Re: HEADS UP: ACPI CHANGES AFFECTING MOST -CURRENT USERS
> > Show us a suitable LISP interpreter, then. > > $ cd ~/lang/Scheme/tinyscm-1.27 > $ size scheme >textdata bss dec hex filename > 6134244763480 69298 10eb2 scheme Is that statically-linked? I'm curious to know the size of the bootloader forth footprint. The loader is about 150k, so I'm sure you could probably fit a nice Scheme interpreter in under that size... ?? > Tinyscheme is a mostly complete R5RS Scheme (R5RS is the You can also conditionally-compile the components to make a smaller footprint. I'm highly in favor of Scheme replacing 4th... It's a very easy language to learn (only 11 special forms) yet still powerful (you can't pass code as data in BASIC ;). If you replace the boot loader interpreter, pick Scheme over LISP. There are lots of implementations: siod, scm, mit-scheme, MzScheme, and tinyscheme are among the better ones. --Rick C. Petty, aka Snoopy [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: HEADS UP: ACPI CHANGES AFFECTING MOST -CURRENT USERS
> > Is that statically-linked? I'm curious to know the size of the bootloader > > forth footprint. The loader is about 150k, so I'm sure you could probably > > fit a nice Scheme interpreter in under that size... ?? > > ie. almost all of the size is the dictionary/runtime library. I'll bet it's comparable to a tiny, stripped-down implementation of Scheme.. Only one way to find out... ;) > It's quite hard to beat this, and to be frank, Scheme's syntax is not much > better than Forth's. 8) That's debatable. At least it's consistant & makes sense. Syntax is only an argument of preference. I like Scheme better than LISP because there's less syntax to learn. But the original concern was not of syntax but of the number of committers who know the language. I'll bet there are quite a few who know/love Scheme. I think that if a choice is made, to move to Scheme over LISP because in theory it should have a smaller footprint. Not that it makes a significant difference so long as the loader fits nicely on /boot and out of the way of the loaded kernel (which loads at over 1 MB). --Rick C. Petty, aka Snoopy [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
Re: HEADS UP: ACPI CHANGES AFFECTING MOST -CURRENT USERS
> > > 6134244763480 69298 10eb2 scheme > > > > Is that statically-linked? I'm curious to know the size of the bootloader > > forth footprint. The loader is about 150k, so I'm sure you could probably > > fit a nice Scheme interpreter in under that size... ?? > > Dynamically linked. Here is the statically linked size: > > $ size scheme >textdata bss dec hex filename > 127659 110929236 147987 24213 scheme Hmm, if it's stripped down a bit, it might fit nicely in the loader, replacing that 40k libficl mess.. ;) > Here is the /boot/loader size for comparison sake: > > textdatabss dec hex > 4096147456 0 151552 25000 > But ultimately someone has to do the actual work for this to > go beyond mere wishful thinking. I'd be happy to help out > (but not take on the whole task) if anyone braves the > naysayers :-) I suppose I could volunteer for this. I've been dissecting the loader for months now and hitting the 4th "fence" has been bothersome.. As far as braving those pesky naysayers, I thought about doing it on my own anyway so if no one wants the change, I'll just keep it for my own systems. =) If nothing else, I'm very curious to see how small I can get a Scheme implementation.. --Rick C. Petty, aka Snoopy [EMAIL PROTECTED] To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message
