[lgerbarg]

Well, having been at Apple for the tail end of 10.1's development, I thought I would just share some thoughts. A few of these are just general things, a few are comments I would have made in the previous weeks, biu I have a policy of not posting in threads that are discussing pirate builds, so I often have no way to post a response in context.

wget:

curl is included and has the same functionality

desktop patterns:

I actually wrote code that can change a desktop pattern at an arbitrary time (instead of at login, or when finder quits, which is what the current dekstop switchers do). If anyone wants the sample code I will see about getting it to them.

16-32-64 bitness:

There are a whole bunch of misconceptions about this. When people talk about the transition from 16 bit to 32 bit it is almost always within a purely intel context. The issue has nothing to do with size of the address or data busses, but the fact that a whole bunch of new instructions, especially new memory management related ones were added to the chip. Thus moving to using these new instructions was done at the same time as moving to 32 bit, making the move very difficult (especially since the intel architecture is very messy, and the memory segmented management model, while clever, was not used natively by any OS).

The macintosh always had "32 bit" chip. The 68000 was internally 32 bit, but they only exposed 24 address lines. That meant that while pointers were 32 bits, only 24 bits of them were ever used. The problem was that when people had a pointer they would sometimes stuff some other data into the 8 unused bits. This meant that when chips that actually used those last 8 address lines came out the code that put stuff into those 8 bits of the pointers would cause errors. Note that if this had not been done, there never would have been an issue about stuff being 32 bit clean code, and all the old stuff would have run just fine.

64 bit support on PPC is a very interesting thing. There are two specs, the original PPC64 spec, which some high end IBM chips use (I think), and Book E (book E also has purely 32 bit implementation spec). I am not going to comment on it more than that, but you can go and get the information if you want.

Anyway, I hope you guys like 10.1 ;-)

Louis

[juanvaldes]

Thanks alot Louis! IT's great to have you around again. Things must have been stressful there these past few months. Send our thanks to the whole team!

BTW: can you tell us what you were working on in particular? You mention you were working on code for a desktop pattern randomize. Is that all you worked on? naturally if what you were working on is yet unreleased...I won't tell.

[lgerbarg]

Originally posted by juanvaldes:
<STRONG>Thanks alot Louis! IT's great to have you around again. Things must have been stressful there these past few months. Send our thanks to the whole team!

BTW: can you tell us what you were working on in particular? You mention you were working on code for a desktop pattern randomize. Is that all you worked on? naturally if what you were working on is yet unreleased...I won't tell. </STRONG>

Nah, that was just something I did in my free time to play with system notification system, not a product, my time, my toy, and no proprietary info in it.

I currently am working in Low Level integration, basicly as a hired gun. I do what needs to get done. Mostly I worked on an internal tool we use for debugging kernel panics. Not terribly sexy, but very useful (and soem of the nastiest code I have ever written, everytime I want to test something I need to cause a kernel panic ;-)

Louis

[Xeo]

Originally posted by lgerbarg:
<STRONG>everytime I want to test something I need to cause a kernel panic ;-)</STRONG>

So what was your tried-and-true method of causing a kernel panic? You must have come up with some certain way in all your testing.

Perhaps two SMB devices mounted at one time?

[lgerbarg]

Originally posted by Xeo:
<STRONG>So what was your tried-and-true method of causing a kernel panic? You must have come up with some certain way in all your testing.

Perhaps two SMB devices mounted at one time? </STRONG>

I have a kernel extension I wrote called Panicker ;-) Kernel extensions can do what ever they want, causing a panic once you are loaded into the kernel is as easy as:

panic("We are panicking here");

I may have put it on the MacHack CD, I don't remember.

Louis

[muchfresh]

I really enjoy getting intelligent technical information. thanks.

I have a couple questions

1) what benefit would a 64bit OSX provide. Other than support for large memory blocks I can't think of any.

2) mysql in 10.0.4 doesn't quit when told to. I read in the darwin Q and A page that pthread_kill (?) function isn't implemented. does it work in 10.1

3) how did they make window resizing faster?

4) There is a big debate on weather renice works or not. can you shed some light

thanks again

[ZnU]

Originally posted by muchfresh:
<STRONG>I really enjoy getting intelligent technical information. thanks.
4) There is a big debate on weather renice works or not. can you shed some light
</STRONG>

I'm not a Darwin developer, but I can answer this one. Renice works in 10.1. The most obvious way to test this is to renice the DVD Player app to something with a low priority and start doing things in other apps. Skip city, but the rest of the system gets a lot more responsive.

[lgerbarg]

Originally posted by muchfresh:
<STRONG>I really enjoy getting intelligent technical information. thanks.

I have a couple questions

1) what benefit would a 64bit OSX provide. Other than support for large memory blocks I can't think of any.

2) mysql in 10.0.4 doesn't quit when told to. I read in the darwin Q and A page that pthread_kill (?) function isn't implemented. does it work in 10.1

3) how did they make window resizing faster?

4) There is a big debate on weather renice works or not. can you shed some light

thanks again</STRONG>

1. Typically thats all it is really useful for. I am not really qualified to comment on it, and I have no idea what our, IBM, or Mot's plans on the matter are.

2. I know that was an issue, I am not personally aware whether or not it still is.

3. No idea, and even if I knew I probably couldn't say...

4. It works.

Louis