GNOME3 in the same situation.

None of the tested DEs were using Kwin, nice trolling on KDE on a thread that has nothing to do with it.

Bug report or it didn't happen.

Everyone around me (we are all IT people) doesn't use any graphical file managers - they are all dog slow, browsing is counter intuitive and involves mouse. You could do the same in bash/zsh/ksh/mc 10 times faster.

Default install of what? For whom, Re Tardo? Do you know anything about their installation options? Stop blabbering.

So, Gnome3/Cinnamon are barely usable without compositing? What a load of shat.

Re Tardo, I proposed a test for Michael - that's why I was talking about KDE.

My friend has better things to do than reporting bugs. A workaround I offered him works just fine: he added `killall compiz` to his crontab and now his compiz hovers around 100MB instead of over 400MB it had grown to recently.

It's amazing how you're trying hard to argue with me, however you're just talking from a different PoV, so you're not actually contradicting. I lol'ed.

[birdie]

Total - (Free + Buffers + Cached)

That's incorrect because some buffers and cache cannot be freed (for instance they back loaded binaries or libraries so there's no way the kernel can free this memory).

[dungeon]

Incorrect or not, that is how GUIs do it so that is what user see, of course you can always look at:

cat /proc/meminfo

And calculate things as you like

[Mavman]

You can talk whatever you want, but gtk2 is unbeatable

Yes, but Windows 3.11 is even better. Gtk2 is mostly dead... Either go with qt or whit gtk3

Also you are running test on 32bit distro, most does not use that ...

I believe most users of lxqt are 32 bit users... But I might be wrong. I surely have one and I am very very happy with lxqt and wouldn't go back to gtk version even if someone would pay me. Lxqt has so many great things added...