Browsers keep crashing
- Vous devez vous identifier ou créer un compte pour écrire des commentaires
Hi all,
After updating Trisquel 7 today, both Abrowser and IceCat suddenly crash when a few webpages are opened.
One in particular is GoogleMaps
Another is a the the webpage of the Jornal of Neurology, Neurosurgery & Psychiatry ( www.jnnp.bmj.com )
Has anyone experienced the same? Are you able to open both pages?
A bit frustrating... :-( I am considering reverting to my latest backup before the (highly recommended security) update...
Update:
- abrowser had no add-ons installed.
- I completely uninstalled abrowser+icecat; sudo apt-get update, update, autoremove.
- rebooted, installed just abrowser and rebooted.
- tried the fresh installation of abrowser; websites mentioned continue crashing.
- I then checked the "disable javascript" option from abrowser --> was able to open the JNNP website, but for obvious reason can't use googlemaps or (worse in my case) tinytinyRSS (which is >50% of what my computer time consists of...)
Next I wanted to try uninstalling and re-installing the javascript components of Trisquel; any idea about the packages involved?
thanks!
Sometimes LibreJS addon crash some browsers. I never could to get acceptable performance with LibreJS, I tryed it on Abrowser, IceCat, Iceweasel. I use NoScript.
An altenative to Google Maps is
https://www.openstreetmap.org/
and allso Marble a Virtual Globe with open street map
https://trisquel.info/fr/wiki/marble-un-globe-virtuel
"It works great here" (TM).
Thank you, I appreciate the input.
Trying to clarify:
- I do not have add-ons installed; I used the "disable javascript" feature from abrowser itself (from about:home)... I don't have LibreJS installed.
- Even though openstreetmap is an alternative, I am still unable to open JNNP's website.
- 24 hours ago I was able to browse through JNNP's website normally.
- After yesterday's system/security updates, every time I try to open JNNP's website, the browsers crash.
- However, if I "disable javascript" from abrowser initial page, JNNP loads.
- ...which makes me think that the update somehow messed up some javascript component; that's why I was interested in trying to reinstall it.
- The option of disabling javascript causes further problems as it is necessary to open tinytinyRSS.
No one so far told me if s/he is able is having any trouble opening JNNP or GoogleMaps (privacy concerns -> Tor is an easy workaround)
Solutions that I currently have:
- Modify my browser habits and alternate disabling/enabling javascript depending on the webpages I am trying to load (***very annoying***) --> not an option in the long run...
- Format the system and do a fresh installation of trisquel, and see if the problem goes away
Any further suggestions?
The neuro surgery site works fine for me (js disabled).
I use noscript to block all scripts (and several other possible issues - noscript is not only a script blocker, it is much more).
EDIT: I shall add that the browser is the seamonkey.
veeeery interesting... I just downloaded seamonkey 2.40 64 bit version, but didn't install the noscript add-on.
seamonkey crashed on loading jnnp.bmj.com :/
on the other hand, googlemaps loaded just fine.... mysteries of nature...
As an extra suggestion: Try switching from TinyTinyRSS to another feed
reader, like Liferea, that's already installed on your copy of Trisquel
(it can be found on Main menu → Internet → "Liferea news reader" or
something similar).
Update: just started using Tor Browser directly from Tor Project's website; works like a charm, no crashes... maybe I will just forget about IceCat/abrowser and only use Tor Browser from now on ;)
The JavaScript engine comes with the Web browser. There are "libmozjs-*" packages but the Web browser does not depend on any of them.
I would recommend not to use IceCat. It doesn't receive timely security updates, and is at version 38.8.0, which has been unsupported for months.
> I would recommend not to use IceCat. It doesn't receive timely security updates, and is at version 38.8.0, which has been unsupported for months.
Same for Seamonkey, but fortunately you can download the latest aurora build, which is FF 51 (minus the bloat).
http://www.seamonkey-project.org/dev/aurora
It is super-stable here, it never hanged/crushed, not once.
I never tried Seamonkey, although I vaguely remember using the Mozilla Suite a long time ago. It's funny that Seamonkey would be less bloated than Firefox despite having more functionality. I use Iceweasel (actually, Firefox + Iceweasel branding extension) and it works well for me, but I don't like the recent feature creep.
Legimet, all the useless anti-feature crappy bloat inserted by FF in the last 4 years is not in the monkey. Yes, the monkey is the direct descendant of the Netscape suite, it has a lot more functionality but 20 times (yes, twenty) less lines of code. On my 2004 laptop FF opens in 30 seconds, Seamonkey in 10. A heavy website (I am thinking of a diaspora tier of gifs) is a pain in the rectum with FF, a smooth pleasure with the monkey. The monkey is much more configurable and customizable.
I think the 20x number is wrong. The repository for Seamonkey does not contain the Gecko code, and you have to checkout the main Mozilla repo (in addition to the Seamonkey/Thunderbird one) to build Seamonkey from Mercurial.
https://developer.mozilla.org/en-US/docs/Mozilla/Developer_guide/Source_Code/Getting_comm-central#Initial_checkout
14017563 / 693285 = 20
https://www.openhub.net/p/firefox
https://www.openhub.net/p/seamonkey
If the openhub website gets it, the math gets it.
If you like functional and smooth, if you don't like bloat, the monkey will be very good to you :)
Read my comment again. The comm-central repository (which openhub is looking at for Seamonkey, with 700,000 lines) does not contain the Gecko code, which has to be fetched from the mozilla-central repository (the Firefox one with 14 million lines) to build Seamonkey. I don't doubt that Seamonkey is less bloated but it's not 20x less code.
Basically, you're comparing the non-Gecko parts of Seamonkey with all of Firefox, including Gecko. That's not a fair comparison to make.
Legimet, I thought openhub lists the **total and exact** amount of LOC of a package.
I ran cloc on the comm-central repository and got the following:
6622 text files.
6369 unique files.
1761 files ignored.
github.com/AlDanial/cloc v 1.68 T=19.10 s (254.7 files/s, 62548.9 lines/s)
---------------------------------------------------------------------------------------
Language files blank comment code
---------------------------------------------------------------------------------------
JavaScript 1715 63888 96296 339046
C++ 440 42833 30365 241642
CSS 653 16794 5946 65239
XML 125 5882 1904 58221
C 105 9420 10691 41145
C/C++ Header 532 11510 14591 33453
IDL 317 4387 0 25811
XHTML 62 3617 231 19697
DTD 389 1830 2566 11400
Python 62 1701 4353 7052
HTML 123 417 110 4062
NAnt script 206 849 0 4054
JSON 11 0 0 3528
Objective C++ 7 459 162 2237
Bourne Shell 11 87 255 1449
INI 65 119 62 1248
Perl 9 567 85 1119
make 17 254 368 789
Windows Module Definition 7 24 156 354
LESS 1 33 52 183
Windows Resource File 5 18 29 76
Markdown 1 10 0 49
Ant 1 6 5 49
m4 1 3 3 21
CMake 1 2 5 0
---------------------------------------------------------------------------------------
SUM: 4866 164710 168235 861924
---------------------------------------------------------------------------------------
cloc may be counting more lines as "code" than openhub, but it is on the same order. For some reason, sloccount isn't counting JS. But anyway, OpenHub is only looking at that repository so that number does not consider the code of Gecko.
***SOLVED***
Hi all,
The issue went away after I manually updated the kernel ( https://trisquel.info/en/wiki/update-linux-libre-kernel )
Life is back to normal - yay :-)
good for you ^^
- Vous devez vous identifier ou créer un compte pour écrire des commentaires