@nexus

When i feel like it, maybe i'll add rearrangable groups for navigation, url & co. Of course it would be trivial to have the whole thing in one line, but i want to do it properly, and not just as a quick hack.

And personally, i hate having too short url strings. :)

@mrodfr

Due to Cairo relative slowness and WebKit complexity, it can't be as fast as Ibrowse, obviously.

Usually rendering a whole page with WebKit + Cairo takes about 50 to 150ms depending on page layout. So you can do the math if you scroll with the wheel for a 3 lines step : scrolling already displayed content is very fast as it's done by the blitter, and then drawing new data through Cairo+WebKit would take something like 30ms or so.

(it can even be timed precisely through an ENV var, as i've added some internal benchmark thing).

Last edited by Fab on 19-Apr-2011 at 11:24 AM.
Last edited by Fab on 19-Apr-2011 at 11:24 AM.

@kas1e

Cool!

@nexus

I think it ise but I haven't tried it. It may be done in a different way that isn't MUI3.x compatible and crash. Don't know.

Edit: Apparently it doesn't, yet

@Mrodfr

I'd say on a normal CSS page it's between IBrowse and RA-OWB. But since MUI-OWB handle transparent CSS objects it becomes extremely slow on some pages, unlike RA-OWB which apparently ignores transparency. Or something like that...

It's over all faster than RA-OWB but not for all sites. But I mean they are both based on the same webkit engine so fairly similar. IBrowse can prepare a complete page as raw image data and just scroll it using accelerated 2D graphics basically, but that is because it ignores all dynamic elements that CSS and iframes can provide. A CSS browser can never be that fast unless you slam lots and lots of raw CPU power at it. IMHO.

Last edited by Deniil715 on 19-Apr-2011 at 11:24 AM.

_________________
- Don't get fooled by my avatar, I'm not like that (anymore, mostly... maybe only sometimes)
> Amiga Classic and OS4 developer for OnyxSoft.

@Deniil715

Well, i don't know what pages you tested with RA OWB, but by design, RA OWB can only be slower.

Then, maybe RA OWB uses a faster (possibly older) Cairo version that makes some thing render faster, but that's it.

Just for the fun, CSS can also be disabled through the content blocker :
you add a rule "http://*#stylesheet", and you get the good old Ibrowse feeling then (still not as fast, though). :)

Last edited by Fab on 19-Apr-2011 at 11:36 AM.
Last edited by Fab on 19-Apr-2011 at 11:32 AM.

@nexus

Quote:

Is it possible to freely re-arrange the "controling elements" of MUI-OWB, like "navigation-buttons, url bar, quicklinks, etc"? I just wonder because there's so much space "wasted" next to the navigation buttons and it would be nice, to locate the complete line with the url-bar, google-search, etc. there.

Maybe Fab will add it at some stage :)

Quote:

As I said, it's just curiosity -- I don't want to complain

Complaining its a way to progress. I.e. feel free to say everything what you think (its in end just port, and all the wish-lists-for-todo going to Fab :) ).

For me personally its ok as it now, but sure, if it will possible to reorganize all the stuff as you wish it will be even better.

@mrodfr

Quote:

- How fast is this version (when scrolling an internet page with the mouse whell) ? (like Ibrowse -> instantly, or like Netsurf -> move more slower).

I think you mean how big step of mouse-wheel scrolling, i.e. did you need to scroll mouse wheel a lot to scroll page a bit, or just smal move to scroll the whole , then the answer is : speed of mouse-wheel is the same as on my FF4 on win32 . The same portions of data are scrolls by one mouse-wheel move.

But if you mean how fast or slow it scrolls at all, then Fab and others already answer on.

In end of all , noone can expect that browser which have support for all the modern stuff with faster than the browser which are 10 years old , and not support 50% of all standards (what mean loads faster, displays faster, renders faster and so on). The "speed" of ibrowse in his limitation. When all the stuff will be added to it, it can be even slower that current ra-own/mui-owb easy.

@dax

Quote:

you'll start being pestered with "give us a beta" demands very soon

We still need to do some very latest fixes/changes, and then we will think about the way to release it fast (we have some ideas already).

@all

There is the latest video of mui-owb in action. Yesterday Deniil fix some bug with Tabs, and now just need to add mouse pointers , fix curl , fix one last small bug and that all. After that we will prepare ready-to-use-archive and will think how to release all of this faster (and in meanwhile maybe adding NPAPI support as well).


@fab & deniil

I think we just need found about 5-10 sites which slower than in RA-OWB, and check why. One of them are html5test.com.

ps. checked binary of ra-owb, and found 1.8.10 right under cairo words (the same version which we use for port as well).

_________________
Join us to improve dopus5!
zerohero's mirror of os4/os3 crosscompiler suites

@nexus

Quote:

Is it possible to freely re-arrange the "controling elements" of MUI-OWB, like "navigation-buttons, url bar, quicklinks, etc"?

I thought always MUI gives that as granted and automaticly without programmers efforts !?

_________________
Rock lobster bit me. My Workbench has always preferences. X1000 + AmigaOS4.1 FE
"Anyone can build a fast CPU. The trick is to build a fast system." -Seymour Cray

@Tomppeli

Of course not. It's up to the programmer to decide which elements should be draggable and where they can be dropped. What's granted and without effort with MUI is rather the Drag'n'Drop operation itself.

A tookit that would allow re-arranging any GUI without any constraint by default would lead to a total mess. :)

@Fab

Quote:

When i feel like it, maybe i'll add rearrangable groups for navigation, url & co. Of course it would be trivial to have the whole thing in one line, but i want to do it properly, and not just as a quick hack.

Yep i follow this feature request, maybe some configurable settings for 1.13

@Kas1e

Excellent !

_________________
BACK FOR THE FUTURE

http://www.betatesting.it/backforthefuture

Sam440ep Flex 800 Mhz 1 GB Ram + AmigaOS 4.1 Update 6
AmigaOne XE G3 800 Mhz - 640 MB Ram - Radeon 9200 SE + AmigaOS 4.1 Update 6

@kas1e
The video was nice

_________________
SamFlex Complete 800Mhz System + AmigaOS 4.1 Update 4
Amiga 2000 DKB 2MB ChipRam GVP G-Force040 Picasso 2 OS3.9 BB2
AmigaCD 32

@kas1e

Quote:

There is the latest video

Impressive!

i hope Fab is going to resolve DNS problem of OWB, that problem is very boring, congratulations to all involved !!

_________________
Sam 460EX, 2Gb Ram, Radeon R7 250, AmigaOS4.1 FE
A4000 PPC604@233, Mediator
A1200 PPC603@160, Mediator
uA1 G3@800, 512 Mb [sold]

@SOFISTISOFTWARE

It's solved since a long time. Both in OWB and in CURL, in fact.

@Fab

What ??? Are you saying the coming OS4 version doesn't have the DNS annoying problem ??? To understand, doesn't it stop if there is anything not load ??

Last edited by SOFISTISOFTWARE on 19-Apr-2011 at 03:31 PM.

_________________
Sam 460EX, 2Gb Ram, Radeon R7 250, AmigaOS4.1 FE
A4000 PPC604@233, Mediator
A1200 PPC603@160, Mediator
uA1 G3@800, 512 Mb [sold]

@SOFISTISOFTWARE

Yes i presume is a Reaction OWB issue (only)

:-/

_________________
BACK FOR THE FUTURE

http://www.betatesting.it/backforthefuture

Sam440ep Flex 800 Mhz 1 GB Ram + AmigaOS 4.1 Update 6
AmigaOne XE G3 800 Mhz - 640 MB Ram - Radeon 9200 SE + AmigaOS 4.1 Update 6

@SOFISTISOFTWARE

In OWB, 2 years ago, i had moved whole network code to a dedicated thread to prevent it from blocking whole UI, but it didn't change the fact DNS could still block the network code itself. Which is why i had later implemented asynchronous DNS resolution in my CURL port.

But from what i can tell, current version of MUI OWB for OS4 has network threading disabled, and DNS resolution is also synchronous in CURL itself.

Last edited by Fab on 19-Apr-2011 at 03:42 PM.

@Fab

But from what i can tell, current version of MUI OWB for OS4 has network threading disabled, and DNS resolution is also synchronous in CURL itself.

what a pity ! I'm sorry for that :(

Last edited by SOFISTISOFTWARE on 19-Apr-2011 at 04:18 PM.

_________________
Sam 460EX, 2Gb Ram, Radeon R7 250, AmigaOS4.1 FE
A4000 PPC604@233, Mediator
A1200 PPC603@160, Mediator
uA1 G3@800, 512 Mb [sold]

@SOFISTISOFTWARE

Right now Deniil rewrite that part which are mos-specific in terms of network-thread , so, it will be enabled as well (and can be disabled from options if user wish).

About adding of asyncio dns resolver to Curl : fab loose his source codes of curl, so we need to reimplement them as well, to make it better.

But even with just enabled "network-thread", it will be better than nothing for sure.

_________________
Join us to improve dopus5!
zerohero's mirror of os4/os3 crosscompiler suites

@kas1e

ohhh nice news !! I hope Fab is finding the source code for Curl, however thanks a lot for the big work

_________________
Sam 460EX, 2Gb Ram, Radeon R7 250, AmigaOS4.1 FE
A4000 PPC604@233, Mediator
A1200 PPC603@160, Mediator
uA1 G3@800, 512 Mb [sold]

@SOFISTISOFTWARE

We can try to add async stuff to curl (because anyway we need rewrite file.c for make URI works fine), so if Fab will help us, and he still not drive into the madness because of those 4 months of annoing, then i think we can do that as well :)

Btw, with network-thread, dns-blocking will be not so problematic. I mean you for example go to some site, then dns-resolver suck and block the stuff, so, with enabled network-thread you just open new tab, and continue your work.

But for sure better to add async to curl as well.

Last edited by kas1e on 19-Apr-2011 at 04:54 PM.

_________________
Join us to improve dopus5!
zerohero's mirror of os4/os3 crosscompiler suites

@kas1e

yes, sure is it better to add async to curl and i think Fab will help you !

_________________
Sam 460EX, 2Gb Ram, Radeon R7 250, AmigaOS4.1 FE
A4000 PPC604@233, Mediator
A1200 PPC603@160, Mediator
uA1 G3@800, 512 Mb [sold]

@kas1e

mui-owb looks fantastic! Can't wait to get my hands on it when it's ready.

_________________
Hate tends to make you look stupid...