6:39 p.m.
Seeing the information on Netsurf 2.5 I got the impression
that the memory theft problem had been sorted.
Leaving it on www.xe.com for about 60 minutes this morning
nevertheless reduced available memory from 47028k to 28k.
--
Russell Hafter - Mailing Lists
rh.lists(a)phone.coop
Need a hotel? <http://www.hrs.com/?client=en__blue&customerId=416873103>
(NB This link needs Firefox to work)
6:42 p.m.
On Mon, 26 Apr 2010 10:39:02 +0100
Russell Hafter - Lists <rh.lists(a)phone.coop> wrote:
Seeing the information on Netsurf 2.5 I got the impression
that the memory theft problem had been sorted.
No. The root cause of the problem is RISC OS's dreadful memory
management. NetSurf 2.5 simply has some measures to mitigate the issue.
B.
6:52 p.m.
On 26 Apr 2010 Rob Kendrick wrote:
On Mon, 26 Apr 2010 10:39:02 +0100
Russell Hafter - Lists <rh.lists(a)phone.coop> wrote:
> Seeing the information on Netsurf 2.5 I got the impression
> that the memory theft problem had been sorted.
No. The root cause of the problem is RISC OS's dreadful memory
management. NetSurf 2.5 simply has some measures to mitigate the issue.
Whether or not RISC OS's memory management is dreadful, most tasks
seem to cope with it.
--
Richard Porter http://www.minijem.plus.com/
mailto:ricp@minijem.plus.com
Confidentiality statement: Imagine 50 lines of pointless verbiage.
6:57 p.m.
On Mon, 26 Apr 2010 10:52:42 +0100
Richard Porter <ricp(a)minijem.plus.com> wrote:
On 26 Apr 2010 Rob Kendrick wrote:
> On Mon, 26 Apr 2010 10:39:02 +0100
> Russell Hafter - Lists <rh.lists(a)phone.coop> wrote:
>> Seeing the information on Netsurf 2.5 I got the impression
>> that the memory theft problem had been sorted.
> No. The root cause of the problem is RISC OS's dreadful memory
> management. NetSurf 2.5 simply has some measures to mitigate the
> issue.
Whether or not RISC OS's memory management is dreadful, most tasks
seem to cope with it.
Most tasks are not handling data structures anywhere near as complex as
NetSurf is. The only other that springs to mind is Firefox, and I think
we'll all agree that NetSurf is far more frugal.
B.
7:51 p.m.
On 26 Apr 2010 Rob Kendrick wrote:
On Mon, 26 Apr 2010 10:52:42 +0100
Richard Porter <ricp(a)minijem.plus.com> wrote:
> Whether or not RISC OS's memory management is dreadful, most
tasks
> seem to cope with it.
Most tasks are not handling data structures anywhere near as complex
as
NetSurf is. The only other that springs to mind is Firefox, and I think
we'll all agree that NetSurf is far more frugal.
Firefox isn't really viable on a RiscPC, but NetSurf didn't have this
problem until around August last year. The 21st July version (r8643)
is better behaved as regards memory utilisation.
--
Richard Porter http://www.minijem.plus.com/
mailto:ricp@minijem.plus.com
Confidentiality statement: Imagine 50 lines of pointless verbiage.
7:56 p.m.
On Mon, 26 Apr 2010 11:51:16 +0100
Richard Porter <ricp(a)minijem.plus.com> wrote:
> Most tasks are not handling data structures anywhere near as
> complex as NetSurf is. The only other that springs to mind is
> Firefox, and I think we'll all agree that NetSurf is far more
> frugal.
Firefox isn't really viable on a RiscPC, but NetSurf didn't have this
problem until around August last year. The 21st July version (r8643)
is better behaved as regards memory utilisation.
Well, you know where the source is. I've just written a longish mail
about what the problems are, and possible solutions to them, on the
developer's mailing list.
B.
7:36 p.m.
I downloaded version 2.5 (18th April 2010) this morning and it STILL
steals memory. It is a great shame that it has been left in this
state.
Michael Bell
--
7:44 p.m.
On Wed, 02 Jun 2010 11:36:01 +0100
Michael Bell <michael(a)beaverbell.co.uk> wrote:
I downloaded version 2.5 (18th April 2010) this morning and it STILL
steals memory. It is a great shame that it has been left in this
state.
Are you asking for a refund?
Alternatively, try the latest test build, which some people have said
is significantly better in this respect.
B.
10:25 p.m.
In message <20100602114415.0ea6b428(a)trite.i.flarn.net.i.flarn.net>
Rob Kendrick <rjek(a)netsurf-browser.org> wrote:
On Wed, 02 Jun 2010 11:36:01 +0100
Michael Bell <michael(a)beaverbell.co.uk> wrote:
> I downloaded version 2.5 (18th April 2010) this morning and it
STILL
> steals memory. It is a great shame that it has been left in this
> state.
Are you asking for a refund?
Alternatively, try the latest test build, which some people have
said
is significantly better in this respect.
B.
I went to where it said "the latest version". Where should I have gone
to?
Michael Bell
--
11:11 p.m.
On Wed, 02 Jun 2010 14:25:56 +0100
Michael Bell <michael(a)beaverbell.co.uk> wrote:
> Alternatively, try the latest test build, which some people
have
> said is significantly better in this respect.
I went to where it said "the latest version". Where should
I have
gone to?
http://www.netsurf-browser.org/downloads/riscos/testbuilds
B.
2:34 p.m.
On 02 Jun, Rob Kendrick <rjek(a)netsurf-browser.org> wrote:
On Wed, 02 Jun 2010 11:36:01 +0100
Michael Bell <michael(a)beaverbell.co.uk> wrote:
> I downloaded version 2.5 (18th April 2010) this morning and it
STILL
> steals memory. It is a great shame that it has been left in this
> state.
Are you asking for a refund?
This made my day :-)
Alternatively, try the latest test build, which some people have
said
is significantly better in this respect.
Yes, I find this to be the case (02 May build) and it is very stable too.
Thanks once again to all the developers
regards,
--
Robert Greenfield <rgr(a)spotswoodcollege.school.nz>
HOD Music
Spotswood College
7:58 p.m.
In article <ecc0052151.michaelbell(a)michael.beaverbell.co.uk>,
Michael Bell <michael(a)beaverbell.co.uk> wrote:
I downloaded version 2.5 (18th April 2010) this morning and it STILL
steals memory.
NetSurf 2.5 will always behave the same, whenever you download it. To see
the current state of things please try running the latest test build:
http://www.netsurf-browser.org/downloads/riscos/testbuilds
--
Michael Drake (tlsa) http://www.netsurf-browser.org/
6:40 p.m.
In message <512107da2ctlsa(a)netsurf-browser.org>
Michael Drake <tlsa(a)netsurf-browser.org> wrote:
In article <ecc0052151.michaelbell(a)michael.beaverbell.co.uk>,
Michael Bell <michael(a)beaverbell.co.uk> wrote:
> I downloaded version 2.5 (18th April 2010) this morning and it
STILL
> steals memory.
NetSurf 2.5 will always behave the same, whenever you download it.
To see
the current state of things please try running the latest test build:
I downloaded this and after a few day's trial it is still running
well. Thank you!
Michael Bell
--
7:10 p.m.
In article
<20100426104206.287b6fa2(a)trite.i.flarn.net.i.flarn.net>,
Rob Kendrick <rjek(a)netsurf-browser.org> wrote:
On Mon, 26 Apr 2010 10:39:02 +0100 Russell Hafter - Lists
<rh.lists(a)phone.coop> wrote:
> Seeing the information on Netsurf 2.5 I got the
> impression that the memory theft problem had been
> sorted.
No. The root cause of the problem is RISC OS's dreadful
memory management. NetSurf 2.5 simply has some measures
to mitigate the issue.
I do not claim to understand memory management, but the
problem did not exist with the earlier versions of NetSurf,
and, AIR, was only introduced around 9 months ago.
So something must have got broken somewhere.
Ver r8240 (1 July 2009) certainly does not have the problem
and I have been using that continously from when the problem
started. The version I still have from that time is r9012.
--
Russell Hafter - Mailing Lists
rh.lists(a)phone.coop
Need a hotel? <http://www.hrs.com/?client=en__blue&customerId=416873103>
(NB This link needs Firefox to work)
12:12 a.m.
On 26/04/2010 10:39, Russell Hafter - Lists wrote:
Seeing the information on Netsurf 2.5 I got the impression
that the memory theft problem had been sorted.
Leaving it on www.xe.com for about 60 minutes this morning
nevertheless reduced available memory from 47028k to 28k.
47MB isn't an unreasonable amount of memory for a browser to use.
It's less than 2.5% of the minimum any sensible system will have, and
not worth the developers of any modern application trying to optimise
away. Memory is cheap, developer's time isn't.
Either expand your Risc PC to the maximum amount of memory it will take
(256MB) and put off the problem for another few hours, or face the fact
that an almost 20 year old computer with a fraction of the processor
power and memory of a low end smartphone, just wont cut it any more.
Cheers
---David
--
David J. Ruck
email: druck(a)druck.org.uk
phone: +44(0)7974 108301
12:16 a.m.
On Mon, 26 Apr 2010 16:12:28 +0100
"David J. Ruck" <druck(a)druck.org.uk> wrote:
47MB isn't an unreasonable amount of memory for a browser to
use.
It's less than 2.5% of the minimum any sensible system will have, and
not worth the developers of any modern application trying to optimise
away. Memory is cheap, developer's time isn't.
Just to be clear, there are still a handful of memory leak bugs in
NetSurf, but 2.5 still greatly reduces its memory usage. (And we do
know what causes the one that Russell has experienced, but we're still
not completely clear on how to fix it yet.)
B.
3:46 a.m.
In article <4BD5AD5C.1070305(a)druck.org.uk>, David J. Ruck
<druck(a)druck.org.uk> wrote:
Either expand your Risc PC to the maximum amount of
memory it will take (256MB) and put off the problem for
another few hours, or face the fact that an almost 20
year old computer with a fraction of the processor power
and memory of a low end smartphone, just wont cut it any
more.
Or, more sensibly, just revert to a version of NetSurf that
does work OK when I want to use NetSurf with www.xe.com
permanently open.
I currently am running ten applications, including the older
version of NetSurf, and still have over half the installed
RAM free. Adding more RAM is a sledgehammer to crack a nut.
If only the XP computer a fraction of the RPC's age ran with
anything even approaching the speed of the RPC.
--
Russell Hafter - Mailing Lists
rh.lists(a)phone.coop
Need a hotel? <http://www.hrs.com/?client=en__blue&customerId=416873103>
(NB This link needs Firefox to work)
3:58 a.m.
New subject: NetSurf 2.5 still steals memory- but works much better elsewhere
After my problem with www.xe.com on NetSurf 2.5 I thought
that I would try and see if another page that has caused
problems in the past with older versions of NetSurf was now
any better...
Some time ago I reported (I think it may well have been on a
csa. newsgroup rather than here) about how NetSurf took
around 8 to 10 minutes to render a page from my German
bank's online banking. This is a list of stored payee
details - around 60 in total I would guess.
I just tried it with NetSurf 2.5, and while I did not
actually time it, it rendered at a usable speed - around 45
sec I would guess, meaning that it is useable on NetSurf
now. This is a major improvement.
--
Russell Hafter - Mailing Lists
rh.lists(a)phone.coop
Need a hotel? <http://www.hrs.com/?client=en__blue&customerId=416873103>
(NB This link needs Firefox to work)
8:34 p.m.
New subject: Was: NetSurf 2.5 still steals memory
On 27 Apr, David J. Ruck <druck(a)druck.org.uk> wrote:
47MB isn't an unreasonable amount of memory for a browser to
use.
It's less than 2.5% of the minimum any sensible system will have, and
not worth the developers of any modern application trying to optimise
away. Memory is cheap, developer's time isn't.
True, but haven't the Netsurf programmers done a great job?
Netsurf runs well on even a good ol' SARPC with a max of 256MB. A web
browser - a complex piece of software... and it runs *this* well on this
little old computer with its little OS, with simplistic memory handling
etc.
We have been spoiled by what can be run in small amounts of memory, I
guess. Web-browsers are hungry creatures.
Either expand your Risc PC to the maximum amount of memory it will
take
(256MB) and put off the problem for another few hours.
Quitting and restarting Netsurf is cheap and effective too.
or face the fact
that an almost 20 year old computer with a fraction of the processor
power and memory of a low end smartphone, just wont cut it any more.
Not for every task, no.
But what an almost 20 year old computer can do is OK. I like Artworks2 (
lot). I like Netsurf. I like PS3 and OvationPro. I like the apps you
made. I like what ROOL appear to be doing. The porting to beagleboard
etc. are interesting and I hope there can be some way forward here.
However I don't (can't) do multitrack recording on RISC OS. I'm not that
upset about it. That's just the way it is.
<whispers> we are allowed to have more than one computer.
regards,
--
Robert Greenfield
via RISC OS
<monster(a)caverock.net.nz> - http://www.webmonster.co.nz/
5:11 p.m.
In article <510df29b0frh.lists(a)phone.coop>,
Russell Hafter - Lists <rh.lists(a)phone.coop> wrote:
Leaving it on www.xe.com for about 60 minutes this morning
nevertheless reduced available memory from 47028k to 28k.
Please try the latest test build:
http://www.netsurf-browser.org/downloads/riscos/testbuilds
--
Michael Drake (tlsa) http://www.netsurf-browser.org/
5:55 p.m.
In article <510ff9e4f4tlsa(a)netsurf-browser.org>, Michael
Drake <tlsa(a)netsurf-browser.org> wrote:
In article <510df29b0frh.lists(a)phone.coop>, Russell
Hafter - Lists <rh.lists(a)phone.coop> wrote:
> Leaving it on www.xe.com for about 60 minutes this
> morning nevertheless reduced available memory from
> 47028k to 28k.
Please try the latest test build:and run it
I have downloaded r10525 and will report back
--
Russell Hafter - Mailing Lists
rh.lists(a)phone.coop
Need a hotel? <http://www.hrs.com/?client=en__blue&customerId=416873103>
(NB This link needs Firefox to work)
6:43 p.m.
In article <510ff9e4f4tlsa(a)netsurf-browser.org>, Michael
Drake <tlsa(a)netsurf-browser.org> wrote:
In article <510df29b0frh.lists(a)phone.coop>, Russell
Hafter - Lists <rh.lists(a)phone.coop> wrote:
> Leaving it on www.xe.com for about 60 minutes this
> morning nevertheless reduced available memory from
> 47028k to 28k.
Please try the latest test build:
Memory is still getting used up, though perhaps not as fast
as before.
When connected to www.xe.com the page reloads automatically
every minute.
Each time the page reloads the available RAM (according to
the FreeClock module on my icon bar) drops by just over
30kB.
From 1012 to 1042 (30 minutes) RAM has dropped from 45 628kB
to
44 632kB
The only thing that actually changes each time the page
reloads is the exchange rate numeric data.
--
Russell Hafter - Mailing Lists
rh.lists(a)phone.coop
Need a hotel? <http://www.hrs.com/?client=en__blue&customerId=416873103>
(NB This link needs Firefox to work)
4862
days inactive
4903
days old
netsurf-users@netsurf-browser.org
21 comments
8 participants
participants (8)
-
David J. Ruck -
Michael Bell -
Michael Drake -
Richard Porter -
Rob Kendrick -
Robert Greenfield -
Robert Greenfield -
Russell Hafter - Lists