chrome browser issues

classic Classic list List threaded Threaded
19 messages Options
Reply | Threaded
Open this post in threaded view
|

chrome browser issues

Amit Kulkarni-5
As part of testing for 4.9 beta, I am just checking out chrome for
amd64. (yes everything is up to date before adding chromium) I just
pkg_add chromium. It gave me a dialog to import settings and make
default browser. I asked it to import settings from firefox and not
set it as default browser. Afterwards, it consistently dumps core. Is
anybody else getting it?

Thanks

$ chrome &
[1] 26879
$ Called PathProviderLinux onceCalled PathProviderLinux
once[0201/114020:WARNING:base/debug_util_posix.cc(230)] Don't know how
to do this
Called PathProviderLinux onceCalled PathProviderLinux
once[26879:-238776:66977413932:FATAL:chrome/common/notification_registrar.cc(102)]
Check failed: current_thread_id == thread_id. called on invalid
thread: 243003904 vs. 131972160

[1] + Abort trap           chrome  (core dumped)
$

Reply | Threaded
Open this post in threaded view
|

Re: chrome browser issues

Landry Breuil-6
On Tue, Feb 01, 2011 at 11:46:17AM -0600, Amit Kulkarni wrote:
> As part of testing for 4.9 beta, I am just checking out chrome for
> amd64. (yes everything is up to date before adding chromium) I just
> pkg_add chromium. It gave me a dialog to import settings and make
> default browser. I asked it to import settings from firefox and not
> set it as default browser. Afterwards, it consistently dumps core. Is
> anybody else getting it?

Chromium never worked perfectly fine on OpenBSD. Now it's badly
outdated, and should have been unlinked from the build..

Landry

Reply | Threaded
Open this post in threaded view
|

Re: chrome browser issues

Robert Nagy
I have a WIP port of a new chromium version so be patient a bit.

On (2011-02-01 22:45), Landry Breuil wrote:

> On Tue, Feb 01, 2011 at 11:46:17AM -0600, Amit Kulkarni wrote:
> > As part of testing for 4.9 beta, I am just checking out chrome for
> > amd64. (yes everything is up to date before adding chromium) I just
> > pkg_add chromium. It gave me a dialog to import settings and make
> > default browser. I asked it to import settings from firefox and not
> > set it as default browser. Afterwards, it consistently dumps core. Is
> > anybody else getting it?
>
> Chromium never worked perfectly fine on OpenBSD. Now it's badly
> outdated, and should have been unlinked from the build..
>
> Landry
>

Reply | Threaded
Open this post in threaded view
|

Re: chrome browser issues

Brad Smith-14
On Tue, Feb 01, 2011 at 11:54:30PM +0100, Robert Nagy wrote:

> I have a WIP port of a new chromium version so be patient a bit.
>
> On (2011-02-01 22:45), Landry Breuil wrote:
> > On Tue, Feb 01, 2011 at 11:46:17AM -0600, Amit Kulkarni wrote:
> > > As part of testing for 4.9 beta, I am just checking out chrome for
> > > amd64. (yes everything is up to date before adding chromium) I just
> > > pkg_add chromium. It gave me a dialog to import settings and make
> > > default browser. I asked it to import settings from firefox and not
> > > set it as default browser. Afterwards, it consistently dumps core. Is
> > > anybody else getting it?
> >
> > Chromium never worked perfectly fine on OpenBSD. Now it's badly
> > outdated, and should have been unlinked from the build..
> >
> > Landry

Sure, after unlock, but shipping packages *now* which are known to be
broken on amd64 is of no use to anyone. Just remove amd64 from
ONLY_FOR_ARCHS and re-add it with your WIP update if it works
properly.

--
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.

Reply | Threaded
Open this post in threaded view
|

Chromium 9.0.597 packages for testing

Robert Nagy
In reply to this post by Amit Kulkarni-5
Chrome 9.0.597 packages for testing:

http://blade2k.humppa.hu/chromium/

The port update is useless for most of the people because you need
an uncommited gcc patch. So you can't build it.

Reply | Threaded
Open this post in threaded view
|

Re: Chromium 9.0.597 packages for testing

Markus Hennecke
On 02/03/11 10:45, Robert Nagy wrote:
> Chrome 9.0.597 packages for testing:
>
> http://blade2k.humppa.hu/chromium/
>
> The port update is useless for most of the people because you need
> an uncommited gcc patch. So you can't build it.

This looks really nice on amd64. I had to upper the limit of fds to make
it display websites with a lot of elements (256 works, 128 didn't) even
when using only one or two tabs.

I'm really looking forward to see this as a port.

Kind regards
   Markus


Reply | Threaded
Open this post in threaded view
|

Re: Chromium 9.0.597 packages for testing

Amit Kulkarni-5
Er, I know its irrelevant for some (maybe all of you) but you need to
do a pkg_add of libxslt + libvpx from current packages before you can
check out chromium.

I made kern.shminfo.shmall=65536 from the default 8192 (I might drop
it down to recommended 32768 if problems crop up) and the login.conf
limits are adequate in my case.

Jumping straight to the most stable chrome available is a dream. I
just did a normal browse for half an hour or so. Checked javascript
websites etc. We need more alternative browsers for end users, gives
more bragging rights. I see robert@ made the 5.0.XXX version only for
i386, looks like it isn't possible for this updated port to go in for
4.9 because of the possible gcc issues. Would somebody comment on this
one?

Thanks

>> The port update is useless for most of the people because you need
>> an uncommited gcc patch. So you can't build it.
>
> This looks really nice on amd64. I had to upper the limit of fds to make it
> display websites with a lot of elements (256 works, 128 didn't) even when
> using only one or two tabs.
>
> I'm really looking forward to see this as a port.

Reply | Threaded
Open this post in threaded view
|

Re: Chromium 9.0.597 packages for testing

Robert Nagy
The issue is that
#praga pop and push is not supported in our in-tree gcc and it has
to be enabled, otherwise chrome will have a lot of unaligned code
compiled and you will have SIGBUSes all the time.

Index: gcc/config/openbsd.h
===================================================================
RCS file: /cvs/src/gnu/gcc/gcc/config/openbsd.h,v
retrieving revision 1.5
diff -u -r1.5 openbsd.h
--- gcc/config/openbsd.h        11 Sep 2010 15:13:17 -0000      1.5
+++ gcc/config/openbsd.h        3 Feb 2011 18:44:01 -0000
@@ -325,6 +325,9 @@
    as this depends on a few other details as well...  */
 #define HANDLE_SYSV_PRAGMA 1
 
+/* Define this so we can compile MS code */
+#define HANDLE_PRAGMA_PACK_PUSH_POP 1
+
 /* Stack is explicitly denied execution rights on OpenBSD platforms.  */
 #define ENABLE_EXECUTE_STACK                                           \
 extern void __enable_execute_stack (void *);                           \

--

If you want you can compile chrome by yourself: https://github.com/rnagy/ports

On (2011-02-03 19:37), Marc Espie wrote:
> Hey, can I get the gcc patch ? I think I haven't got it on hackers@...
On (2011-02-03 12:21), Amit Kulkarni wrote:

> Er, I know its irrelevant for some (maybe all of you) but you need to
> do a pkg_add of libxslt + libvpx from current packages before you can
> check out chromium.
>
> I made kern.shminfo.shmall=65536 from the default 8192 (I might drop
> it down to recommended 32768 if problems crop up) and the login.conf
> limits are adequate in my case.
>
> Jumping straight to the most stable chrome available is a dream. I
> just did a normal browse for half an hour or so. Checked javascript
> websites etc. We need more alternative browsers for end users, gives
> more bragging rights. I see robert@ made the 5.0.XXX version only for
> i386, looks like it isn't possible for this updated port to go in for
> 4.9 because of the possible gcc issues. Would somebody comment on this
> one?
>
> Thanks
>
> >> The port update is useless for most of the people because you need
> >> an uncommited gcc patch. So you can't build it.
> >
> > This looks really nice on amd64. I had to upper the limit of fds to make it
> > display websites with a lot of elements (256 works, 128 didn't) even when
> > using only one or two tabs.
> >
> > I'm really looking forward to see this as a port.

Reply | Threaded
Open this post in threaded view
|

Re: Chromium 9.0.597 packages for testing

Amit Kulkarni-5
I am compiling all kernel, userland and X on amd64 with latest
sources, and will let you know if there are any problems wit this
diff. Looks very small but I don't know if it will affect anything.

Please if it works right on all arch'es, let's get the port in!

Thanks

On Thu, Feb 3, 2011 at 1:27 PM, Robert Nagy <[hidden email]> wrote:

> The issue is that
> #praga pop and push is not supported in our in-tree gcc and it has
> to be enabled, otherwise chrome will have a lot of unaligned code
> compiled and you will have SIGBUSes all the time.
>
> Index: gcc/config/openbsd.h
> ===================================================================
> RCS file: /cvs/src/gnu/gcc/gcc/config/openbsd.h,v
> retrieving revision 1.5
> diff -u -r1.5 openbsd.h
> --- gcc/config/openbsd.h        11 Sep 2010 15:13:17 -0000      1.5
> +++ gcc/config/openbsd.h        3 Feb 2011 18:44:01 -0000
> @@ -325,6 +325,9 @@
>    as this depends on a few other details as well...  */
>  #define HANDLE_SYSV_PRAGMA 1
>
> +/* Define this so we can compile MS code */
> +#define HANDLE_PRAGMA_PACK_PUSH_POP 1
> +
>  /* Stack is explicitly denied execution rights on OpenBSD platforms.  */
>  #define ENABLE_EXECUTE_STACK                                           \
>  extern void __enable_execute_stack (void *);                           \
>
> --
>
> If you want you can compile chrome by yourself: https://github.com/rnagy/ports
>
> On (2011-02-03 19:37), Marc Espie wrote:
>> Hey, can I get the gcc patch ? I think I haven't got it on hackers@...
> On (2011-02-03 12:21), Amit Kulkarni wrote:
>> Er, I know its irrelevant for some (maybe all of you) but you need to
>> do a pkg_add of libxslt + libvpx from current packages before you can
>> check out chromium.
>>
>> I made kern.shminfo.shmall=65536 from the default 8192 (I might drop
>> it down to recommended 32768 if problems crop up) and the login.conf
>> limits are adequate in my case.
>>
>> Jumping straight to the most stable chrome available is a dream. I
>> just did a normal browse for half an hour or so. Checked javascript
>> websites etc. We need more alternative browsers for end users, gives
>> more bragging rights. I see robert@ made the 5.0.XXX version only for
>> i386, looks like it isn't possible for this updated port to go in for
>> 4.9 because of the possible gcc issues. Would somebody comment on this
>> one?
>>
>> Thanks
>>
>> >> The port update is useless for most of the people because you need
>> >> an uncommited gcc patch. So you can't build it.
>> >
>> > This looks really nice on amd64. I had to upper the limit of fds to make it
>> > display websites with a lot of elements (256 works, 128 didn't) even when
>> > using only one or two tabs.
>> >
>> > I'm really looking forward to see this as a port.
>

Reply | Threaded
Open this post in threaded view
|

Re: Chromium 9.0.597 packages for testing

Robert Nagy
No we can't it's too late for this release.

On (2011-02-03 13:44), Amit Kulkarni wrote:

> I am compiling all kernel, userland and X on amd64 with latest
> sources, and will let you know if there are any problems wit this
> diff. Looks very small but I don't know if it will affect anything.
>
> Please if it works right on all arch'es, let's get the port in!
>
> Thanks
>
> On Thu, Feb 3, 2011 at 1:27 PM, Robert Nagy <[hidden email]> wrote:
> > The issue is that
> > #praga pop and push is not supported in our in-tree gcc and it has
> > to be enabled, otherwise chrome will have a lot of unaligned code
> > compiled and you will have SIGBUSes all the time.
> >
> > Index: gcc/config/openbsd.h
> > ===================================================================
> > RCS file: /cvs/src/gnu/gcc/gcc/config/openbsd.h,v
> > retrieving revision 1.5
> > diff -u -r1.5 openbsd.h
> > --- gcc/config/openbsd.h        11 Sep 2010 15:13:17 -0000      1.5
> > +++ gcc/config/openbsd.h        3 Feb 2011 18:44:01 -0000
> > @@ -325,6 +325,9 @@
> >    as this depends on a few other details as well...  */
> >  #define HANDLE_SYSV_PRAGMA 1
> >
> > +/* Define this so we can compile MS code */
> > +#define HANDLE_PRAGMA_PACK_PUSH_POP 1
> > +
> >  /* Stack is explicitly denied execution rights on OpenBSD platforms.  */
> >  #define ENABLE_EXECUTE_STACK                                           \
> >  extern void __enable_execute_stack (void *);                           \
> >
> > --
> >
> > If you want you can compile chrome by yourself: https://github.com/rnagy/ports
> >
> > On (2011-02-03 19:37), Marc Espie wrote:
> >> Hey, can I get the gcc patch ? I think I haven't got it on hackers@...
> > On (2011-02-03 12:21), Amit Kulkarni wrote:
> >> Er, I know its irrelevant for some (maybe all of you) but you need to
> >> do a pkg_add of libxslt + libvpx from current packages before you can
> >> check out chromium.
> >>
> >> I made kern.shminfo.shmall=65536 from the default 8192 (I might drop
> >> it down to recommended 32768 if problems crop up) and the login.conf
> >> limits are adequate in my case.
> >>
> >> Jumping straight to the most stable chrome available is a dream. I
> >> just did a normal browse for half an hour or so. Checked javascript
> >> websites etc. We need more alternative browsers for end users, gives
> >> more bragging rights. I see robert@ made the 5.0.XXX version only for
> >> i386, looks like it isn't possible for this updated port to go in for
> >> 4.9 because of the possible gcc issues. Would somebody comment on this
> >> one?
> >>
> >> Thanks
> >>
> >> >> The port update is useless for most of the people because you need
> >> >> an uncommited gcc patch. So you can't build it.
> >> >
> >> > This looks really nice on amd64. I had to upper the limit of fds to make it
> >> > display websites with a lot of elements (256 works, 128 didn't) even when
> >> > using only one or two tabs.
> >> >
> >> > I'm really looking forward to see this as a port.
> >

Reply | Threaded
Open this post in threaded view
|

Re: Chromium 9.0.597 packages for testing

Amit Kulkarni-5
doesn't affect me personally but I can understand.

compiled everything fine.

started working and everything looks normal on amd64.

thanks

2011/2/3 Robert Nagy <[hidden email]>:

> No we can't it's too late for this release.
>
> On (2011-02-03 13:44), Amit Kulkarni wrote:
>> I am compiling all kernel, userland and X on amd64 with latest
>> sources, and will let you know if there are any problems wit this
>> diff. Looks very small but I don't know if it will affect anything.
>>
>> Please if it works right on all arch'es, let's get the port in!
>>
>> Thanks
>>
>> On Thu, Feb 3, 2011 at 1:27 PM, Robert Nagy <[hidden email]> wrote:
>> > The issue is that
>> > #praga pop and push is not supported in our in-tree gcc and it has
>> > to be enabled, otherwise chrome will have a lot of unaligned code
>> > compiled and you will have SIGBUSes all the time.
>> >
>> > Index: gcc/config/openbsd.h
>> > ===================================================================
>> > RCS file: /cvs/src/gnu/gcc/gcc/config/openbsd.h,v
>> > retrieving revision 1.5
>> > diff -u -r1.5 openbsd.h
>> > --- gcc/config/openbsd.h        11 Sep 2010 15:13:17 -0000      1.5
>> > +++ gcc/config/openbsd.h        3 Feb 2011 18:44:01 -0000
>> > @@ -325,6 +325,9 @@
>> >    as this depends on a few other details as well...  */
>> >  #define HANDLE_SYSV_PRAGMA 1
>> >
>> > +/* Define this so we can compile MS code */
>> > +#define HANDLE_PRAGMA_PACK_PUSH_POP 1
>> > +
>> >  /* Stack is explicitly denied execution rights on OpenBSD platforms.  */
>> >  #define ENABLE_EXECUTE_STACK                                           \
>> >  extern void __enable_execute_stack (void *);                           \
>> >
>> > --
>> >
>> > If you want you can compile chrome by yourself: https://github.com/rnagy/ports
>> >
>> > On (2011-02-03 19:37), Marc Espie wrote:
>> >> Hey, can I get the gcc patch ? I think I haven't got it on hackers@...
>> > On (2011-02-03 12:21), Amit Kulkarni wrote:
>> >> Er, I know its irrelevant for some (maybe all of you) but you need to
>> >> do a pkg_add of libxslt + libvpx from current packages before you can
>> >> check out chromium.
>> >>
>> >> I made kern.shminfo.shmall=65536 from the default 8192 (I might drop
>> >> it down to recommended 32768 if problems crop up) and the login.conf
>> >> limits are adequate in my case.
>> >>
>> >> Jumping straight to the most stable chrome available is a dream. I
>> >> just did a normal browse for half an hour or so. Checked javascript
>> >> websites etc. We need more alternative browsers for end users, gives
>> >> more bragging rights. I see robert@ made the 5.0.XXX version only for
>> >> i386, looks like it isn't possible for this updated port to go in for
>> >> 4.9 because of the possible gcc issues. Would somebody comment on this
>> >> one?
>> >>
>> >> Thanks
>> >>
>> >> >> The port update is useless for most of the people because you need
>> >> >> an uncommited gcc patch. So you can't build it.
>> >> >
>> >> > This looks really nice on amd64. I had to upper the limit of fds to make it
>> >> > display websites with a lot of elements (256 works, 128 didn't) even when
>> >> > using only one or two tabs.
>> >> >
>> >> > I'm really looking forward to see this as a port.
>> >
>

Reply | Threaded
Open this post in threaded view
|

Re: Chromium 9.0.597 packages for testing

Marc Espie-2
In reply to this post by Amit Kulkarni-5
On Thu, Feb 03, 2011 at 01:44:20PM -0600, Amit Kulkarni wrote:
> I am compiling all kernel, userland and X on amd64 with latest
> sources, and will let you know if there are any problems wit this
> diff. Looks very small but I don't know if it will affect anything.
>
> Please if it works right on all arch'es, let's get the port in!
>
> Thanks

The gcc patch definitely WON'T GO into src/ this late in the release process.

Reply | Threaded
Open this post in threaded view
|

Re: Chromium 9.0.597 packages for testing

stanleylieber
In reply to this post by Amit Kulkarni-5
> Chrome 9.0.597 packages for testing:
>
> http://blade2k.humppa.hu/chromium/

Installed on two i386 machines running a current snapshot (one hardware,
one vmware on xp). I've been surfing and installing plugins for several hours.
Only hiccup so far is the usual DNS resolution issues also exhibited by the
previous 5.x port.

-sl

Reply | Threaded
Open this post in threaded view
|

Re: Chromium 9.0.597 packages for testing

Stuart Henderson
On 2011/02/03 16:42, Stanley Lieber wrote:
> > Chrome 9.0.597 packages for testing:
> >
> > http://blade2k.humppa.hu/chromium/
>
> Installed on two i386 machines running a current snapshot (one hardware,
> one vmware on xp). I've been surfing and installing plugins for several hours.
> Only hiccup so far is the usual DNS resolution issues also exhibited by the
> previous 5.x port.

care to elaborate? I don't think that's been mentioned before.

Reply | Threaded
Open this post in threaded view
|

Re: Chromium 9.0.597 packages for testing

Vadim Zhukov
In reply to this post by Robert Nagy
On 3 February 2011 г. 12:45:25 Robert Nagy wrote:
> Chrome 9.0.597 packages for testing:
>
> http://blade2k.humppa.hu/chromium/
>
> The port update is useless for most of the people because you need
> an uncommited gcc patch. So you can't build it.

Work fine here on i386, except that no HTML5 video codecs support seem
to be there. Is this by purpose?

Thanks for your work (especially from my wife that liked even current
chromium)!

--
  Best wishes,
    Vadim Zhukov

A: Because it messes up the order in which people normally read text.
Q: Why is top-posting such a bad thing?
A: Top-posting.
Q: What is the most annoying thing in e-mail?

Reply | Threaded
Open this post in threaded view
|

Re: Chromium 9.0.597 packages for testing

matteo filippetto
In reply to this post by Stuart Henderson
2011/2/4 Stuart Henderson <[hidden email]>:

> On 2011/02/03 16:42, Stanley Lieber wrote:
>> > Chrome 9.0.597 packages for testing:
>> >
>> > http://blade2k.humppa.hu/chromium/
>>
>> Installed on two i386 machines running a current snapshot (one hardware,
>> one vmware on xp). I've been surfing and installing plugins for several hours.
>> Only hiccup so far is the usual DNS resolution issues also exhibited by the
>> previous 5.x port.
>
> care to elaborate? I don't think that's been mentioned before.
>
>

Maybe he refers to this

http://isc.sans.edu/diary.html?storyid=10312

Best regards


--
Matteo Filippetto
http://op83.blogspot.com

Reply | Threaded
Open this post in threaded view
|

Re: Chromium 9.0.597 packages for testing

stanleylieber
In reply to this post by Robert Nagy
I believe my previous attempt to send this message
was garbled.


> On 2011/02/03 16:42, Stanley Lieber wrote:
>> > Chrome 9.0.597 packages for testing:
>> >
>> > http://blade2k.humppa.hu/chromium/
>>
>> Installed on two i386 machines running a current snapshot (one hardware,
>> one vmware on xp). I've been surfing and installing plugins for several hours.
>> Only hiccup so far is the usual DNS resolution issues also exhibited by the
>> previous 5.x port.
>
> care to elaborate? I don't think that's been mentioned before.

Under the 5.x port, DNS resolution would occasionally stall for thirty seconds
or more with the message "Resolving host..." in the status bar. This wasn't
unique to the OpenBSD build; users on multiple operating systems reported
similar behavior.

Some *nix users reported that removing the "search" line from /etc/hosts
resolved the problem. Others reported disabling "Pre-fetching" in their
preferences eliminated the delays. My own tests produced inconclusive
results on several different machines (all OpenBSD/i386). After making a
change to /etc/hosts or preferences, the problem would disappear for a
while, only to return unexpectedly. On some systems I never experienced
the delays at all.

The new package exhibits similar behavior. The status bar message
now appears as "Searching..." It's a maddeningly inconsistent failure;
sometimes it happens and sometimes it doesn't.

On the plus side, 9.0.597 seems much more stable.

References:
http://old.nabble.com/Re%3A-chromium-port-update-p28726821.html
http://code.google.com/p/chromium/issues/detail?id=12754

-sl

Reply | Threaded
Open this post in threaded view
|

Re: Chromium 9.0.597 packages for testing

stanleylieber
In reply to this post by Robert Nagy
I've been hammering on the i386 package pretty hard over the last couple
of days. It's proven surprisingly stable on my single user test machines:

AMD Sempron(tm) Processor 3300+ ("AuthenticAMD" 686-class,
128KB L2 cache) 2.01 GHz, avail mem = 1572626432 (1499MB)

Intel(R) Pentium(R) M processor 1.60GHz ("GenuineIntel" 686-class)
1.60 GHz, avail mem = 2092380160 (1995MB)

/etc/login.conf:
staff:\
        :datasize=infinity:\
        :maxproc=infinity:\
        :openfiles-cur=1024:\
        :ignorenologin:\
        :requirehome@:\
        :tc=default:

/etc/sysctl.conf:
kern.shminfo.shmall=32768

I run an average of 10-15 tabs open pretty much continuously
(gmail, google reader, netvibes, tumblr, flickr, etc.). Subjectively,
performance is comparable to Firefox 3.x. Objectively, RAM usage
hovers at around 75%, CPU at 5-15% with the tabs left open and
the browser idle.

So far, no graphics glitches or crashes.

Problems:

- Japanese text displays as square boxes. This is a known issue
with Chrome that some users have rectified by uninstalling the
Droid fonts from their system. Japanese text displays correctly
in Firefox.

Thanks a lot for your hard work.

-sl

Reply | Threaded
Open this post in threaded view
|

Re: Chromium 9.0.597 packages for testing

Amit Kulkarni-5
Thanks everybody for getting in the latest version of chrome for 4.9.

Muchas gracias Robert!

On Sun, Feb 6, 2011 at 3:44 AM, Stanley Lieber <[hidden email]> wrote:

> I've been hammering on the i386 package pretty hard over the last couple
> of days. It's proven surprisingly stable on my single user test machines:
>
> AMD Sempron(tm) Processor 3300+ ("AuthenticAMD" 686-class,
> 128KB L2 cache) 2.01 GHz, avail mem = 1572626432 (1499MB)
>
> Intel(R) Pentium(R) M processor 1.60GHz ("GenuineIntel" 686-class)
> 1.60 GHz, avail mem = 2092380160 (1995MB)
>
> /etc/login.conf:
> staff:\
>        :datasize=infinity:\
>        :maxproc=infinity:\
>        :openfiles-cur=1024:\
>        :ignorenologin:\
>        :requirehome@:\
>        :tc=default:
>
> /etc/sysctl.conf:
> kern.shminfo.shmall=32768
>
> I run an average of 10-15 tabs open pretty much continuously
> (gmail, google reader, netvibes, tumblr, flickr, etc.). Subjectively,
> performance is comparable to Firefox 3.x. Objectively, RAM usage
> hovers at around 75%, CPU at 5-15% with the tabs left open and
> the browser idle.
>
> So far, no graphics glitches or crashes.
>
> Problems:
>
> - Japanese text displays as square boxes. This is a known issue
> with Chrome that some users have rectified by uninstalling the
> Droid fonts from their system. Japanese text displays correctly
> in Firefox.
>
> Thanks a lot for your hard work.
>
> -sl
>
>