Odyssey 1.23 progress: r5 RC2

@K-L

High likely because the WebKit core of the latest MorphOS version is a bit more newer compared to the core used in the latest version for AmigaOS4

@samo79

Ah yes, probably. I thought the latest version of Odyssey OS4 had the most recent webkit used by MorphOS' version.

@K-L

Compared to the 1.23 for AOS4 the 1.24/1.25 for MorphOS use a more recent core by about 1 year

@kas1e

Any news or betas on an Odyssey update?

@khayoz

Yes an updated webkit would be really cool. Nowadays, more and more websites became unreadable with our current version of Odyssey. This makes me sad to not even be able to read my emails through the webmail of my isp anymore.

@all
As were said many times updated core of webkit in odyssey is not possible: they changed it too much. Even on morphos they go from scratch with Wayfarer.

All what can be updated in Odyssey is 3d party libs, gui code, some fixes in webkit core where possible, but not whole core: its still will be same 10 years old core, sadly (for that new browser should be done based on new webkit core).

So, do we need fixes not related to core updating, like updated 3party libs, recompiling over clib4 and new binutils, etc ?

Just stumbled upon this project:

https://github.com/arthurchoung/RELISH

Besides the OS4 browser project, could that be another approach...?

@Primax

Quote:

We don't have Objective-C, and it does not look like full web browser. If it is, then lots other project also needs to be ported.

Okay, thanks.

But maybe generally, this is also the right way for OS4: not a full, but fast and slim web browser which does not have to support "everything"...?

For example, I would do without YouTube (and any other video), but would like to have a browser which supports HTML5, CSS und JavaScript.

@Primax

Yeh… think maybe best thing is just focus on JavaScript, JIT compiler and JavaScript Binary format. It seams to me good way go about be looking at NodeJS, also we other browsers like NetSurf I think can be good.

I think part of problem, is that we want latest stuff now, and as result we keep taking shortcuts. Not saying retrofitting new components, parallel development is good idea, as it ends in incompatible side projects.
On the other hand, one problem with EUAE, is the maintainer kept porting over newer code, but did not work on integration with OS, as result EUAE sucks on many parts like network support, protection bits etc, let’s say user experience can be better, then what it is now. The same you have any large project; we need to customize it our needs. It can’t just be recompiled. Well technically first modern web browsers we had, was SDL based, without any integration at all.

Search with 'Duckduck' does not seem to work anymore with Odyssey. If I input any word in the search field, there appears only a blank page with this text on it:

"If this persists, please email us. Our support email address includes an anonymized error code that helps us understand the context of your search."

Is there any fix for this? The link defined on 'search engines' window is currently "https://duckduckgo.com/html/?q=%s".

@Gregor
With time more and more things will not work with Odyssey. If you need any particular site fixes, try to invistegate why exactly it didnt work, and maybe you will be able to fix it by scripts (as Walkero do for example for github, to some extent).

@all
As i say, core update near to imposible with their massive changes, and only new browser will help, but any of you can try to do it anyway, Odyssey code open, with all how-to-build, so feel free. But reporting sites issues make no sense, as core will be still old. Even if i will update it to some better version, like one in 1.25 version for aros, it still will be too old to render modern sites today. So realistically speaking : new core mean new browser.

From my side i only can update 3d party libs, rebuild with clib4/new binutils, add/fix some bits there and there, but nothing hardcore like core update so to have desktop-youtube, telegram and whatever else. Do you need that ?

@kas1e

Quote:

Certainly yes, generally even just an update to the libraries brings benefits in term of speed and stability

One question, why does upgrading to a new core mean rewriting the entire browser? Couldn't we reuse the current Odyssey's MUI interface by making to it only the necessary changes to use the new core?

@samo79

I’m think the same unless complete API’s have changed.
can be gradual thing, where you can’t just take A and apply it on B, but have rewrite things step by step.

@samo79
Quote:
Since our version webkit just changed very much: layout, api, connection with gui, everything. Adapting current gui to different core (which later webkit are) mean the same as writing new gui, but writing new gui always better and easy in compare with rewriting pieces of other's code.

And basically GUI not the problem for adequate gui ptogrammer who know how to wrote UI. Problem is core.

Everything possible of course, code of Odyssey open, with as much as possible desribed steps of how to build it - maybe someone have skills to update to later webkit, but so far no one.

Talks is cheap, and currently only George (Walkero) with help from others tryied to port at least core of webkit from Morphos's Wayfarer. And new GUI still need to be written. Of course an Odyssey's one can be taken as reference, but as i say GUI is not problem, just GUI should be written by one who know UI and didnt introduce "amiga only" logic :)

@Gregor

Try


I get the blank reults page from time to time, too.
Nothing much you can do, close Odyssey, restart and hope for the best

@kas1e

Yeah, aniway must be said that even rewriting a new GUI that has the same numerous features as Odyssey must not be simple... rewriting a GUI from scratch every time means always having a more modern browser but with fewer features ...
In this sense, re-adapting an already existing GUI certainly has its real advantages... after all, when browsers are updated on other platforms I think they do exactly this

@samo79
No needs to argue, try yourself or find someone who prove me wrong by commit to Odysseys repo updated core working with exiting gui :) i will then accept i was wrong and will pay such person 1000$ right away as "thanks".

@kas1e

I wasn't arguing, I just expressed a consideration :)

@RazielQuote:

Works nicely! Thank you, Raziel .