Re: [Freedos-devel] Volunteering - Arachne

2009-03-30 Thread Bernhard Eriksson
Jim wrote:
 Also: IMHO, Arachne should not be a file manager or email client. I
 think the most progress could be made by simplifying the code, and
 putting the focus on the web browser.
   
The browser is in core.exe, filemanager in wwwman.exe and insight.exe 
(formerly mailman.exe) handles email.
I agree on simplifying the code, Ray has done a great job on this and 
has reduced the code but still keep most of the functionality (XT and 
CGA support are two things his port has dropped that very few people 
use). Unfortunately due to time contraints from almost all of the 
developers little else is happening.

-- 
Bernhard Eriksson, Wermlandsdata
Fryxellsgatan 2, 652 22 Karlstad
054 - 15 69 00, http://www.wermlandsdata.se/
Datorer, tillbehör, service, programmering mm. 



--
___
Freedos-devel mailing list
Freedos-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/freedos-devel


Re: [Freedos-devel] Volunteering - Arachne

2009-03-29 Thread Michael Reichenbach
By the way I also think Arachne is one of the DOS flagship projects.

Unfortunately there are some very basic things which makes this software
less useful and afaik there is currently no one continuing the
development of this project.

1) no SSL support
- latest DOSLynx implemented it

2) Linux port
- From Arachne v1.93 there is also a Linux port.
- Unfortunately it's also not completely synchronized with v1.95.
- I wonder why there are two different source packages, the DOS and
Linux source should be unified as any multi platform application which
can be compiled in a few steps for different platforms does this so. It
also ensures also that all ports are always up to date.

3) uses still real mode and xswap
- Udo Kuhnt made an alpha version with DPMI out of v1.90J1 while latest
version in the maintree is v1.95 but I think backporting the few
changes shouldn't be a big deal
- I am not sure whenever Udo Kuhnt's version uses 16 or 32 bit DPMI but
it I think it's 16.
- 32 bit DPMI would be better.
- As I think xswap will not work in native Linux (no emulation) as there
is no xms, ems so it must have been ported already. Why the DOS version
uses still xswap then?

4) compiler
- still dependent to Borland C, port to OW and/or DJGPP would be good
because more people are using it
- Makes it point to port it from C to C++?

5) to many ports
- There are to many ports with different features floating around, all
should be merged, there was even a windows port form 1.6x or so.

6) graphics backend
- Dr WebSpyder and Lineo Embrowser (unfortunately no source code) ported
to Allegro and gave Arachne a speed boost.

--
___
Freedos-devel mailing list
Freedos-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/freedos-devel


Re: [Freedos-devel] Volunteering - Arachne

2009-03-29 Thread Eric Auer

Hi Michael,

 By the way I also think Arachne is one of the DOS flagship projects.

I agree.

 1) no SSL support
 - latest DOSLynx implemented it

I believe there are also eLinks based projects with SSL for DOS?

 2) Linux port
 - From Arachne v1.93 there is also a Linux port.
 - Unfortunately it's also not completely synchronized with v1.95.
 - I wonder why there are two different source packages, the DOS and
 Linux source should be unified as any multi platform application..

Arachne is not as cool as the big browsers among those who CAN
run the big browsers - But it could be, in particular because
it has the potential to run with little RAM... At the moment,
alas, it still needs big RAM or big temp directory for the
rendering of JPEG and other images as far as I remember...?

Another good thing to have would be basic Javascript, and I think
there were projects in that direction...

 3) uses still real mode and xswap
 - Udo Kuhnt made an alpha version with DPMI out of v1.90J1 while latest
 version in the maintree is v1.95 but I think backporting the few
 changes shouldn't be a big deal

Not sure...

 - As I think xswap will not work in native Linux

You will not need any of that if you simply compile Arachne
with a 32bit compiler in the first place. I do not think any
286 or older PC could run Arachne at acceptable speed anyway.

 4) compiler
 - still dependent to Borland C, port to OW and/or DJGPP would be good

I agree.

 - Makes it point to port it from C to C++?

No. Only if there is much of hard-to-read object-related code
and only if you have the idea that this code would be easier
to implement in object oriented languages.

 6) graphics backend
 - Dr WebSpyder and Lineo Embrowser (unfortunately no source code)
 ported to Allegro and gave Arachne a speed boost.

Arachne has GPL license so you can force WebSpyder/Lineo to
make their Arachne modifications public.

I agree that Arachne (and MPXPLAY) are nice DOS projects which
can become even nicer if there are more volunteers to help them.

Eric



--
___
Freedos-devel mailing list
Freedos-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/freedos-devel


Re: [Freedos-devel] Volunteering - Arachne

2009-03-29 Thread lyricalnanoha


On Sun, 29 Mar 2009, Eric Auer wrote:


 Hi Michael,


snips here and there

 1) no SSL support
 - latest DOSLynx implemented it

 I believe there are also eLinks based projects with SSL for DOS?

Dunno if the elinks port does SSL, I haven't tried that aspect.

 4) compiler
 - still dependent to Borland C, port to OW and/or DJGPP would be good

 I agree.

DJGPP's libs are very Borland-like to begin with.  Nice feature.

 - Makes it point to port it from C to C++?

 No. Only if there is much of hard-to-read object-related code
 and only if you have the idea that this code would be easier
 to implement in object oriented languages.

Agreed: there's almost never a need to translate anything from C to C++.

-uso.

--
___
Freedos-devel mailing list
Freedos-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/freedos-devel


Re: [Freedos-devel] Volunteering - Arachne

2009-03-29 Thread Michael Reichenbach
Eric Auer schrieb:
 - As I think xswap will not work in native Linux
 
 You will not need any of that if you simply compile Arachne
 with a 32bit compiler in the first place.

Why it can't be also simple compiled on DOS with 32 bit compiler?

 6) graphics backend
 - Dr WebSpyder and Lineo Embrowser (unfortunately no source code)
 ported to Allegro and gave Arachne a speed boost.
 
 Arachne has GPL license so you can force WebSpyder/Lineo to
 make their Arachne modifications public.

No, them have purchased a proprietary license from the copyright holders
(original developers). GPL doesn't surrender rights. It's dual licensing
like with Qt.

-mr

--
___
Freedos-devel mailing list
Freedos-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/freedos-devel


Re: [Freedos-devel] Volunteering - Arachne

2009-03-29 Thread Eric Auer

Hi,

 - As I think xswap will not work in native Linux
 You will not need any of that if you simply compile Arachne
 with a 32bit compiler in the first place.
 
 Why it can't be also simple compiled on DOS with 32 bit compiler?

That is what I meant - port Arachne to GNU C / DJGPP and
then you have better performance in DOS - and an easier
life in porting to Linux. Two good things for one effort.

 - Dr WebSpyder and Lineo Embrowser (unfortunately no source code)
 ported to Allegro and gave Arachne a speed boost.
 Arachne has GPL license so you can force WebSpyder/Lineo to
 make their Arachne modifications public.
 
 No, them have purchased a proprietary license from the copyright holders
 (original developers). GPL doesn't surrender rights. It's dual licensing
 like with Qt.

Oh I understand. That is bad luck for fans then, fans
will have to port Arachne to Allegro a second time if
they want a free open source Allegro based Arachne...

Eric



--
___
Freedos-devel mailing list
Freedos-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/freedos-devel


Re: [Freedos-devel] Volunteering - Arachne

2009-03-29 Thread Geraldo Netto
Hi all,

wouldn't be better porting dillo for freedos?
imho, the main bottleneck is the gui part that should be re written


http://www.dillo.org/

See Ya,

Geraldo
Sapere Aude
Non ducor, duco
São Paulo, Brasil, -3gmt
site: http://exdev.sf.net/
msn: geraldo_boca_at_hotmail.com
skype: geraldo-netto
icq: 145-061-456



2009/3/29 Eric Auer e.a...@jpberlin.de:

 Hi,

 - As I think xswap will not work in native Linux
 You will not need any of that if you simply compile Arachne
 with a 32bit compiler in the first place.

 Why it can't be also simple compiled on DOS with 32 bit compiler?

 That is what I meant - port Arachne to GNU C / DJGPP and
 then you have better performance in DOS - and an easier
 life in porting to Linux. Two good things for one effort.

 - Dr WebSpyder and Lineo Embrowser (unfortunately no source code)
 ported to Allegro and gave Arachne a speed boost.
 Arachne has GPL license so you can force WebSpyder/Lineo to
 make their Arachne modifications public.

 No, them have purchased a proprietary license from the copyright holders
 (original developers). GPL doesn't surrender rights. It's dual licensing
 like with Qt.

 Oh I understand. That is bad luck for fans then, fans
 will have to port Arachne to Allegro a second time if
 they want a free open source Allegro based Arachne...

 Eric



 --
 ___
 Freedos-devel mailing list
 Freedos-devel@lists.sourceforge.net
 https://lists.sourceforge.net/lists/listinfo/freedos-devel


--
___
Freedos-devel mailing list
Freedos-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/freedos-devel


Re: [Freedos-devel] Volunteering - Arachne

2009-03-29 Thread Bernhard Eriksson




Michael wrote:

  
Arachne has GPL license so you can force WebSpyder/Lineo to
make their Arachne modifications public.

  
  
No, them have purchased a proprietary license from the copyright holders
(original developers). GPL doesn't surrender rights. It's dual licensing
like with Qt.
  

Additionally Arachne wasn't GPLed when Michael Polak sold the code to
Caldera (who released DR-Webspyder) or Suntech. Orignally if you wanted
to get a hold of the source code you would need to sign a NDA and sent
it to Michael. I did so after I told Michael I had copied his code (I
still have an account at arachne.cz BTW), which influenced him to later
release it as GPL.
-- 
Bernhard Eriksson, Wermlandsdata
Fryxellsgatan 2, 652 22 Karlstad
054 - 15 69 00, http://www.wermlandsdata.se/
Datorer, tillbehr, service, programmering mm. 




--
___
Freedos-devel mailing list
Freedos-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/freedos-devel


Re: [Freedos-devel] Volunteering - Arachne

2009-03-29 Thread Christian Masloch
 - I am not sure whenever Udo Kuhnt's version uses 16 or 32 bit DPMI but
 it I think it's 16.

The difference isn't that big, anyway. Default code operation size  
changes, but this doesn't require changes to the actual code (even in .ASM  
source files) if the assembler or compiler supports both 16- and 32-bit  
output. The code segment can be larger than 64 KiB, which might be the  
main advantage.

 - 32 bit DPMI would be better.

Not necessarily. Some applications even run dual-mode (RM or 16-bit PM) so  
even 16-bit has it's advantages. The 64 KiB code segment limit is the only  
real disadvantage. (Or are code segments with 16-bit default operation  
size not actually limited to 64 KiB?)

Regards,
Christian

--
___
Freedos-devel mailing list
Freedos-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/freedos-devel


Re: [Freedos-devel] Volunteering - Arachne

2009-03-29 Thread Blair Campbell
 - 32 bit DPMI would be better.

 Not necessarily. Some applications even run dual-mode (RM or 16-bit PM) so
 even 16-bit has it's advantages. The 64 KiB code segment limit is the only
 real disadvantage. (Or are code segments with 16-bit default operation
 size not actually limited to 64 KiB?)

I agree that 32-bit DPMI is better because 32-bit is more common and
people with 286s are highly unlikely to be browsing the internet with
a graphical browser (if at all).  And in 16-bit code code segments
IIRC don't have to be limited to 64kb; for example if you were using a
compiler that supported huge pointers, it would generate function
calls to increment a pointer if you try to access anything  64kb in
one memory block, but that could considerable slow down an application
and of course the generated code size would be much bigger.  An
application utilizing huge pointers could potentially access 1 MB of
memory IIRC.

 Regards,
 Christian

 --
 ___
 Freedos-devel mailing list
 Freedos-devel@lists.sourceforge.net
 https://lists.sourceforge.net/lists/listinfo/freedos-devel


--
___
Freedos-devel mailing list
Freedos-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/freedos-devel