I kinda agree. But I think it's the obliviousness that will haunt us more than the language spec itself. Mostly my comment was just meant to reduce the number of people oblivious to this mess.
It is nice that I can map a "int" to a register on DSPs (instead of having to have my 20-bit architecture emulate a 32-bit one with 2 registers).
The worst thing is that C++, AFAIK, is still even more of a mess than C in this regard.
At least with C, it is possible to write portable code. With C++, it's much much harder. Since the C++ standard apparently still ignored this issue, you're best off using the (awesome!) Boost libraries that bring the best parts of C to C++, such as their
cstdin.hpp
header file that takes care of fixed-sized types.-
If it doesn't, that's really a shortcoming in the kernel itself.
C is intentionally a little vague about the sizes of ints.
Using your 37-bit CPU example:
If all you're doing is manipulating values from 1 to 1000, you really don't want your C compiler doing software-checks for 32-bit overflow and underflow every time you use your registers. You really want a type that matches the native register.
If this kernel you're trying to compile really needs a variable with math that has predictable 32-bit underflow and overflow, you want a int_32t.
Parent wrote:"The C Standard says nothing about the sizes of any of the base types in comparison to any of the various sizes a 64-bit platform might choose to support, rather referring only to comparisons: char may not be longer than int, short may not be longer than int, long may not be shorter than int, et cetera."
Nonsense. ANSI C's description of requires that int's have a range of INT_MIN to INT_MAX, with a maximum value for INT_MIN being -32767 and INT_MAX being 32767. This implies a requirement of at least 16 bits for an int. (Yes, both end in a '7', but this still implies at least 16 bits).
Similarly the range between LONG_MAX and LONG_MIN imply at least 32 bits for a long.
While we're at it, the C standard (ISO/IEC 9899, or C99) also specifies a range for
LLONG_MAX >= 9,223,372,036,854,775,807 and
LLONG_MIN
For more...
http://www.ssec.wisc.edu/~dglo/c_class/fundamental _types.html
http://std.dkuug.dk/JTC1/SC22/WG14/
http://www.dinkumware.com/htm_cl/limits.html
http://home.tiscalinet.ch/t_wolf/tw/c/c9x_changes. html
PS: It's been years since the standard came out... but many people still think K&R C, which AFAIK had many fewer restrictions.
"int" should but in't required be whatever type is native to the processor; but is required to be at least 16 bits (thanks to a requirement that it can hold values from INT_MIN to INT_MAX whose minimum and maximum values are -32767 and 32767 respectively.
If your code cares beyond that, use the standard C types (defined in stdint.h) that specify the sizes.
From the many year old C99 standard (part B.17 Integer types ):
If you want an exactly 64-bit type use an int64_t.
If you want an exactly 32-bit type, use an int32_t.
If you want a fast type that's at least 32 bits, use a int_least64_t.
Not for a long time. Considering Yahoo Finance thinks they have $2.5 Billion cash, they'll be around a while. And with a market cap of > $13.8 Billion they can just buy some small company like Apple (who'se only worth about half as much as sun).
Then they'd have a BSD based OS comfortable for those of us who liked good old SunOS 4.X.
Parent post wrote: "Dr. Forking is extremely healthy"
The good news is that even if our dear doctor gets sick and can't work on his open source project anymore -- other developers can take over the project.
I don't think Dr. Forking's health is really that big a thread to open source, since it's so easy for others to take over. Now if it were commercial software, it might hvae been harder.
" There should NEVER be an api to mess with the phone numbers and dialing."
You have got to be kidding. That's _ALL_ I want my phone to do.
Just let me control the software that deals with those APIs by explicitly installing it; instead of having the instant-message-chatroom-client that i never wanted install junk for me.
It seems to me Solaris hasn't really been focusing on the desktop, so I'm guessing they're trying to sell the message: "Solaris is the OS for Servers and Linux is the OS for Desktops".
Quite a few companies I've seen actually do work this way - engineers use Linux for development, while QA and production servers are on some commercial Unix. If you are in such a development environment, your choices are either (a) Suns on the desktop (expensive), (b) shared Sun for developers (doesn't scale well), (c) Linux for developers (some pain 'porting' to QA, but cost effective).
Apples would work well in this environment too, now, so I wouldn't be surprised to see Sun buying Apple if their Linux efforts don't pan out.
With HP (hpux, pa-risc) and Dec (ultrix/alpha) both apparently going Linux, it'd be pretty cool to see the two last non-Wintel-Linux computer companies join forces.
"Look at Blender. I find the UI frustrating because I have never learned it before."
Many of the special-purpose and/or high-end packages have "interesting" interfaces that deviate from Windows/Gnome/Mac/whatever. Consider other 3d programs like Maya, or the higher-end video editing programs like Pinnacle Edition and you get some very wierd (at first) UIs. Use them for a while, and you realize that what was wierd at first becomes _very_ efficient for the special purpose the tool is trying to address.
I think it'd be smart for Wal*Mart to make it support both Linux & Microsoft -- just to have a better negotiating position when trying to get attractive OEM windows pricing.
Interesting to see if they need Microsoft more or less than Microsoft needs them for this product.
Frantic global purchase of notebook personal computers has prompted Hewlett Packard (HP) and Dell, the world's two leading PC brands, to increase their purchases in Taiwan.
The two companies' combined procurements from Taiwan are expected to top NT$1 trillion (US$29.7 billion) in 2004."
It's as valuable an "education" as any for the CS department to support an OS across their enterprise.
Would I rather hire an OS guru who'se experience was "yeah, I learned how to call RedHat's tech support number" or "I learned how to support the OS across the whole campus"?
"and once they were retrieved, the market value of their contents would plummet."
Not if the owners of the asteroid controlled the supply. Perhaps like DeBeers supposedly manages the price of diamonds and OPEC supposedly manges oil prices.
http://java.sun.com/people/jag/green/ "Java, whose original name was Oak, was developed as a part of the Green project at Sun. It was started in December '90
Looks like they were asking for one of Gosling , Naughton, or Sheridan.
parent wrote: To take another recent example, jobs that are requiring an SAT score.
According to the
San Jose Mercury News "Google once required SAT scores of job candidates". And if you're trying to hire "only the smartest of employees", that test sounds as reasonable as many other possible tests of smartness.
if a computer is compromised, never believe the logs.
Depends on the logs. IMHO you should have syslog set up to forward to a more secure system so if your desktop's compromised, you may still have logs that aren't.
Slashdot Mirror
Free PCs, but not a free lunch: "Free-PC.com of Pasadena, California, plans to give away 10,000 Compaq ".
It is nice that I can map a "int" to a register on DSPs (instead of having to have my 20-bit architecture emulate a 32-bit one with 2 registers).
The worst thing is that C++, AFAIK, is still even more of a mess than C in this regard.
At least with C, it is possible to write portable code. With C++, it's much much harder. Since the C++ standard apparently still ignored this issue, you're best off using the (awesome!) Boost libraries that bring the best parts of C to C++, such as their cstdin.hpp header file that takes care of fixed-sized types.-
C is intentionally a little vague about the sizes of ints.
Using your 37-bit CPU example:
If all you're doing is manipulating values from 1 to 1000, you really don't want your C compiler doing software-checks for 32-bit overflow and underflow every time you use your registers. You really want a type that matches the native register.
If this kernel you're trying to compile really needs a variable with math that has predictable 32-bit underflow and overflow, you want a int_32t.
PS: Is that 37-bit example real? I know of ones with 24 bit (some motorolla dsps), 20 bit (Zoran ZR 38001), and 40-bit (Texas Instruments 67XX.
Wonder what they say about other GM fish like these fish that were genetically engineered to grow faster
"int_64t" where we require 64 bits, and
"int_32t" where we require 32 bits, and
"int" where we require a native size >= 16 bits.
More trivia, Some cool processors (like some TI DSPs) have 40-bit "long" datatypes, so sizeof(long)==5. Pretty cool?
Nonsense. ANSI C's description of requires that int's have a range of INT_MIN to INT_MAX, with a maximum value for INT_MIN being -32767 and INT_MAX being 32767. This implies a requirement of at least 16 bits for an int. (Yes, both end in a '7', but this still implies at least 16 bits). Similarly the range between LONG_MAX and LONG_MIN imply at least 32 bits for a long.
While we're at it, the C standard (ISO/IEC 9899, or C99) also specifies a range for
LLONG_MAX >= 9,223,372,036,854,775,807 and
LLONG_MIN
For more...l _types.html. html
http://www.ssec.wisc.edu/~dglo/c_class/fundamenta
http://std.dkuug.dk/JTC1/SC22/WG14/
http://www.dinkumware.com/htm_cl/limits.html
http://home.tiscalinet.ch/t_wolf/tw/c/c9x_changes
PS: It's been years since the standard came out... but many people still think K&R C, which AFAIK had many fewer restrictions.
If your code cares beyond that, use the standard C types (defined in stdint.h) that specify the sizes.
From the many year old C99 standard (part B.17 Integer types ):
Then they'd have a BSD based OS comfortable for those of us who liked good old SunOS 4.X.
"Dr. Forking is extremely healthy"
The good news is that even if our dear doctor gets sick and can't work on his open source project anymore -- other developers can take over the project.
I don't think Dr. Forking's health is really that big a thread to open source, since it's so easy for others to take over. Now if it were commercial software, it might hvae been harder.
My driver is a parody of their driver. :-)
You have got to be kidding. That's _ALL_ I want my phone to do.
Just let me control the software that deals with those APIs by explicitly installing it; instead of having the instant-message-chatroom-client that i never wanted install junk for me.
It seems to me Solaris hasn't really been focusing on the desktop, so I'm guessing they're trying to sell the message: "Solaris is the OS for Servers and Linux is the OS for Desktops".
Quite a few companies I've seen actually do work this way - engineers use Linux for development, while QA and production servers are on some commercial Unix. If you are in such a development environment, your choices are either (a) Suns on the desktop (expensive), (b) shared Sun for developers (doesn't scale well), (c) Linux for developers (some pain 'porting' to QA, but cost effective).
Apples would work well in this environment too, now, so I wouldn't be surprised to see Sun buying Apple if their Linux efforts don't pan out.
And www.masters.org (a golf tournament) will be called what?!?
With HP (hpux, pa-risc) and Dec (ultrix/alpha) both apparently going Linux, it'd be pretty cool to see the two last non-Wintel-Linux computer companies join forces.
Many of the special-purpose and/or high-end packages have "interesting" interfaces that deviate from Windows/Gnome/Mac/whatever. Consider other 3d programs like Maya, or the higher-end video editing programs like Pinnacle Edition and you get some very wierd (at first) UIs. Use them for a while, and you realize that what was wierd at first becomes _very_ efficient for the special purpose the tool is trying to address.
Interesting to see if they need Microsoft more or less than Microsoft needs them for this product.
http://www.taiwanheadlines.com/20030929/20030929b5 .html
"Sep 29 2003... Taiwan procurements by HP, Dell to top NT$1 trillion: analysts
Frantic global purchase of notebook personal computers has prompted Hewlett Packard (HP) and Dell, the world's two leading PC brands, to increase their purchases in Taiwan.
The two companies' combined procurements from Taiwan are expected to top NT$1 trillion (US$29.7 billion) in 2004."
It's as valuable an "education" as any for the CS department to support an OS across their enterprise.
Would I rather hire an OS guru who'se experience was "yeah, I learned how to call RedHat's tech support number" or "I learned how to support the OS across the whole campus"?
Not if the owners of the asteroid controlled the supply. Perhaps like DeBeers supposedly manages the price of diamonds and OPEC supposedly manges oil prices.
Reply with the the email addreses of other spammers :-)
http://java.sun.com/people/jag/green/
"Java, whose original name was Oak, was developed as a part of the Green project at Sun. It was started in December '90
Looks like they were asking for one of Gosling , Naughton, or Sheridan.
According to the San Jose Mercury News "Google once required SAT scores of job candidates". And if you're trying to hire "only the smartest of employees", that test sounds as reasonable as many other possible tests of smartness.
For those who didn't get the 'funny' part, check out these guys with internet explorer.
Depends on the logs. IMHO you should have syslog set up to forward to a more secure system so if your desktop's compromised, you may still have logs that aren't.
Microsoft has some interesting pages on Configuring UNIX Computers to Forward Syslog Messages .
And here's a doc on how to How to Monitor Windows NT from Unix
(I don't know enough about msft, but I assume they can also remotelly log their own events).
"cat
or
"copy
or
'open("/dev/hda1",O_RDONLY)'
even if it doesn't have a filesystem.
What makes you think otherwise?
Try "sudo strings /dev/hdaX" on your swap device. It'll work fine, and give you fun insights into privacy you might not have had.