From: Adam Thompson <arthompson1990@gmail.com>
To: Karl Dahlke <eklhad@comcast.net>
Cc: edbrowse-dev@edbrowse.org
Subject: Re: [edbrowse-dev] Abort from within duktape when visiting...
Date: Sun, 8 Dec 2019 16:57:16 +0000 [thread overview]
Message-ID: <20191208165716.GG194728@toaster> (raw)
In-Reply-To: <20191108102636.eklhad@comcast.net>
On Sun, Dec 08, 2019 at 10:26:36AM -0500, Karl Dahlke wrote:
> Yes, we didn't need the 20 thousand lines of output, just the last line, as it does what I thought it might be doing, starting a low level calland not finishing it.
> core dumps and gdb won't help, even if you did have the duktape symbols, it is a deliberate abort because of some runtime issue not protected.
Yeah, but when I first saw the error I wasn't sure where it'd blown up. Anyway that
was my reading of the output.
> I've sometimes thought I have to protect *every* duktape call we make, but so much overhead, in software and in performance.
> Not sure what to do about that.
Hmmm, not sure. I can see something to be said for protecting all the
duktape calls but I'm not sure how we can contain them in something where we
can avoid all the extra code. How large is the overhead of these in
reality?
> As a verification you can comment out html.c line 926 and it runs.
> I don't just want to protect the call set_property_string, though I may have to, I'd really like to know what is going wrong here.
> In jdb, try setting document.readyState = "complete"
> This is protected so won't core dump, but neither does it show any errors, even with dberr, and neither does it change the value of readyState,
> so the web page has done something to it.
> either it's not a string any more (getter setter), or it is not writeable.
Should that be allowed or can we make that something only we can change?
From what I've read I'm not sure why scripts would be touching this or if
they should even be allowed to.
Cheers,
Adam.
next prev parent reply other threads:[~2019-12-08 16:57 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-12-07 17:28 [edbrowse-dev] Abort from within duktape when visiting https://www.digicatapult.org.uk/careers Adam Thompson
2019-12-07 22:14 ` Karl Dahlke
2019-12-08 13:53 ` Adam Thompson
2019-12-08 15:26 ` [edbrowse-dev] Abort from within duktape when visiting Karl Dahlke
2019-12-08 16:57 ` Adam Thompson [this message]
2019-12-09 10:41 ` Karl Dahlke
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20191208165716.GG194728@toaster \
--to=arthompson1990@gmail.com \
--cc=edbrowse-dev@edbrowse.org \
--cc=eklhad@comcast.net \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).