Absolutely. Most of Unix is plain C, which translates well to Pascal. C++ is harder, of course.

"Gerard N/A" wrote

As with many things, it only depends of the time you have. But as you grow older you get a different idea of life expectancy and the use of your remaing time.

This is an argument only after considering the time versus the goal. Is the goal worth the time?

An operating system based on FPC code would be one way to give FPC more (well deserved) attention, one way to make a wider range of people know that the language is not dead, but significantly updated and a stronger alternative than most people think.

But it is the best way? An OS kernel is not the most visible source code that makes a lot of people gather to have a look. Showing off the technology can be done in more time-efficient ways.

Another goal is to make a better OS. FPC programs use less memory and starts faster. That can be a good thing. Would it be significant? I'm not sure.

You are perfectly right in that we must choose our battles and spend our time on things that are worth the time. I don't rule out that this could be one. "Porting Linux" doesn't have to imply that every little program is ported.

/Ingemar

If everybody does a little bit it could go quite quickly.

Question is:

Which linux distro ? ;)

Maybe even some BSD version so that closed-source os-es could be done as well.

Though I onced tried FreeBSD... I couldn't even figure out the gui or how to start an app..

So maybe linux more user friendly...

So first someone needs to find a good/user friendly linux distro which is not to large.

Also for kicks maybe a floppy-disk based linux distro could be turned into pascal as well.

So could even be multiple linux distro's turned into pascal...

Different flavors for everybody :)

Most important would be to be able to play with the to-be-pascal sources of the linux distro so people can try out things with the os... ;)

So focus would be:

1. Linux distro(s)/kernel(s).

2. Linux gui(s).

3. Linux editor for editing pascal files.

4. Free pascal compiler probably already available.

Then anything else can be turned into pascal later on :)

Maybe even drivers and applications and such ;)

Bye,
Skybuck.

Skybuck Flying wrote

If everybody does a little bit it could go quite quickly.

Some (arbitrary numbers) from http://www.ohloh.net/p/linux: Codebase 10,679,927 lines
Effort (est.) 3,396 Person Years

So, if everybody on this list (maybe 300 persons) work on it, then it can be done in just over 11 years. Not so quickly, IMHO.

Vincent

Ok, so lots of code...

Two possibilities come to mind:

1. Using automatic conversion from C to Pascal but then the code would still have to be checked by humans.

2. Only convert certain portions which are most interesting to people.

For example:

Linux's tcp/ip stack.

Linux gui.

So that for example Linux gui might be improved by pascal programmers ;) :)

Is that possible ? ;)

Bye,
Skybuck.

Hello all,

Although the effort is huge, and the topic is senseless, porting is much faster then creating/coding/debugging. The numbers above do not reflect a port and are pointless.

If you want to do some large work to increase the use of FPC I would recommend creating a Window Manager instead, probably with fpgui.

The work is imensely smaller, althougth still large, and a window manager usually comes with lot's of useful gui software, so this would be an opportunity to distribute all kinds of software made with fpc and lazarus.

My open source project, for example, the Virtual Magnifying Glass, suffers in linux because KDE only wishes to distribute C++/Qt software and Gnome only distributes C and C# software, so it get's hard to be popular.

And don't expect people from the list to join your initiative. Usually you have to start the project alone and get something working before one or two people join in. Don't expect everyone to stop what they are doing to help you in what you think is important.

Ingemar Ragnemalm wrote

This is an argument only after considering the time versus the goal. Is the goal worth the time?

No, and that was exactly my point. Life is too short to spend it rewriting Linux in Pascal. <g>

Another goal is to make a better OS. FPC programs use less memory and starts faster. That can be a good thing. Would it be significant? I'm not sure.

Then maybe it would better to focus on something different from Linux.

Not every little program, but only for the kernel + drivers + filesystems the task is daunting.

Gerard.

Hello Everyone,

While I can't honestly say I share any enthusiasm for writing a OS in FreePascal, I do think there is merit in looking for projects of similar scope and interest that can demonstrate the power of the language and the tremendous tools that have been developed by this group over the past four years.

I do think that if someone were to interested in testing their mettle for writing something like an OS, he or she might first take a crack at writing a few of the GNU tools in FP. Say for example the BusyBox suite. Once they had done that, there are a few more code foothills they could climb to build up enough additional undestanding and programming techniques others could then learn and then join them later on more ambitious OS projects.

My current projects are to port the FP compiler to the OpenSolaris platform and to eventually write an Eclipse plug-in oe a Mozilla based XUL application (like Komodo) for FP. Ultimately, I am interested in exploring more of the ARM platform support in FP.

Prince

Well, you said "porting linux". Linux to be only references the Kernel only. That is the real OS part. All the other stuff like X11, and 1000's of utilities and libraries are 3rd party stuff. Having a Linux system without a GUI is just as much a Linux system, as one with a GUI.

So you first need to decide what you want to port? The kernel only, or a whole distro? The latter is absurd, as it will take a lifetime to complete, plus - what's the point?

As for a OS that can boot a computer, implement a basic filesystem and copy/rename files is doable. I had friends that had to write such a simple OS as a end-of-year project in there studies. I guess you can take a peak at the Linux code to accomplish that, but again, what's the point. Other that being able to say "yes it can be done with Free Pascal as well". Maybe you have more free time available than I do. :-)

Regards,
- Graeme -

_______________________________________________
fpGUI - a cross-platform Free Pascal GUI toolkit
http://opensoft.homeip.net/fpgui/

Felipe Monteiro de Carvalho wrote

If you want to do some large work to increase the use of FPC I would recommend creating a Window Manager instead, probably with fpgui.

How far did you guys get with the 'fpwm' project? Did it actually run at some point. I see the last code changes was 2 years ago.

Regards,
- Graeme -

_______________________________________________
fpGUI - a cross-platform Free Pascal GUI toolkit
http://opensoft.homeip.net/fpgui/

This was true in the 70's. Then Pascal wasn't modular (C wasn't but hacked it in header files, and still does), it simply wasn't adapted for programming of that kind yet. And C was optimized for being similar to the machine code of the CPU's of the time. And it still is.

Today, the biggest difference between the languages are missing high-level features in C (which Pascal has had for decades), and that the syntax of C has a big range of obvious flaws. Neither of these make C a better language for anything.

The advantage to be able to mix operations on the same line, like the ++ operator, had importance for performance back then, before pipelining and caches. Today, it simply doesn't matter.

IMHO, C was simply the *first* language replacing assembly language on the OS level. And the first often gets chosen as "standard", the "only way".

Now, to return to the OS part: How could Linus Torvalds write the core of Linux in rather short time, single-handed, if it is such a huge task just to port it?

The window manager part is what I find really interesting. If Gnome is language-locked, there are two ways to change that: Either convince the Gnome team to open it up in our directions (just a matter of interfaces to them) or to make an FPC branch of Gnome (assuming that Gnome is under appropriate licenses), where appropriate parts are ported to FPC. I would call it "Jedi Gnome", analogous to Jedi-SDL. Now, is there anyone more than me visualizing Yoda with a lightsabre as logotype? :-)

/Ingemar

> How could Linus Torvalds write the core

of Linux in rather short time, single-handed, if it is such a huge task just to port it?

Because he didn't really write it from scratch, instead he just improves what's already there (Unix source code, maybe?) and poing! Linux comes up.

Skybuck Flying wrote

1. Using automatic conversion from C to Pascal but then the code would still have to be checked by humans.

I can just imagine it will generate ugly code. I would much rather imagine that a clean design with a good objects structure would be a much better solution in the long term.

2. Only convert certain portions which are most interesting to people. For example: Linux's tcp/ip stack. Linux gui.

You are jumping the gun here! You FIRST need a kernel. Then you need file system support. Then you can start with networking, utilities and GUI applications.

Regards,
- Graeme -

_______________________________________________
fpGUI - a cross-platform Free Pascal GUI toolkit
http://opensoft.homeip.net/fpgui/

Graeme Geldenhuys wrote

How far did you guys get with the 'fpwm' project? Did it actually run at some point. I see the last code changes was 2 years ago.

I think it was able to run, but wouldn't do much.

Too many other tasks at hand, and this one isn't making any money, so it has very low priority...

I just checked out fpwm and updated it and fpgui locally so it can compile. It does almost nothing atm but can reparent and close windows :) not alot. I was thinking I might write a widget for resizing and moving windows. Currently the window "title bar" is a Label and a Button that has an "X" for it's text and stdimg.quit as the image. I might work on it some in the coming week if I can find the time.

Regards,

Andrew

Hello

Like ti weigh in on this thread in the discussion regarding increasing the GUI-ness of FP.
Has anyone looked into writing an Eclipse plug-in for FP (as an was done for Python and Ruby)?

Given the huge base of Eclipse users this might be a way to reach the goal with both a better developer tool and plugging into a big audience of potential users.

Prince

Umm... I have though of trying my hand at the X11 tutorial on creating a simply window manager, but never got around to it. The last time (years ago) when I tried to compile 'fpwm', I managed to get it to compile, but didn't know how to test/run it.

Could you give some hints on the latter, or is there a readme file in the fpwm project that explains this? How do I temporarily test a new window manager? I'm using Ubuntu, and have quite a few window managers available from GDM (sessions menu), but I have no idea how those options got there, or how to add my own. :-)

Regards,
- Graeme -

_______________________________________________
fpGUI - a cross-platform Free Pascal GUI toolkit
http://opensoft.homeip.net/fpgui/

In our previous episode, Prince Riley said:

Afaik there was a more editor like plugin at one time. But at that time designer support only existed for C++ and Java. Haven't heard much since.

Given the huge base of Eclipse users this might be a way to reach the goal with both a better developer tool and plugging into a big audience of potential users.

Afaik most Eclipse users use Java?

Thank you for your reply.

Eclipse was first released as a Java development IDE several years ago, but if you take a look at the Eclipse web site www.eclipse.org you'll see that since then Eclipse has evolved and been extended to support Ruby, Python, C/C++.

The suggestion I was making is there is a very large base of programmers, developer, and others who now use Eclipse as their primary IDE in much the same way UNIX folks use emacs and
if there were a Eclipse FPC plugin they could discover and learn the language.

Prince

In our previous episode, Prince Riley said:

I know that, and that was not the point. You say "support", I say "usage".

That project is such a mess (my personal opinion). Just the other day I wanted to download the Eclipse IDE, to see what has changed over the last few years since I , after browsing their
website for 30 minutes, I still couldn't find a clear "download Eclipse IDE here" link! What's up with that!!! I simply gave up and moved on.

Regards,
- Graeme -

_______________________________________________
fpGUI - a cross-platform Free Pascal GUI toolkit
http://opensoft.homeip.net/fpgui/

Hi Graeme,

www.eclipse.org
On the upper menu, 5th item, "downloads".
Then choose whatever you wnat to download, and choose the mirror, and the download starts.
lazarus.freepascal.org on the other way seldom works lately and when it does it's so slow it can't be used.

Graeme,

I find that very interesting. The Eclipse group has redesigned the web site and it now has a much cleaner and easier to navigate 'front door' that leads most visitors directly to the major Java or Rich Client IDE.

Prince

Graeme

I just took this snapshot of the www.eclipse.org kanding page .. you say you weren't able to find a clear download link .... lol .. ok well it appears directly in the center of the page ...

Let me know what you think after you download Eclipse ..

Prince

I've seen people using Eclipse for developing in Fortran too...

I did the exact same thing. It gave me serious doubts about their ability to understand what people actually need. An IDE doesn't do anything intrinsically itself, it is an enabler. It is supposed to make things that you can already do easier. You'd think that easy downloading should be a part of that.

hmm. That's not very Freepascal related. How about if I say that Eclipse is an excellent example of the direction that FPC should not go :)

Hello,

Please also move this Eclipse discussion to fpc-other (unless it's specifically on how to set up Eclipse for use with FPC or so).

Thanks,

Jonas
FPC lists moderator

Well, this means only lazarus has too much success :)

I can't find a Eclipse for pascal either :>

Prince Riley wrote

I find that very interesting. The Eclipse group has redesigned the web site and it now has a much cleaner and easier to navigate 'front door' that leads most visitors directly to the major Java or Rich Client IDE.

OK, I just went to the www.eclipse.org and the website looks totally different to the one I was at a few months ago. Yes, it's much easier to find and download now. :-) Clearly they noticed their error in the previous web design.

Regards,
- Graeme -

_______________________________________________
fpGUI - a cross-platform Free Pascal GUI toolkit
http://opensoft.homeip.net/fpgui/

Prince Riley wrote

I just took this snapshot of the www.eclipse.org kanding page .. you say you weren't able to find a clear download link .... lol .. ok well it appears directly in the center of the page ...

hehehe... Yes it seems funny now, but like I said, the last time I checked the website didn't look like is how it did look (Using the Web Archive's WayBack Machine).

Archive in February 2008
http://web.archive.org/web/20080214194714/http://www.eclipse.org/

This one was not so clear. ;-)

Regards,
- Graeme -

_______________________________________________
fpGUI - a cross-platform Free Pascal GUI toolkit
http://opensoft.homeip.net/fpgui/