On Mon, 11 Jul 2022 07:05:57 +0200 Pierre Couderc via enlightenment-users
<enlightenment-users@lists.sourceforge.net> said:

> 
> On 7/9/22 09:00, Carsten Haitzler wrote:
> > On Sat, 9 Jul 2022 07:24:59 +0200 Pierre Couderc via enlightenment-users
> > <enlightenment-users@lists.sourceforge.net> said:
> >
> >> On 6/29/22 11:20, Carsten Haitzler wrote:
> >>> On Wed, 29 Jun 2022 10:19:12 +0200 Pierre Couderc via enlightenment-users
> >>> <enlightenment-users@lists.sourceforge.net> said:
> >>>
> >>>> in e 0.25.3 (installed from debian)
> >>>>
> >>>> When I start e, many times but not always, on the "second" screen,
> >>>> backlight is very low, I have to go in "Backlight Settings", change
> >>>> something then backlight turns back normal, the "something" could be
> >>>> changing Dim Backlight form 84 to 83, then  applying..
> >>> This can be a DDC backlight thing. Sometimes monitors when they are in the
> >>> middle of changing modes don't respond to DDC backlight change messages.
> >>> The backlight code does have some retry logic but it eventually gives up.
> >>> If you don't use the DDC stuff this wouldn't happen. It's kind of iffy
> >>> and your mileage varies based on who makes your monitor.
> >> OK, I disabble ddc ans I see what occurs...
> No more problem for the moment since I have removed ddc...
> >>
> >> (Sorry, I have not answerd before because I was away, with olny 1
> >> screen...)
> >>
> >>>> It seems to me that the saving of the previous state is not correctly
> >>>> reloaded. I have some other bugs of this kind. Particularly, if I try to
> >>>> exit e, I have a crash (see crashdump.txt below). And it seems to me
> >>>> that these bugs are  linked to the presence of 2 screens, the crash does
> >>>> not occurs with 1 screen.
> >>>>
> >>>>
> >>>> cf :  https://www.couderc.eu/download/.e-crashdump.txt
> >>> hmm it's almost always in the desktop gadgets module (gadman);
> >>>
> >>> Thread 1 (Thread 0x7f90f3635b40 (LWP 7625) "enlightenment"):
> >>> #0  0x00007f90f5417932 in pause () at /lib/x86_64-linux-gnu/libpthread.so.
> >>> #0 1  0x00007f90f5418200 in <signal handler called> ()
> >>> #at /lib/x86_64-linux-gnu/libpthread.so.0 2  0x00007f90dc12cab0 in  ()
> >>> at 
> >>> /usr/lib/x86_64-linux-gnu/enlightenment/modules/gadman/linux-gnu-x86_64-0.25.1/module.so
> >>>
> >>> etc. etc. ... try unload the gadman module?
> >>>
> >>> You don't have debug information so I can't say much other than "it seems
> >>> to often complain about some crash starting in gadman". No idea which
> >>> file, function or line to look at.
> >> OK, this is a "true" bug occuring at each e exit. How do I enable debug
> >> information...?
> > you read the code, add printf's int he code you're interested in looking at
> > what it's thinking and then read your ~/.e-log.log
> Sorry, I want to help by reporting errors, but I am not ready now to 
> enter e code and debug it...

that's how you get logs for something interesting. :) e otherwise logs what it
normally does to ~/.e-log.log

if e crashes and can grab a backtrace with gdb you get ~/.e-crashdump.txt and
it appends the latest backtrace there.

if you have crashes - the best thing to do is to rebuild efl and e with asan
(in arch i made efl-git-asan and enlightenment-git-asan AUR packages just for
this). Then ~/.e-log.log will contain nice asan dumps of where the memory
problem is. i highly suggest looking into asan (enlightenment.org has
information on how to build efl/e with asan - also applies to terminology, rage
etc. too)

> >> And how do I "unload gadman module" ?
> > settings -> extensions -> modules -> core -> gadgets ... press "unload"
> > below. you wont have desktop gadgets anymore. :)
> Unloading gadmen removes the crash.But iy turns back when I load it again...

??? unloading is saved. it should not load it again until you choose to load it
again unless something is very broken with saving config... ?

> _______________________________________________
> enlightenment-users mailing list
> enlightenment-users@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/enlightenment-users


-- 
------------- Codito, ergo sum - "I code, therefore I am" --------------
Carsten Haitzler - ras...@rasterman.com



_______________________________________________
enlightenment-users mailing list
enlightenment-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/enlightenment-users

Reply via email to