Mac OS X 10.4.7 Seeded

[FF_productions]

I usually don't rely on my Mac for the calculator feature such as adding 695.22222324982989262476240672046029 + 228940208.000248240842=blah blah...
I'll just use an actual calculator like a TI-84 Calculator if anything...

 

[Lixivial]

In a previously fixed bug, adding a BCC field after you had already started a message would crash Mail and lose the message. Thank goodness that one was fixed. I still feel that moment of hesitation before adding or removing the BCC field, but it's now been working reliably for a couple of Mail versions.

Interesting. I believe I recall hearing something about this bug, but I've only recently migrated fully to Mail so I never encountered it. I migrated around 10.4.3 or so, when I did a final push to Tiger and opted to converge my gmail, hotmail and personal accounts together into one program.

Anyway, the bug I'm referring to is a bug that I've had since I began with Mail, and it's still happening in 10.4.6 (Mail v. 2.0.7). The bug is pretty easy to replicate, it simply requires a user to have a signature auto-attaching to new emails/replies/forwards.

Here's how to recreate it:

1. Create a new message, place your cursor to the left of the first character in your signature and press [return] a few times.

2. Begin typing anywhere between where you signature originally was and where it is now. This [where it originally was] should likely be, if it's a blank message, the second line. This is because Mail automatically adds a line just above the signature where a user writes their message.

3. Finally, in the signatures drop-down select: "None" and your message should disappear, along with the signature. My guess is that Mail assigns some sort of hidden character to the line where the signature begins and where it ends, so that when a user prompts to change their signature it knows what lines to remove/alter.

I have an issue because I like to create extra padding between my message and whatever pre-existing text there is as I'm creating the message. I usually forget to change the signature, too, and it's the worst in replies where I'm editing the replied message to coincide with my response and I create space to paste the stuff into. User error? Yeah. Program error? That too.

 

[ejl10]

An update for syncing? Like maybe iSync? Like maybe it will support my hot new Samsung T509? Or is that just a pipe dream?

Its frustrating that some of the nicest, most Apple-esque phones aren't supported by iSync or Salling Clicker. Especially when I see that cute commercial with the hip Mac guy saying everything just works with a Mac.

 

[Catfish_Man]

But programmers (like me) are dumb enough to think they are smarter then the accumulated brain power of 5,000 years worth of wheel making experts and we build or own from scratch and they break.

QFGT. I now force myself to look for libraries implementing stuff (or papers discussing implementation issues) before writing it. Even some relatively sophisticated stuff has issues with this; iirc Java doesn't support -0, for example, and that messes up certain very specific types of apps.

 

[thefunkymunky]

Can anyone who can get access to 10.4.7 and have an early 15-inch MBP tell me wether it adds the two-finger right-click functionality to the trackpad?

Later build 15-inch MBP (Week 16/17 onwards) already have this.

 

[netdog]

Can anyone who can get access to 10.4.7 and have an early 15-inch MBP tell me wether it adds the two-finger right-click functionality to the trackpad?

Later build 15-inch MBP (Week 16/17 onwards) already have this.

I keep reading that you have it now.

Edit...sorry, didnt see last bit.

 

[Malcster]

3. Finally, in the signatures drop-down select: "None" and your message should disappear, along with the signature. My guess is that Mail assigns some sort of hidden character to the line where the signature begins and where it ends, so that when a user prompts to change their signature it knows what lines to remove/alter.

I have experienced this one too, i reported it to Apple through the OSX Feedback page a long time ago. (10.3. something..) It is still there.

Maybe we should both re-submit it!

 

[savar]

You'd think calculator would be a pretty easy app to get right...

A computer does floating point arithmetic in base 2, but the calculator does floating point arithmetic in base 10. This leads to some subtle issues, and if the calculator was written by some intern or otherwise less experienced developer, its possible that there could be minute approximation errors. I doubt this is anything serioius.

Edit: OK I read that linked thread, if that's all true then there are some serious problems. Looks like the developer really wasn't thinking when he wrote this app.

 

[savar]

I think there are only two options
(1) Write the software in such a way that you can _prove_ it's correctness. This is very hard and requires PhD level computer scientests on staff

That's a bit of an exaggeration. Proving an algorithm is no different than proving a mathematical theorem, and millions of students across the world are doing that every day.

 

[billyboy]

Try this in Calculator:

654,654.321 - 987.312

You'll get 653,667.00899999996 instead of the correct 653,667.009

(weird, copy pasting the result out of Calculator gave me the correct number)

It's a side effect of the binaire nature of computers.

I got the correct answer. Do I have a gifted calculator?

Also if I paste your numbers, because I have a Spanish keyboard, your number reads as 653.654321

 

[ChrisA]

That's a bit of an exaggeration. Proving an algorithm is no different than proving a mathematical theorem, and millions of students across the world are doing that every day.

Yes, I can prove things like the "prime number therom" the one that says all integers can be factored to a unique set of primes. But have you seen the proof of the five color map problem. I think it went on to 270 pages. That was just for a simple thing like coloring a map such that no two areas that share a border have the same color. Now try to prove something harder: that a C compiler will accept all valid C programs or even something so simple as that the Ptreads library will never deadlock. OK you want an easy one: Prove that a calculatoer will give correct answers for all possable inputs.

I claim that the calculaor proof is well past the ability of almost all working profesional software engineers. You can prove me wrong only be example: Ask a large group of experianced professionsals who have only a four year degree in CS, or math if they can do the calculator proof. If more then a handfull come up with a valid proof that stands up to peer review then you've proved me wrong.

I am familar with a project that tried to build a provably correct non-triveal system. It was to be a kind of "micro kerrnal" that would control a message exchange They had about 200 software engineers working for two to three years. Nothing but technical problems basically no one knows how to do this except for toy problems. Amoung the problems are how to verify the correctness of a proof that runs on to thousands of pages. Who can read it? but the hardest part is writing a formal specification. You need to invent a mathmatical language that can specify what a non-trieal progoram, lkie a micro kernel is to do (and not do) and then your proof is that your code will meet the spec for all possable input files. Today there are very few (maybe zero) people who can do this for even the calclator. Again prove me wrong by listing more than a few names.

 

[electronboy]

I really hope they fix the annoying errors we get when saving Word and InDesign files to AFP volumes! 10.4.5 and 10.4.6 were a complete disaster. Don't upgrade!

 

[Rocketman]

Try this in Calculator:

654,654.321 - 987.312

You'll get 653,667.00899999996 instead of the correct 653,667.009

(weird, copy pasting the result out of Calculator gave me the correct number)

It's a side effect of the binaire nature of computers.

Of course since it is a "rounding error", the solution is to round the answer itself by a digit or two.

You simply cannot display as many digits as are actually calculated.

Rocketman

P.S. A Mars probe hit "really hard" because of an error of units, pounds to Newtons.

 

[artifex]

Y'know, my iMac G5 has been having some sleep issues lately as well. Never associated it with that update, but it probably was around that same time...

It's just worried that it's going to be replaced, soon

 

[dextertangocci]

I hope they update ical! Every time I try to switch to week view, it just unexpectedly quits! Month and day view work fine, and I've tried everything from repairing permissions to reinstalling the app. Nothing works! Anyone knows whats wrong?

 

[howesey]

I think you misunderstood him: computers are calculators so a calculator function is the easiest app to build, compared to say, image editing. Image editing uses mathematical calculations to create and manipulate images, much more complex.

any programmer (I am not one) knows that a good place to start programming is to create a simple calculator. the difference here is that apple are looking at inaccuracies in incredibly complex calculations which take the computers to their limits.

Computers can only add though. They cannot divide, multiply, subtract or any other function. All the results have to be calculated by adding. It's more complicated than asking a computer to do 00100 - 00001, as they cannot do this.

However, it's odd it gets the wrong results, I wrote a calculator at school when I was 12 in IT classes using BASIC, FORTRAN and machine code, and they got the correct results.

 

[Arbiter]

I also suspect that the calculator is internally convertint the number to an IEEE standard SP or DP number wich can cause some problems unfortunately. I have had similiar problems with matlab and it was ver weird.. I read numbers from a file and did some calculations on them. However when I inserted them by hand I would get some other result.. the difference was very very small but there was a difference between them.

 

[Plymouthbreezer]

I usually don't rely on my Mac for the calculator feature such as adding 695.22222324982989262476240672046029 + 228940208.000248240842=blah blah...
I'll just use an actual calculator like a TI-84 Calculator if anything...

Hah. You beat me to suggesting that if you're requiring this kind of stuff, why not just use a high-end calculator?

 

[steve_hill4]

They could very well go to 10.4.10, 10.4.11, ..., 10.4.100 etc. Do not confuse decimal numbers with version numbering.

Whilst I agree, I also think that we are getting closer to Leopard. We may end Tiger with 10.4.12 for arguments sake, but I think Leopard will be possibly announced and start shipping around the time of 10.4.9.

It strikes me that Apple are releasing a lot of updates at the moment. This is both good and bad, but while it appears that way, I don't think they are coming any faster than they did with Panther or Jaguar. Can anyone confirm dates for those minor update cycles, (e.g 10.2.3, 10.2.4 etc)?

 

[Peace]

Leopard will go out to developers at WWDC2006 in August.Not before.
Builds of Tiger will probably go up to 10.4.9.

 

[JFreak]

When will we be able to download 10.4.7?

whenever it's ready. what's wrong in waiting?

 

[dextertangocci]

whenever it's ready. what's wrong in waiting?

A few of my apps, like ical, are a bit corrupted, and I'm hoping that they get fixed with 10.4.7.

 

[bigandy]

A few of my apps, like ical, are a bit corrupted, and I'm hoping that they get fixed with 10.4.7.

umm, why not just reinstall them?

 

[CTYankee]

Nope, I get 653667.009

>Hmmm, I get 4 8 15 16 23 42...

 

[dextertangocci]

umm, why not just reinstall them?

Yes, that's what I thought I could do. So reinstalled them, and the problems still remain. Like with ical, it just randomly quits, and I can't go to week view without it quiting unexpectedly. Tried repairing permissions etc, so I have no idea whats wrond.