Account Options

  1. Sign in
The old Google Groups will be going away soon, but your browser is incompatible with the new version.
Google Groups Home
« Groups Home
v8-dev
Home
+ new post
About this group
Join this group
Message from discussion Better reporting of call stack overflows
The group you are posting to is a Usenet group. Messages posted to this group will make your email address visible to anyone on the Internet.
Your reply message has not been sent.
Your post was successful
 
From:
To:
Cc:
Followup To:
Add Cc | Add Followup-to | Edit Subject
Subject:
Validation:
For verification purposes please type the characters you see in the picture below or the numbers you hear by clicking the accessibility icon. Listen and type the numbers you hear
 
Dave Smith  
View profile  
 More options Nov 7 2012, 1:21 pm
From: Dave Smith <d...@singly.com>
Date: Wed, 7 Nov 2012 11:21:32 -0700
Local: Wed, Nov 7 2012 1:21 pm
Subject: Re: [v8-dev] Better reporting of call stack overflows
Print | View thread | Show original | Report this message | Find messages by this author
Hi Yang,

Thanks for the update. I've applied the patch against node 0.8.14's
embedded version of V8 (which is, I think, 3.11.10.25) -- the patch
applied mostly cleanly except one reference to "global_object()" in
isolate.cc, which seems to be just a new name  for "global()".

For my simplest test case, the patch definitely helps. However for the
production test case, I'm still not seeing any stack trace. A bit of
digging reveals that it's exiting here:

+  Handle<Object> error = GetProperty(global(), "Error");
+  if (!error->IsJSObject()) return Failure::Exception();

So it seems like there are still situations where no stack trace will
be provided -- can you elaborate on those a bit, please?

Thanks,

D.

On Wed, Nov 7, 2012 at 4:05 AM, Yang Guo <yang...@chromium.org> wrote:
> Hi Dave,

> I completed a patch fixing this problem. It's currently under review:
> http://codereview.chromium.org/11275186/

> You are welcome to point out any short-comings in this patch.

> The problem with your approach is that it reimplements the stack trace
> formatting in C++ while we already have it in javascript. It is also not
> compatible with our stack trace API.

> Regards,

> Yang

> On Mon, Nov 5, 2012 at 1:27 PM, Dave Smith <d...@singly.com> wrote:

>> Hi Yang,

>> Thanks for the feedback. If you have the time/inclination to describe
>> (roughly) the sort of patch you'd like to see and the subtle problems,
>> I'd be happy to revise my patch and try to help move things forward.

>> Thanks,

>> d.

>> On Mon, Nov 5, 2012 at 4:37 AM, Yang Guo <yang...@chromium.org> wrote:
>> > (previously sent without actually finishing the mail...)

>> > Hi Dave,

>> > I wasn't aware of this problem before. We currently already have
>> > redundant
>> > ways to capture stack trace (at throw site and at creation site of the
>> > Error
>> > object). I've been wanting to consolidate this, which would also enable
>> > me
>> > to solve this issue.

>> > Your current approach has some subtle problems. If you have some
>> > patience,
>> > as I will be working on this issue, it will be solved eventually. I'll
>> > file
>> > a bug for this.

>> > Yang

>> > On Mon, Nov 5, 2012 at 12:36 PM, Yang Guo <yang...@chromium.org> wrote:

>> >> Hi Dave,

>> >> I wasn't aware of this problem before. We currently already have
>> >> redundant
>> >> ways to capture stack trace (at throw site and at creation site of the
>> >> Error
>> >> object). I wanted to consolidate those

>> >> On Sat, Nov 3, 2012 at 4:02 AM, Dave Smith <diz...@singly.com> wrote:

>> >>> Hi,

>> >>> I've been using node.js (and hence, v8) for a little bit and one thing
>> >>> that has driven me nuts is the lack of reporting when the call stack
>> >>> overflows. So, in keeping with open source tradition, I've knocked
>> >>> together
>> >>> a (simplistic) patch to fix this specific problem. My patch is here:

>> >>> https://github.com/dizzyd/node/commit/40434019540ffc17e984ff0653500a3...

>> >>> (note that it's against the fork of the node codebase, which embeds v8
>> >>> directly).

>> >>> My lack of familiarity with v8 is probably grossly evident, but
>> >>> perhaps
>> >>> someone could provide some direction on improvements if this is an
>> >>> itch
>> >>> bothering other people. :)

>> >>> Thanks!

>> >>> D.

>> >>> --
>> >>> v8-dev mailing list
>> >>> v8-dev@googlegroups.com
>> >>> http://groups.google.com/group/v8-dev


 
You must Sign in before you can post messages.
To post a message you must first join this group.
Please update your nickname on the subscription settings page before posting.
You do not have the permission required to post.
Create a group - Google Groups - Google Home - Terms of Service - Privacy Policy
©2013 Google