OSD - IDE

[Solar]

An IDE would be used to implament these new ideas into an existing os structure...

Won't work. My ideas don't fit into "an existing os structure", and neither do those of several others here. You would be targeting only those people willing to be limited by the OS structure superimposed by your IDE... unless they ignore your IDE and use the OSKit / OSLib instead...

...providing useful code snippets and code wizards to do menial tasks...

The useful code snippets are in the Wiki. And... code wizards? Like "I want an O(1) scheduler" and there's your code?

The IDE would be to speed thing up, and help stop errors (with the wizards).

I don't think so. It would be a huge workload to get even a marginally useful IDE put together, and I don't think the benefit will outweight the effort.

Also with a lot of hard work a code checker could be created to check for common errors...

A great many of those made in C can be caught by lint and its ilk already. The rest are, as Curufir pointed out, perhaps not errors at all... we're speaking of one hell of artificial intelligence here.

Do you have any idea of how difficult it can be to just parse a programming language, let alone looking for semantical errors?

[KieranFoot]

It wouldn't be a error as such it would appear in a sugestions box n the IDE.

anyway iif you used that code:-

Code: Select all

push ax
push bx
...
jmp 0000:0fc0

once you get to that code you would still need

Code: Select all

pop bx
pop ax
...

[Candy]

It wouldn't be a error as such it would appear in a sugestions box n the IDE.

anyway iif you used that code:-

Code: Select all

push ax
push bx
...
jmp 0000:0fc0

once you get to that code you would still need

Code: Select all

pop bx
pop ax
...

If I may suggest these two:

Code: Select all

push eax
push ebx
...
iret

and

Code: Select all

push eax
push ebx
...
pop ebx
pop eax
iret

Which one is good? (hint: both).

[KieranFoot]

true...

well then you can look for the nearest label above the push command, and find any call references to that label and then look for the pops underneath... This might actually take a bit of time...

It was just and idea it might still prove usefull.

[KieranFoot]

in reply to solar, yes i do know how hard it is to write a programming language parser i tried t make my own for a small language of my own, but failed. Until i found a good program called gold parser builder, this creates a parser which can be exported to many languages and runs from an activex dll.
So it would be possible to have the parser file for most common languages built into the OSIDE...

This would be a bit trickey in the case of ASM becuse of machine dependent codes... but basic 8080/0806 assembler could be supported.

[Pype.Clicker]

actually, one of the goal of "high-level" languages is to ease the detection of incorrect structures by making them "explicit" ... you cannot apply things like "check if the stack frame has been cleaned when exitting the procedure" without telling the langage what is the stack frame and when you're exitting it ...

Virtually all the good editors have code-coloring features ... All the good compilers have warning that prevents you from doing auwful things ... What kind of things a dedicated IDE could add ? Warning because "you're not acknowledging interrupt to 8259a before IRET" ?

[KieranFoot]

It wouldn't check for really complicated things, it would check for basic things and then report them in a hints/warning box, which may be toggled on or off

[KieranFoot]

So heres what the IDE will do so far....

• SIMPLE Code error checking and warning system (ASM & C)
• Source file Categorization
• Source versioning system
• Template file headers
• Simple ASM Wizards
• Interrupt wizard (ASM ,C, C++)
• Generation of linker scripts
• ToDo subprogram
• Usefull hardware documents eg, int reff, VESA, APM...
• Quick reference library
• Full reconfigurable build system
• Project managment
• Generate BOCHS ini files

And much more as i think of it, all options will be deselectable

As many people r saying this IDE is doing alot of simple stuff which thay deen unnecessary, well in answer, this IDE will help to open operating system development up to new people & new ideas/concepts...

[piranha]

I like the Idea of an IDE for OSdev. There is a problem: If it was really easy to make an OS (like you are indirectly suggesting) then everyone and their aunt Sally will try to make one. This isn't bad, but most people would waste time, and Free web hosting would be overridden with these projects...

I would use it though, good idea. I just feel like being critical today.

[Solar]

If it was really easy to make an OS (like you are indirectly suggesting) then everyone and their aunt Sally will try to make one.

You mean, worse than it already is?

You might note this thread is two and a half years old...

[AndrewAPrice]

How about creating an image? Like a FAT12 floppy image with grub, the include the kernel.

The IDE should allow seperate user projects, a way for seperating libraries away from programs. A more mature OS should move their userland programming over to a standard IDE, not a OS IDE.

[pcmattman]

An IDE for OSDev is like ISO putting standards on operating systems. It'll just make all the important code the same and basically do everything for you.

I think someone should just create a website with pre-made disk images and tutorials and such. But hey, there is one, it's called osdever.net...

[Tyler]

An IDE for OSDev is like ISO putting standards on operating systems. It'll just make all the important code the same and basically do everything for you.

I think someone should just create a website with pre-made disk images and tutorials and such. But hey, there is one, it's called osdever.net...

You really need to work on your comparisons. How is ISO putting a standard on OS' even slightly similar to building a development enviroment for the purpose of OS Development. Your comparison is akin to the following - "Buttering Bread with a knife is like ISO putting standards on bread making ingredients." Also, read up on IDE's, they help the development process but they don't do any development.

On a topic related note, an IDE for OS Development would still be mostly standard IDE. It would be more useful to design some add-ins (i don't know enough about any IDE's to be sure what i am talking about here) that add functionality specific to OS. For example addition of Virtaul Machine/Emulator debugging integration, far more powerful Linking system for defining the numerous Binaries the project will produce in a fashion skin to Linker Scripts and, well anything else anyone would think was useful for OS Development.

[pcmattman]

In the context of what everyone else was saying, such as wizards for code and automatically creating disk images, it felt to me like creating a standard.

I shouldn't have said 'IDE for OSDev', instead I should have made it clear I was talking about the actual features people wanted.