[Edbrowse-dev] seg fault in edbrowse proper

[Edbrowse-dev] seg fault in edbrowse proper

[Karl Dahlke]

Not the js process but edbrowse itself; though you have to have js active to get the seg fault.

http://www.rawstory.com/2017/01/keith-olbermann-pleads-with-trump-supporters-his-illness-is-putting-your-family-at-risk

Karl Dahlke

Re: [Edbrowse-dev] seg fault in edbrowse proper

[Kevin Carhart]

I pulled up the core dump from the issue raised by rawstory.com but am not 
sure what it means.  I think there is a bad string which can't be 
processed by variants of printf.  I don't know what the entry point is 
from edbrowse files to system files.

After the first few below, it gives short entries like 4,5,6, forever.



#0  __GI__IO_default_xsputn (f=0x7ffda0961660, data=0x456080, n=0) at 
genops.c:448
#1  0x00007f60f66ad867 in _IO_vfprintf_internal (s=s@entry=0x7ffda0961660,
     format=format@entry=0x456080 "%c%d*", ap=ap@entry=0x7ffda0961778) at 
vfprintf.c:1334
#2  0x00007f60f66d59f4 in __IO_vsprintf (string=0x7ffda0961860 
"`>\346\001",
     format=0x456080 "%c%d*", args=args@entry=0x7ffda0961778) at 
iovsprintf.c:42
#3  0x00007f60f66b8127 in __sprintf (s=<optimized out>, format=<optimized 
out>) at sprintf.c:32
#4  0x000000000042a39b in ?? ()
#5  0x000000000042a6d8 in ?? ()
#6  0x000000000043ccc7 in ?? ()





  On Tue, 10 Jan 2017, Karl Dahlke wrote:

> Not the js process but edbrowse itself; though you have to have js active to get the seg fault.
>
> http://www.rawstory.com/2017/01/keith-olbermann-pleads-with-trump-supporters-his-illness-is-putting-your-family-at-risk
>
> Karl Dahlke
> _______________________________________________
> Edbrowse-dev mailing list
> Edbrowse-dev@lists.the-brannons.com
> http://lists.the-brannons.com/mailman/listinfo/edbrowse-dev
>

--------
Kevin Carhart * 415 225 5306 * The Ten Ninety Nihilists

Re: [Edbrowse-dev] seg fault in edbrowse proper

[Chris Brannon]

Even more interesting, I cannot reproduce this at debug level 5.

-- Chris

Re: [Edbrowse-dev] seg fault in edbrowse proper

[Kevin Carhart]

I think I noticed this.  I think the "kill javascript on any javascript 
runtime" code from ebjs is triggered prior to the seg fault.  But it's 
tricky to grasp what happened, because of what we've been saying in the 
recent messages about this ebjs code: it gets buried by the rest of the 
db5 output.  But I think Karl just went ahead with changing this to quit 
eb immediately.  Now, I think if you go through these motions (javascript 
disabled but no seg fault) you actually 
can indeed read Raw Story articles.  So it is sort of like the right thing 
happening for the wrong reasons.  It still could be seen as a good thing 
but it is capricious, because Raw Story did not happen to hold their plain 
text hostage to successful JS completion, but unfortunately a lot of other 
sites do.

Kevin