Source code to Awesomium?

[Falagard]

I think that I'm hitting some VS 2008 problems in release mode that might have to do with STL.

In debug it works, but I compiled in release to see if I could get transparency working (I'm hitting a crash when setting transparency on - it may be an assert as you said, but it's being reported as an unhandled exception - access violation from somewhere in Awesomium_d.dll) and in release the first time a callback is called it's crashing when I attempt to access JSArguments passed to the callback.

I'm only passing one argument to the callback, but the args.size() returns 4255283963 and the array contains garbage data.

What's the ETA on releasing the source code for Awesomium so we can compile it in VS 2008 and look at some of the bugs ourselves?

Particularly for our project we need cookie and cache written to disk so we may add that feature ourselves, but also it'd be helpful to be able to find and fix bugs as well.

As I might have mentioned before, we have a fairly huge project already using Astral extensively and would like to move it over to Awesomium as soon as possible (or at least start moving it over and see if we hit any major roadblocks that prevent us from using it).

[adam]

I will be releasing the source code when I get home today (I'm writing this from class). Be aware that you will need to have a Jan. 9th checkout of Chromium (I built it using VS2005 but they seem to have begun supporting VS2008 to some degree) to build Awesomium from source.

EDIT: Quite unfortunately, it seems my apartment's internet is down and so I'll have to delay the commit until it comes back up (hopefully later tonite or tomorrow).

[Falagard]

Have you made many changes to Chromium itself?

I'm currently getting both Chromium trunk as well as the last updated version of Chromium on Jan 9th.

[adam]

Have you made many changes to Chromium itself?

I'm currently getting both Chromium trunk as well as the last updated version of Chromium on Jan 9th.

Nope-- Awesomium (at this time) requires no changes/patches to Chromium. Just fully build Chrome (both debug and release) and you should be ready to go.

Oh also, the Awesomium solution expects a directory structure of something like this:

Code: Select all

|
-- Awesomium
|   |
|   -- Awesomium.sln
-- chromium
    |
    -- chromium
       |
       -- src

[Falagard]

Excellent, that'll make it easier to keep Chromium up to date.

[adam]

Okay, the source has been released: http://princeofcode.com/forums/viewtopic.php?f=8&t=31

[Cutter]

First of all, great work, adam! I have updated Navi from pre-Astral to the newest svn version and it`s really smooth to use!

But I have the same problem as Falagard - an event called from Javascript with one parameter works fine in Debug Mode, but crashes in Release Mode (the vector is full of garbage). Are there any news on this?

[adam]

First of all, great work, adam! I have updated Navi from pre-Astral to the newest svn version and it`s really smooth to use!

But I have the same problem as Falagard - an event called from Javascript with one parameter works fine in Debug Mode, but crashes in Release Mode (the vector is full of garbage). Are there any news on this?

Thanks!

As far as the issue with WebViewListener::onCallback, it seems to be confined to appearing only in MSVC9 Release. I don't have a solution but I have a theory: because the SDK is built with MSVC8 with static CRT, it could be that Awesomium's version of std::vector<T> is different from the hosting application's version and so we're experiencing undefined behavior (garbage values being returned and such).

To prove/disprove this theory, we just need someone to build Chromium and Awesomium using nothing but MSVC9 (which I think Falagard just did in another thread) and test if the issue still persists.

[Falagard]

I haven't compiled Chromium in release yet, but I was thinking along the same lines in regards to the std::vector template. I'll compile in VC9 release and let you know if it fixes it.

[adam]

I haven't compiled Chromium in release yet, but I was thinking along the same lines in regards to the std::vector template. I'll compile in VC9 release and let you know if it fixes it.

Thanks.