Indeed never used, coming from a template i suppose

No, there is no reason to not list a language for an existing po file, surely who added those files simply forgot to update LINGUAS. In some cases it could be me, maybe I added he and ps but I'm not sure :)

Closed by commit 4c31f653f5

It's not afaik. Looking into logs I see that the last real update was on 2014-01-29, it was committed by me and I received that file in e-intl mailing list from Páder Rezső <rezso@rezso.net>. I would edit the file including the author's name and setting copyright to the E team and license as the efl package. What do you think?

I tried setting "only dialogs with focused owner" and indeed with that the problem on notifications does not occur anymore. It still happens on desktop change, though, and anyway I would still like to have "all new windows focused" :-)

Why a fullscreen window should revert to unfullscreen when another window grabs focus? I do want that all new windows has focus, not only dialogs, and also want that dialogs can be on top even of the fullscreen window. This is the correct behaviour, but some times it works this way and other times it doesn't work anymore, I didn't realize in what particular situation (if there is any).
Also that doesn't explain why now it goes unfullscreen even on desktop switch. That shouldn't involve focuse change, even if on the desktop where rage is there is also a terminology window (normal stacking ad size).

I can confirm that it works now, thanks :)

Ok, I've never thought the problem could be that because they were all apps started at login and all at desktops different than 0, which of course is the current desktop at startup.
Now I've tried remembering gedit on desktop 0 (position, size and desktop) and libreoffice on desktop 4 (position, size and desktop) , and see that both apps are placed in the right position only if I'm on the desktop where they are started. So the answer is definitely: it happens with all apps but only if they are remembered on the non-current desktop.

I'll try some other apps. What do you mean as "non-current desktop"? Every desktop which is not 0? Or what else?

Yes, it's still an issue as I've just wiped out all window remembers to try, logged out, logged in and restored settings as I like but position is still not remembered.

My local git was from friday but looks like it was a few time before, as it's indeed solved now :)

Yes, I'm using git master. Before the patch I had errors both on EVP_CIPHER_CTX and EVP_MD_CTX, now it's only on EVP_MD_CTX. See the log attached (sorry for the italian localization):

I still get the same compilation error, not just deprecated warnings.

You're right, disabling that option make it works even with sna acceleration. So an option that should enhance the stability, or at least that was what I thought it's intended for, instead triggers an issue :)

Just done ;)

@raster
Yes, that option is enabled so fullscreen windows is not composited

Regarding my problem, I tried switching to software compositing and it doesn't occur. My card is an Intel GMA 4500M.

e17 is just the name of the dir under /opt where I install e and efl :)

Thanks for your info, I followed the instructions and made a backtrace, it's attached along with e-crashdump in case it's useful.
After quitting gdb I went back to the display where E was, and of course the white popup was also, and hit F1 but it did exit just like I've had choose F12.
Let me know if I can do something else.

Sorry for the late reply, I've had problems running xinit, probably due to a debian bug, and after a short digging I've had to setuid Xorg to make it able to open a virtual console where to run.
At the end it was all useless because under valgrind the problem does not occur :( I don't know how to help in other ways, sorry.

I don't know if it crashes, surely it doesn't segfault as no e-crashdump is generated. Should I run E through valgrind to dump some useful informations? How?