E26 crashes when pausing video

Just wanting to find out if this is just my machine or if it's an Enlightenment issue...i.e does anyone else see this too?

Occasionally the E desktop crashes (with a message about a saved .e-crashdump.txt file) and requires a restart with F1 or F2. If I do not do anything, I notice that all E window decorations are gone and that windows cannot be raised anymore but the menu bars (if available) do react to mouse clicks.

On my machine (running E26) I think it often happens when pausing video-playback on SMplayer and/or Firefox ....but only after a longer period of time. Again I'm not sure. :frowning:

Before reporting this bug to Enlightenment devs I'd like to make sure it's not Elive related or just my setup.


If you've used or are using E26, check if you've got a '.e-crashdump.txt' file on your home-directory.
If there's also a '.e-log.log' or 'e-log.log.old' of the same creation date as the crashdump there, I'd be quite interested in the output of "grep INH .e-log.log" or "grep INH .e-log.log.old"


  • make sure your crashdump file is removed before to try to trigger the issue, this file is important to report the issue to E but if you don't do that you may have it as an accumulative-mode, and you only want the last/actual error messages
  • install enlightenment-dbg and libefl-dbg to improve the crashdump data
  • GL or Software rendering behaves VERY differently, for stability is better to use the second one, but in any case both ones should be reported to E

I suggest to simply report the issue to E and see what happens :thinking:, include the crashdump data so its useful for them to find where the issue is

I wasn't sure if it was my machine hardware that was playing up or software. Hence me asking the crashdumps and logs.
In he mean time I've discovered that it's my hardware ... a bad connection on my screen. slowly getting worse :frowning:

It replaces the previous crashdump so it's best to save the file (including a date) before it gets overwritten by a newer one. The log files rotate after 2x time.

if is a bad connection with the screen I don't think there should be a crash anyways :thinking: (but makes sense that it can happen due to unexpected things) so still a bug to report

Actually it also made my CPU temp go skyhigh (up to 100˚C) when watching videos on Youtube.
Didn't happen if I used Smplayer to play the link, so I blamed bad HTML5 programming. :angry:

It's an integrated GPU (intel) and after dismantling and re-asserting all the screen connections everything seems alright again. No more crashing, overheating or freezing for the moment but it doesn't bode well for the laptop in question. :cry:

mmmh, sounds more related to GL, vsync, or the rendering engine of E... and yes it is an important issue

check your Compositor settings and play with all of them to see if you find the one causing the issue, remember to "ctrl + alt + end" to reload E before trying it

No more crashing, overheating or freezing

so bad, we cannot reproduce it in order to report the "E crashes when the screen connector fails" (which is a likely scenario) :slight_smile:

Hah, not that bad. The machine wouldn't even boot with the Mobo giving all sorts of error sounds.
I'm happy it's working again. :happybounce:

On top it doesn't make a difference whether on E16 or E26. :thinking:

1 Like

I have to retract that bit about the browser heating the CPU when playing Youtube videos, alas.
It's a bit better but still heats up to about 80˚C albeit it takes a while longer (4-5 min) to do so. :frowning_face:

Compositor engines plays the most important role in that consumption, how it is in E16 ? also disabling compositor in E16 will have a strong impact too, basically try with all possible options to comparison:

  • e16 with compositor (stable and default)
  • e16 without compositor (should be the fastest option, direct graphical card rendering without wrappers)
  • e26+ with software-rendering compositor (most stable option for new E)
  • e26+ with GL rendering compositor (should be faster on many laptops, especially if the GL is faster than the CPU, but can have some unstable behaviours)
  • e26 in GL mode but trying the different options (GL supports many options, which many of them are not available in software-mode like vertical-sync, these options affects speed and stability, just try every option to see differences)

heating cpu / consuming resources to play videos is a very important topic in elive, which aims to be as light as possible for all / any computers, so this benchmark is an important one

AFAIK it's down to to hardware acceleration i.e the browsers cannot do that (yet) while smplayer can access the hardware directly. :thinking:
So I can play with settings as you suggest but reality is that I do NOT have these issues on other (even lower specs) machines... but OTOH I do NOT use the 'other' machines as extensively as this particular one.

  • Next week I'll have more data on a very similar machine (similar Lenovo X1 carbon 7th Gen) and a newer machine (8th gen with a 3840x2160 screen).

The reality is that I can watch any video I like, for as long as I want, on any platform, without heating up ... as long it's not Youtube. :frowning_face:

I think they should, they use GL libraries and so they should render with accelerated gl too perfectly, the problem may be more like which type of rendering they are using and how it is communicating with the compositor layer

...and a newer machine (8th gen with a 3840x2160 screen).

I recently saw a very powerful laptop from a friend and I was in shock when I saw the cpu benchmarks that his i9 machine was actually 10x times faster than my work laptop lol

The reality is that I can watch any video I like, for as long as I want, on any platform, without heating up ... as long it's not Youtube.

Well I think you should report the issue to E, but since its a delicated topic I think the best is to record a video with your phone "showing" what is happening (showing htop resources or whatever gadget, showing how it incrases with the youtube but NOT with other players, adding machine specs like from "inxi -v7" result, etc.. and upload the video in a temporal unlisted youtube video to add to the report. Something is happening but we don't know if the issue is because of your hardware, the drivers, the kernel, or E, in any case these are questions to ask first:

  • does the issue happens with software-mode rendering too?
  • does the issue happens in E16 too?
  • what from a different kernel?
  • and especially, what from a bullseye build? which includes different drivers and libs of everything (version <= 3.8.27 )

It's not so much he browser but how Youtube handles the videos. (Note: I'm not sure yet in how far audio is part of the issue :thinking: )
For instance if I move to another source through the address bar, the heating immediately stops and even cools down.

I do this by changing i.e https://www.youtube.com/watch?v=bPymwLj_cG4 to use invidious by replacing 'youtube.com' with 'yewtu.be'.
Using the pager on Elive E16 minimal - Invidious

Where the link even pans out better in in this post :slight_smile:

everything goes in layers

  • youtube handle the video
  • video images are sent to the browser
  • the browser renders them
  • the desktop receives the rendering requests from the browser
  • the desktop sends the data to the compositor layer to handle the rendering requests
  • the compositor selects which driver to use, and how to render them
  • you can view the rendering results

Youtube clearly does a lot more than that, in ways of gathering data and tracking. Remember: You are NOT the customer there, you are the merchandise baited by the videos.

Rendering the very same videos, from the very same source (Youtube) through another carrier like 'invidious' or copying the link directly to 'smplayer' does NOT heat up the CPU.

After moving my SSD to another machine, this 'heating up' is still there, like on every other machine but maxes at around 60˚C in place of near 100˚C.... It clearly is/was a bad hardware connection caused by friction when opening/shutting the laptop lid.

The browsers settings can have a say in that, if wanted:

yes but we are talking only about the rendering issue, tracking and ads has no relation relation, especially since only happens in e26+

smplayer renders the video entirely different than youtube, as mentioned before the website of youtube renders it in a specific using html5 and/or other technologies that goes into the browser to render it, which goes to the compositor, etc... basically:

  • if the render issue happens in every desktop, the issue is on the browser
  • if happens in e26+ only, the issue is on the render of the desktop which is the compositor

not necessarily, a different machine is an entirely different driver and how is managed, the driver and/or how is managed by E26+ is likely where the issue relies on, did you tried the other options like booting on 3.8.27 (bullseye) and installing e26 to verify if happens too?

No, it was exactly the same machine from the same series built in the same year. Practically a carbon copy.

As said: It does not.

No, happens on E16 too... It really was the hardware.

Exactly my point but not specifically related to this video rendering issue. It is the root cause of higher than nesecary CPU temperatures, though.

The thing is that E26 has the thermometer directly visible in the ibar, that way a temperature rise is immediately visible. Previously on E16 I'd only notice when fans started spinning up....never knowing that the CPU got as hot as it did.

oh, that's interesting, in any case make sure it is really the same graphic card, maybe even an updated model or its firmware can behave differently, what about the bios? (updated firmware can fix a bug too), also check the results from commands like "inxi -v7":

  Device-1: Intel Skylake GT2 [HD Graphics 520] vendor: Lenovo driver: i915
    v: kernel arch: Gen-9 ports: active: eDP-1 empty: DP-1, DP-2, HDMI-A-1,
    HDMI-A-2 bus-ID: 00:02.0 chip-ID: 8086:1916 class-ID: 0300
  API: OpenGL v: 4.6 Mesa 22.3.6 renderer: Mesa Intel HD Graphics 520 (SKL
    GT2) direct-render: Yes

or the output of some commands and then "diff" it with the ones of the other machine, like:

  • xvinfo
  • glxinfo
  • vdpauinfo

note that the e17 widget for temperature was bad in UX terms, for example it looked to be high (high and red) when the temperature was normal, a critical temperature in red should be around 95C, orange maybe 85

I think you missed that:
After disconnecting, reconnecting and making sure all cables were well seated ..... the issue has gone away. It behaves like all the other laptops do.

It probably was a consequence of me often 'manhandling' the laptop by simply slapping the laptop shut and putting it aside if I needed to.
I'm more careful now and actually avoid closing the lid on that laptop. :slight_smile: