Maybe we should do a port of the gnu tools like rust is doing... lazarus / 21-General ---------------------------------------------------------------------- nomad · 2025-05-08 19:22 Please do not laugh at my suggestion. Its just a crazy idea. Free pascal has been around a long time now and well established like c and c++. If rust is doing this then I believe it could be done in free pascal. This is a good place to test free pascal in a operating development setting. If the porting of the gnu tools can be done in free pascal then the next thing would be to port the linux kernel into free pascal. Learn from all the mistakes linux made in the past. There has to be a good way to port c/c++ drivers to free pascal. This would be wonderful if it could happen and the license could be LGPL like the lazarus ide. Make it free, open and commercial for everyone. I am not linux kernel expert. I am just making a suggestion. At least if we could port the kernel and gnu tools to free pascal then we could have a full fledged terminal server to host websites and stuff. I hope the idea makes sense.  https://www.youtube.com/watch?v=UKCKPEbOjJU ---------------------------------------------------------------------- PascalDragon · 2025-05-08 19:37 Quote from: nomad on May 08, 2025, 09:22:38 pmPlease do not laugh at my suggestion. Its just a crazy idea. Do you even remotely have an idea how much is involved with these tools? Let alone the Linux kernel? Yes, you can port the tools, but aside from being able to learn things through that there isn't really any benefit in doing this. And writing an operating system is always an operating system is always very educational, but to even remotly reach a functionality like Linux provides will take decades. ---------------------------------------------------------------------- nomad · 2025-05-08 20:06 As I said. It is just a crazy idea. I was watching how the rust developers are attempting to port the gnu tools. Just saying maybe free pascal could be used to do the same since it is a mature and established language. ---------------------------------------------------------------------- nomad · 2025-05-08 20:14 I find it to be very sad no one is writing a decent home use desktop operating system in free pascal. Free pascal is a excellent language that has been battle tested for a long time now. Just saying. ---------------------------------------------------------------------- nomad · 2025-05-08 20:29 Another reason is I would like to have a operating systems that is posix compatible that is not under the gnu license. Something like what freebsd has. I like the LGPL license lazarus has. I am sure if there is a operating system written in free pascal that is stable and allows small companies to make commercial products then they would surely use it instead of linux. that is the whole point of the video. the companies want to bypass the gnu license. I wish I had the skills to pull this off. Maybe I will make an attempt int he future. I am currently maintaining my own small linux distro. Its nothing fancy. I am waiting for all the woke nonsense to come to an end with that is happening in the linux community. This iso is out of date but is a prototype of what I have. Been working on programs written in lazarus for the next release of the distro. I wish I could have a operating system base that allows me to have all the goodies like linux has without the commercial restrictions. https://kjv-endtimes.neocities.org/ ---------------------------------------------------------------------- 440bx · 2025-05-08 20:31 Quote from: nomad on May 08, 2025, 10:14:33 pmI find it to be very sad no one is writing a decent home use desktop operating system in free pascal. Free pascal is a excellent language that has been battle tested for a long time now. Just saying. As PascalDragon "hinted" to you, you seem to believe that writing a "decent home use desktop operating system" is a "little" one year project for someone who's got nothing else to do with their time. If you want a rough idea of the amount of work involved, look at the ReactOS code.  There are a large number of highly talented programmers who have invested _years_ of their time to get to the current beta version. Just gathering the manpower for such a project would be quite a challenge. Heck, the wine project, which is a quite a bit simpler than writing an O/S is already quite a project to undertake. ---------------------------------------------------------------------- nomad · 2025-05-08 20:43 I understand what you are saying. I look at the posts on this forum and I see very very talented programmers here. Lots of them. There are operating systems written in free pascal. Since you bring up wine then I am sure we could implement wine for windows applications, implement a posix translation layer for linux apps like what freebsd does and we have dosbox for dos applications. Just saying it is not impossible. Maybe one of these projects could be setup to do this. https://wiki.freepascal.org/Operating_Systems_written_in_FPC ---------------------------------------------------------------------- paule32 · 2025-05-09 09:01 for a Homework and for such that working with Windows 10/11, there could be a good idea, to written a Desktop Shell with Login and Icons... So you remove the shell item (svchost.exe or: (explorer.exe which loads the Desktop + Window Manager) ) from the registry to start this Desktop Shell, when Windows starts and go into the loggin Session. But wait ! - It is a better Idea, to "not" simply replace Windows Shell by the Coders own. - It is a better Idea, to written a Frameless Window, in which you simulate the Operations of the Operating stuff as you like. You will realize, that the posix Filepath's (also /C/Users/) will be work like (C:\Users\) together. If you replace win32api Calls through your own, then you have a Change for the next step. Till then, happy coding... paule32 ---------------------------------------------------------------------- Thaddy · 2025-05-09 09:48 /mnt/c/users/ /mnt and all lowercase here. ---------------------------------------------------------------------- paule32 · 2025-05-09 10:28 yes. this is true. But Windows old Artefact does not caseinsentive at make or read directories. The only different Part is: C:\oldWay   <-- or:  C:\OldWay  are the same. So, you can not create a File or Directory with the same Name and different letter spell. Since Windows goes to the Internet - since win95, it used/supports Posix Path's like: /C/oldway/  <-- or:  /c/oldway   are the same one, too. You can learn or doing Posix Files under windows with MSYS2 (MinGW) 32- and 64-Bit. Or near the Border of Posix, you can use CYGWIN, wich is for Windows, too. With CYGWIN, you can port the most used Software. Because the ./configure File is supported there (what MSYS2 does not). for old Codebase, you can try: 1. # ~/configure 2. # make 3. # make install for new Codebase, you can try: # cd /c/source && mkdir build && cd build && cmake .. Please don't blame we - I don't know all in detail - I am a Hobbit, not Pro ---------------------------------------------------------------------- zeljko · 2025-05-09 11:15 Question about topic title: Who are we ? ---------------------------------------------------------------------- marcov · 2025-05-09 11:21 Also I think that imitations that don't have  a market waiting will not drive FPC forward.  Rust devels had already infiltrated various Linux and GNU projects, and could provide an audience for what they made (and a motivation, buffer overflows). What would a FPC port provide, and who would use it?  p.s. Paule32:  IMHO such things (cygwin) are the worst of both world. Few things recompile out of the box without windows specific release engineering going into it, and you are forever using some strange hybrid and dealing with the fallout. If you want a Linux compatible, use Linux (or OS X if you are kinky). It makes no sense to build a house of cards on top of Windows. ---------------------------------------------------------------------- Thaddy · 2025-05-09 11:40 @Marcov There is also the "just because I can" issue, which is often tempting.... I guess Rust also started like that regarding to the GNU tools. I for one did the simplest ones many moons ago. ---------------------------------------------------------------------- marcov · 2025-05-09 11:47 People that love that kind of stuff can work on making pascal versions of build tools for windows.  dlltool, AS,AR,LD and windres have been already done. And with fpdebug another one, GDB. Cleaning out the remaining 2005 GNU binaries from the Windows ports would be a good thing. Things like grep would also be intellectually challenging enough. ---------------------------------------------------------------------- Thaddy · 2025-05-09 14:48 There is a grep by the late Rudy Velthuis - I helped a little, old -, but it follows Borland, not GNU syntax and was never satisfactory. There is a pascal grammar for that too somewhere in my house, because I wrote it. The grammar is OK. Our efforts were only released in a very small circle, some of the Borland/Inprise beta-testers and developers. He also did a whereis, that is still on his website. I also did some smaller ones like ls, chown and chmod but heavily windows oriented and I must admit only 60-70% compatible and needs elevation (called the same, does not always work the same), also very old NT/XP time stuff. All of this is at least 25+ years old and on a Quantum BigFoot...(Have to find an IDE interface for it) ---------------------------------------------------------------------- domasz · 2025-05-09 16:57 Well, you don't have to build a clone of Debian. One can start with making a clone of BusyBox: https://pl.wikipedia.org/wiki/BusyBox ---------------------------------------------------------------------- PascalDragon · 2025-05-11 14:14 Quote from: nomad on May 08, 2025, 10:43:12 pmI understand what you are saying. I look at the posts on this forum and I see very very talented programmers here. Lots of them. And those people have more worthwhile things to do than what you are suggesting, not to mention who says that they would even be interested in developing something like this? Developing command line utilities or operating systems is not everyone's passion, just like compiler development. ---------------------------------------------------------------------- bobby100 · 2025-05-12 16:04 Quote from: 440bx on May 08, 2025, 10:31:36 pmQuote from: nomad on May 08, 2025, 10:14:33 pmI find it to be very sad no one is writing a decent home use desktop operating system in free pascal. Free pascal is a excellent language that has been battle tested for a long time now. Just saying. As PascalDragon "hinted" to you, you seem to believe that writing a "decent home use desktop operating system" is a "little" one year project for someone who's got nothing else to do with their time. If you want a rough idea of the amount of work involved, look at the ReactOS code.  There are a large number of highly talented programmers who have invested _years_ of their time to get to the current beta version. Just gathering the manpower for such a project would be quite a challenge. Heck, the wine project, which is a quite a bit simpler than writing an O/S is already quite a project to undertake. I still wonder how did Robert Szeleney developed SkyOS as a single developer. I know the project is dead, but just look at the huge mass, produced by one man in just a couple of years. ---------------------------------------------------------------------- 440bx · 2025-05-12 16:35 Quote from: bobby100 on May 12, 2025, 06:04:46 pmI still wonder how did Robert Szeleney developed SkyOS as a single developer. I know the project is dead, but just look at the huge mass, produced by one man in just a couple of years. Last thing I ever want to do is not give due credit to a talented and hard working individual but, while it is quite likely that he did a lot of the work, it's also quite likely that some of the code is derived from other code.  Also, while the first release may have been done in a year or two, it's only after a number of years that the O/S had genuine potential. As I said, I don't want to take credit away from a talented individual but, the project died because he couldn't keep up and, that's no surprise.  It's quite a feat that he accomplished that much but, a _capable_ O/S isn't a one-man project (unless you're talking about MS- DOS or some other "barely an O/S" O/S.) One of things about Linux that is often overlooked is the _huge_ contribution made by GNU.  Without all the GNU stuff, which was developed by more than one individual, the O/S would be close to useless (if not totally useless.) ---------------------------------------------------------------------- marcov · 2025-05-12 16:54 Quote from: 440bx on May 12, 2025, 06:35:03 pm One of things about Linux that is often overlooked is the _huge_ contribution made by GNU.  Without all the GNU stuff, which was developed by more than one individual, the O/S would be close to useless (if not totally useless.) (there was also tooling by BSD. It couldn't be taken then since BSD was involved in lawsuits, but in 1994 that was settled, so not having GNU would have been a delay of 2-3 years at the most) ---------------------------------------------------------------------- VisualLab · 2025-05-12 19:18 Quote from: bobby100 on May 12, 2025, 06:04:46 pmQuote from: 440bx on May 08, 2025, 10:31:36 pmQuote from: nomad on May 08, 2025, 10:14:33 pmI find it to be very sad no one is writing a decent home use desktop operating system in free pascal. Free pascal is a excellent language that has been battle tested for a long time now. Just saying. As PascalDragon "hinted" to you, you seem to believe that writing a "decent home use desktop operating system" is a "little" one year project for someone who's got nothing else to do with their time. If you want a rough idea of the amount of work involved, look at the ReactOS code.  There are a large number of highly talented programmers who have invested _years_ of their time to get to the current beta version. Just gathering the manpower for such a project would be quite a challenge. Heck, the wine project, which is a quite a bit simpler than writing an O/S is already quite a project to undertake. I still wonder how did Robert Szeleney developed SkyOS as a single developer. I know the project is dead, but just look at the huge mass, produced by one man in just a couple of years. I think another OS project called AtheOS, developed by Kurt Skauen (inspired by BeOS), had a similar story. After the author abandoned it, several other programmers tried to develop it further, under the name Syllable. It is no longer being developed either. ---------------------------------------------------------------------- VisualLab · 2025-05-12 20:40 Quote from: nomad on May 08, 2025, 10:14:33 pmI find it to be very sad no one is writing a decent home use desktop operating system in free pascal. Free pascal is a excellent language that has been battle tested for a long time now. Just saying. Yes, I agree, a decent OS for the desktop (workstation) would be useful. Quote from: nomad on May 08, 2025, 10:29:20 pmAnother reason is I would like to have a operating systems that is posix compatible that is not under the gnu license. At this point I don't understand you. Why POSIX on the desktop? After so many decades, POSIX on the desktop is needed like a "pain in the ass". POSIX may be useful on some servers and routers, because they still use systems derived from UNIX. And there OK, it makes some sense. But on the desktop? It's almost a museum piece! Assuming someone would like to create a modern desktop OS, basing it on POSIX after so many decades after its development makes no sense whatsoever. At most, it makes sense to borrow some ideas from it (and very general ones at that). Quote from: nomad on May 08, 2025, 10:29:20 pmIts nothing fancy. I am waiting for all the woke nonsense to come to an end with that is happening in the linux community. That there will finally be a "year of Linux on the desktop"? That will never happen (a long time ago I also had such naive hopes, but I quickly realized that this would not be the case). It is impossible, for at least several reasons: 1) architecture - it is based on the assumptions of a system developed for mainframe computers (Unix), over 50 years ago! Drivers for many devices are in the kernel instead of in user space, the kernel is a bloated behemoth (blob), archaic and annoying directory system, configuration files scattered across various directories, different formats of configuration files, mess in accessing devices and services from the programmer's point of view (especially through reading "virtual" text files), lots of different strange script files launching various processes in the system, graphics and multimedia support is not part of the OS, 2) it uses an ancient programming language (C), which has some advantages (its compiler generates optimized machine code), but also has many disadvantages (e.g. macros, headers, archaic and clumsy build system), the main kernel programmer would rather "rip someone's heart out" than let C be replaced by some other programming language, 3) The community of developers and hobbyists developing this OS doesn't give a s..t about issues like: (a) ergonomics for the average user, (b) backward compatibility, (c) stable API for hardware drivers, (d) stable API for developers. Making another (nobody needs it) version of Linux using FPC/Lazarus would be an incredible waste of time and electricity. ---------------------------------------------------------------------- paule32 · 2025-05-13 05:02 for a minimal system, you can take a look at my repro from 2022: https://github.com/paule32/JustFunOS_win32 It was created on Windows 10/11 with Windows Tools (MSYS2). Since COFF .o bject Files are used, you can link with .o bject Files created by FPC ? Happy coding