@Raffaele

Quote:

Raffaele wrote: @OneTimer1 All I know about Javascript endianess issue comes from various threads that explain Fabien Coeurjoly can't release MorphOS OWB 1.25 because he is facing a very high number of problems due to last version of Webkit Javascript. Also there is Deadwood thread on Aros-Exec.org where he launches the idea of the bounty for resolving Javascript issues. http://aros-exec.org/modules/newbb/viewtopic.php?post_id=97679#forumpost97679

endian problems may manifest themselves differently, wrong colors on rendered pages, not reading or saving data in usable way, application not starting or crashing, also, being slow or getting stuck i guess. certainly it would be better to educate oneself of the kind of problems and their scale before requesting people to take care of it. in the past while testing, when i was simply reporting this or that problem in many cases i have not realized that there is yet a long way for a developer to find a couse of it.

Last edited by wawa on 29-Feb-2016 at 10:29 AM.

@Xenic

Being able to switch endian on the fly could be useful. To isolate blocks of code. Of course they would need to work in a closed system where the work is done in one section before it is used externally.

Those instructions you list are exactly what I was thinking an endian aware compiler would use.

I wonder what compiler would use them? One that is smart enough to detect a word reversal?

@Raffaele

What if there was less emphasis on PPC? And the point made to be endian agnostic?

Apart from that are they interested in writing clean portable code? Perhaps not.

@Belxjander
@wawa

Absolutely, I should have mentioned that Chris' has been kindly helping me enormously and that without his help, I would be nowhere now

I would argue that some preconfigured all-in-one virtual machine would help developers get started. I remember a story (but cannot find a link to it at the moment) that some developers chose Chrome over Firefox because Chrome was so easy to build.

Cheers!

_________________
Tygre
Scientific Progress Goes Boing!

@BSzili

I think the best is just ask them on Twitter: https://twitter.com/webkit. I think it should do our developer(s), to discuss details with WebKit team, what they can do or what should be done on our side.

@radzik
Yup, that's what I was alluding to. Still, people will keep stating as a fact that WebKit maintainers won't accept patches from us without bothering to actually ask them about it :)

_________________
This is just like television, only you can see much further.

@BSzili

so someone ask them already, instead talking about it here. and it better be someone who understands the problem.

@Hypex


Quote:

Those instructions you list are exactly what I was thinking an endian aware compiler would use

I don't think a compiler can be endian aware. Binary data is just binary numbers in memory or registers. I can't imagine any way to determine the format of a binary number. Since most of the work in a program is performed in the C library, I would think that you would need a C library that is compiled to use little-endian too. However, it's all just speculation unless someone actually attempts to do it.

_________________
X1000 with 2GB memory & OS4.1FE

@Hypex

Quote:

Hypex wrote: @Raffaele What if there was less emphasis on PPC? And the point made to be endian agnostic? Apart from that are they interested in writing clean portable code? Perhaps not.

Anyone who wants he is free to made a fork project dealing with endian agnostic code...

If just debugging will take almost six months of work for a single person (according to Deadwood esteem)... Then write OWB Webkit engine with agnostic endian code will require IMHO almost two years of coding, not counting debugging.

It is just a matter of time and effort.

Personally I think that if we succeded in collectiong almost 6000 dollars and give 2000 to any member of a 3 people team, in one month or two the team will release a working 1.25 OWB update, free of any Javascript endianess issues.

Last edited by Raffaele on 01-Mar-2016 at 01:02 PM.
Last edited by Raffaele on 01-Mar-2016 at 01:01 PM.
Last edited by Raffaele on 01-Mar-2016 at 01:00 PM.

_________________
"When the Amiga came out, everyone [at Apple] was scared as hell." (J.L. Gassée, former CEO of Apple France and chief of devs of Mac II-fx, interviewed by Amazing Computing, Nov 1996).

@Raffaele

How many hours a day for a month or two? Are we talking full time or an hour at night? you need to take all of that into consideration when you make estimates.

_________________
Support AROS sponsor a developer.

"AROS is prolly illegal ~ Evert Carton" intentionally quoted out of context for dramatic effect

@Raffaele

your guess would mean 2 months fulltime... not much money expecially if you have to assume that most devs here are either self-employed having their paid projects or working for someone else (and thus cannot make vacation for 2 months for that project)

also you would need to clone Deadwood before... because he has both the skills and deep sourcecode insight knowledge

for everybody else it would need certainly more time

How about a Patreon to keep pay whoever is going to maintain the project? Maybe even the Webkit Teamwould say yes then. I can start by promising 5€/month from me.

_________________
Amiga Forum - Sweden's best Amiga Magazine

My MorphOS Blog

"Free speech includes other peoples right to offend you."

@terminills

Dunno... ask OlafS who is in direct contact with Deadwood... Deadwood could solve your dubts.

_________________
"When the Amiga came out, everyone [at Apple] was scared as hell." (J.L. Gassée, former CEO of Apple France and chief of devs of Mac II-fx, interviewed by Amazing Computing, Nov 1996).

@Thread


ATTENTION ALL!

I received mail from a guy called Internetzel who committed its internet safari patch 32 bit on bounty:

Quote:

Here's my current patch from Leopard WebKit for JavaScriptCore. It's been done on the branch "safari-601-branch" which was originally branched off at trunk revision 186671. Built with the attached patch (with full optimization on gcc 5.3) on a 32 bit big endian PowerPC machine the JavaScriptCore C-Loop Interpreter passes all tests that I suppose it should pass. Most chunks in the patch shouldn't be needed for MorphOS but someone who knows how to deal with it, like Fabien Coeurjoly, will easily be able to strip the unneeded parts out.

Could someone with programming skills please extract debugged Javascript code and embed it into OWB?

If patch works so then I could declare this guy eligible to get the bounty!

_________________
"When the Amiga came out, everyone [at Apple] was scared as hell." (J.L. Gassée, former CEO of Apple France and chief of devs of Mac II-fx, interviewed by Amazing Computing, Nov 1996).

@Raffaele

Is it message from Tobias Netzel? I'm in contact with him as well. But I think our developers should be contacted with him? I'm thinking about : deadwood, kas1e and Fab (if he is still in Amiga world).

@Xenic

Quote:

I don't think a compiler can be endian aware. Binary data is just binary numbers in memory or registers.

Yes that is true but what I was proposing is an attribute than can be applied to specify a data word as little endian. So natively on PPC the compiler would generate standard load/store word instructions. This attribute would mark a datatype of an entire structure or variable as being reversed so that reversed load/store word instructions need to be used.

Quote:

Since most of the work in a program is performed in the C library, I would think that you would need a C library that is compiled to use little-endian too.

That may be too high level unless the C library also reads/writes data words. Usually it would be coded as part of the program instruction stream.

@Raffaele

Quote:

Anyone who wants he is free to made a fork project dealing with endian agnostic code...

Yes that could be done but that wouldn't fix the other code. It seems to me that if the high level code relies on a certain memory model then it must be broken. Portable code shouldn't do that. Okay they don't need it to be portable but Intel or based isn't the only CPU anymore so they should still keep it clean as code on different compilers can behave differently.

Quote:

If just debugging will take almost six months of work for a single person (according to Deadwood esteem)... Then write OWB Webkit engine with agnostic endian code will require IMHO almost two years of coding, not counting debugging.

That's the problem. All this effort, by one person or more, with financial incentive, will be put into porting the code to be endian agnostic. And then that code version will go out of date the problem reappears. A more future proof solution is needed I think. Whatever that may be.

If we look at Timberwolf as an example, a host of effort and work was put into that to port it to OS4. And then by then end what was left is an incomplete port of a then obsolete code base. That was too old to work with and not suitable for forward-porting.

@radzik

Quote:

radzik wrote: @Raffaele Is it message from Tobias Netzel? I'm in contact with him as well. But I think our developers should be contacted with him? I'm thinking about : deadwood, kas1e and Fab (if he is still in Amiga world).

I don't know.

If you are in contact with Tobias please ask him...

[EDIT]

I made some search and I found on Sourceforge that the guy called Internetzel is right Tobias Netzel and he committed some days ago a version of Safari Webkit for PPC MacOS Snow Leopard.

https://sourceforge.net/u/internetzel/profile/

Last edited by Raffaele on 02-Mar-2016 at 10:00 PM.

_________________
"When the Amiga came out, everyone [at Apple] was scared as hell." (J.L. Gassée, former CEO of Apple France and chief of devs of Mac II-fx, interviewed by Amazing Computing, Nov 1996).

@Raffaele

Tobias isnt upgrading leopard webkit from months probably he is try to find a way for have the jit there... but who know ?
Another one good programmer can be Cameron Kraiser (teen4fox) but it is Geko and not webkit)

_________________
I love Amiga and new hope by AmigaNG
A 500 + ; CDTV; CD32;
PowerMac G5 Quad 8GB,SSD,SSHD,7800gtx,Radeon R5 230 2GB;
MacBook Pro Retina I7 2.3ghz;
#nomorea-eoninmyhome

@tlosm

You posted while I was editing...

Last edited by Raffaele on 02-Mar-2016 at 10:02 PM.

_________________
"When the Amiga came out, everyone [at Apple] was scared as hell." (J.L. Gassée, former CEO of Apple France and chief of devs of Mac II-fx, interviewed by Amazing Computing, Nov 1996).