On Tue, Feb 17, 2009 at 10:31 AM, Evan Martin <[email protected]> wrote:
> On Tue, Feb 17, 2009 at 10:20 AM, Darin Fisher <[email protected]> wrote: > > What are the reasons to not like V8 in the browser? > > To summarize the arguments: > - reduce memory fragmentation in browser by not creating a V8 heap in > the browser for PAC processing > - reduce browser's dependencies (purely from a build perspective). > That significantly grows our test executables. > - that one other reason Marc-Antoine mentioned off-list that is > security-sensitive > - forces the browser process 32-bit, which is a huge pain to install > on 64-bit Ubuntu (you have to manually unpack .deb files and create > symlinks in /usr/lib, etc.). I'm very confused by this one. The browser process already calls on some webkit functions, which means that it brings in code that depends on V8. Also, I assume you intend to ship a static binary that links in most of our code as we do on Windows. Experience with Firefox on Linux showed this to be a huge gain for Firefox as well. I'd be very surprised if it were not. Could you ease some of the installation pain by having a 64-bit stub executable? -Darin > > > If the ship is already heading this direction it's perhaps not worth > trying to turn it around. I was just selfishly hoping for the last > reason. For an example, Dean found that the reason we weren't able to > render Flash on our 64-bit boxes but could on our 32-bit ones was that > the 64-bit and 32-bit libgtk packages had skewed to different > versions. > --~--~---------~--~----~------------~-------~--~----~ Chromium Developers mailing list: [email protected] View archives, change email options, or unsubscribe: http://groups.google.com/group/chromium-dev -~----------~----~----~----~------~----~------~--~---
