Linux-Advocacy Digest #609, Volume #33 Sat, 14 Apr 01 16:13:03 EDT
Contents:
Re: there's always a bigger fool (The Ghost In The Machine)
Re: NT is stagnant while Linux explodes (Roy Culley)
Re: there's always a bigger fool ("Ayende Rahien")
Re: OT: Treason (was Re: Communism) (Chad Everett)
Re: OT: Treason (was Re: Communism) (Chad Everett)
Re: Microsoft: Closed source is more secure (Chad Everett)
Re: Microsoft: Closed source is more secure (Chad Everett)
Re: To Eric FunkenBush (GreyCloud)
Re: To Eric FunkenBush (GreyCloud)
Re: To Eric FunkenBush (Chad Everett)
Re: hmm getting tired of this! (Chad Everett)
Re: Blame it all on Microsoft ("Joseph T. Adams")
Re: Undeniable proof that Aaron R. Kulkis is a hypocrite, and a (The Ghost In The
Machine)
Re: Blame it all on Microsoft (Chad Everett)
Re: OT: Treason (was Re: Communism) ("Joseph T. Adams")
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (The Ghost In The Machine)
Crossposted-To: alt.destroy.microsoft,comp.os.ms-windows.advocacy,soc.singles
Subject: Re: there's always a bigger fool
Date: Sat, 14 Apr 2001 18:50:24 GMT
In comp.os.linux.advocacy, Matthew Gardiner
<[EMAIL PROTECTED]>
wrote
on Sun, 15 Apr 2001 00:39:09 +1200
<[EMAIL PROTECTED]>:
>I updated my glibc from 2.1 to 2.2 and did not require a reboot.
This may illustrate why Linux works better than Windows; one issue
here is that executables requiring the services of glibc are
pointing to one copy of the file, whereas anything started up
after the installation of the new file will, unless they explicitly
specify 2.1 as a version (presumably, this is done during link time
or during dlopen() time by appending the version number), get
the new version.
Heaven help the sysadmin types if there are incompatibilities between
2.1 and 2.2. However, that's extremely rare, though I suspect
something like that has happened on Unix.
The only way to deal with that is to kill all daemons referring to
glibc -- unfortunately, that includes init. Reboot time! But
again, that's extremely rare -- and its even rarer because Linux
doesn't support the notion of a global shared data segment between all
processes using a certain library; if a process wants functionality
like that, one has to use mmap(), read it from a file, or otherwise
deal with this lack, and most Unix utilities functions are
downward-compatible, which has the downside of a number of
deprecated switches on occasion (make(1), for example,
ignores the switches -b and -m). One hopes in the case of mmap()
that all writers respect the original version and that the file
itself can be upgraded once the system reboots. (Note that glibc
implements mmap(), but AFAIK no call within glibc uses it.)
An example of a system which might have had shared-data per library
was the Amiga, which had true shared libraries (but no
memory mapping). However, it wasn't automatic; the library had to
set it up and maintain it itself.
By contrast, NT doesn't implement versioning on their DLLs beyond
the relatively crude 'msvcrt20.dll' and 'mfc30.dll'. Presumably,
the installer moves them out of the way, but the DLL loader -- I
think -- will load exactly one copy of a DLL, and, until needed,
will not unload it; NT therefore thinks that the old copy of the DLL
is the same as the new and won't load the new one. I think
the name is 'MSVCRT20' or 'MFC30'.
(Linux does the same, but the name is the full file pathname,
which means the control is finer and much more reliable. Also,
I think the device and inode gets involved in there somewhere; two
different files will not have the same device and inode.
I'd have to look, but AFAIK Linux can share physical code pages,
as can NT.)
>
>Matthew Gardiner
>
>Ayende Rahien wrote:
>>
>> "Matthew Gardiner" <[EMAIL PROTECTED]> wrote in message
>> news:[EMAIL PROTECTED]...
>> > Kelsey Bjarnason wrote:
>> > >
>> > > [snips]
>> > >
>> > > "Aaron R. Kulkis" <[EMAIL PROTECTED]> wrote in message
>> > > news:[EMAIL PROTECTED]...
>> > >
>> > > > Since there is no junior-high-designed "registry" to get corrupted,
>> > > > there is ZERO need to shut down other programs to install a new app.
>> > >
>> > > Umm... the registry is not the reason apps require reboots. Generally
>> > > speaking, there are three reasons an install wants to reboot:
>> > >
>> > > 1) It has attempted to update a locked file, such as a system file.
>> > > 2) It has installed something expected to run during the startup phase
>> of
>> > > operations
>> > > 3) It is installing a service (not in the NT sense, but in the sense of
>> a
>> > > background application or library) which needs to be launched in order
>> to
>> > > work
>> > >
>> > > One might suggest that in the case of #3, a better approach could be
>> used.
>> > > I agree. In the cases of #2 and #1, I'm not sure how, given such an
>> > > application being installed, Linux would handle this any better; how
>> exactly
>> > > does Linux handle live patching of the kernel, for example, without
>> > > rebooting?
>> > >
>> > Just in regards to:
>> >
>> > 1. Linux has most, if not all libraries installed required for most
>> > software, however, when more libraries need to be installed, no reboot
>> > is required. Under O/S 2 Warp 4, you had to reboot after applying a
>> > service pack (because it included low-level drivers, kernel updates
>> > etc), however, in an application sense, it should not be required.
>>
>> What about needing to update those libraries?
>>
>> > 2. Can be started then and there. Just like if I want to add support for
>> > a USB Zip drive, I simply drop into shell and type: insmod usb-storage,
>> > and voila, instant access to my hardware. I have installed Solaris
>> > patches and the only time you need to reboot is when the kernel has been
>> > updated, however, it is not forced, unlike Microsoft Windows.
>>
>> I don't think he meant this, there are several stuff that *need* reboot,
>> because they can't be done when the computer is running.
>> Converting the FS type of the system partition, or making changes to it,
>> frex.
>> I don't think that you can do that in linux without reboot either.
>>
>> > 3. Services, yet another thing that can be started on the fly. For
>> > example, I could start up Apache without needing to reboot. I could
>> > start Squid proxy with out a reboot.
>>
>> Dito for NT. There is nothing to prevent you from starting & shutting down
>> services whenever you like it.
>>
>> >
>> > What I am pointing out is that, if you take a kernel upgrade out of the
>> > equation, a OS should never need to be rebooted, and hopefully once
>> > hot-plug PCI becomes more mainstream, even needing to shut down the
>> > computer to install hardware will become a thing of the past.
>>
>> I fully agree.
>> Something to point out, though. Ninety nine precent of the applications that
>> want reboot, don't need it. It's something that had been there because of
>> 9x. And I certnaily close none of my applications to install new ones.
>> (Well, I log in as another user via TS, but that is another matter.)
>
>--
>I am the resident BOFH (Bastard Operater from Hell)
>
>If you donot like it go [#rm -rf /home/luser] yourself
>
>Running SuSE Linux 7.1 Pro w/ Kernel 2.4.2
>
>SuSE, the best of German engineering, now in software form :)
--
[EMAIL PROTECTED] -- insert random misquote here
EAC code #191 8d:10h:21m actually running Linux.
Use the source, Luke.
------------------------------
From: [EMAIL PROTECTED] (Roy Culley)
Crossposted-To: alt.destroy.microsoft,comp.os.ms-windows.advocacy
Subject: Re: NT is stagnant while Linux explodes
Date: Sat, 14 Apr 2001 19:52:16 +0200
Reply-To: [EMAIL PROTECTED]
In article <mXUB6.2153$[EMAIL PROTECTED]>,
"Kelsey Bjarnason" <[EMAIL PROTECTED]> writes:
> <[EMAIL PROTECTED]> wrote in message news:[EMAIL PROTECTED]...
>>
>> TCO for any Unix is small compared to Windows (what an absurd name for
>> server systems). Last year alone there were over 100 security bugs found
>> in Microsoft SW. The worst offender was IIS. Windows admins just can't
>> keep up. Each new service pack introduces yet more bugs. The number of
>> security bugs found in Unix systems is tiny compared to Microsoft and
>> patches are normally made available far faster than patches from Microsoft.
>
> The number of bugs _found_ is less than in Windows... but is the number of
> bugs that _exist_ less? What are the comparitive number of Unix systems
> (server and client) in use compared to the number of Windows systems (server
> and client) in use? Doesn't Windows clock in with just a few more
> installations?
Not in the Internet server market. Security in Microsoft server OS's
is appalling. On the desktop it is laughable.
------------------------------
From: "Ayende Rahien" <[EMAIL PROTECTED]>
Crossposted-To: alt.destroy.microsoft,comp.os.ms-windows.advocacy,soc.singles
Subject: Re: there's always a bigger fool
Date: Sat, 14 Apr 2001 21:58:07 +0200
"The Ghost In The Machine" <[EMAIL PROTECTED]> wrote in
message news:[EMAIL PROTECTED]...
> By contrast, NT doesn't implement versioning on their DLLs beyond
> the relatively crude 'msvcrt20.dll' and 'mfc30.dll'. Presumably,
> the installer moves them out of the way, but the DLL loader -- I
> think -- will load exactly one copy of a DLL, and, until needed,
> will not unload it; NT therefore thinks that the old copy of the DLL
> is the same as the new and won't load the new one. I think
> the name is 'MSVCRT20' or 'MFC30'.
Actually, one of the real troublemaker is MFC42.DLL, IIRC
You can get DLL's version without refering to the name, do you know if it's
used on NT?
------------------------------
From: [EMAIL PROTECTED] (Chad Everett)
Crossposted-To: alt.destroy.microsoft,us.military.army,soc.singles
Subject: Re: OT: Treason (was Re: Communism)
Reply-To: [EMAIL PROTECTED]
Date: 14 Apr 2001 13:48:24 -0500
On 14 Apr 2001 17:46:36 GMT, Roberto Alsina <[EMAIL PROTECTED]> wrote:
>billh <[EMAIL PROTECTED]> wrote:
>>
>>"T. Max Devlin"
>>
>>> >The truer translation is "You shall not murder". We've been through
>>this.
>>>
>>> I thought it was "you shall not slay." How do you know which it is?
>>
>>Take your pick. Killing in war is neither illegal, unethical, nor immoral.
>
>And here�s something, in your own words that may clear why you are full
>of odorous remainders of the digestive process:
>
>"killing in war is not illegal".
>
>a) That is not true in general
>b) That requires some standard: "law" against which it can be declared
> "legal" does it not?
>
>In 3000BC, the standard included massive rape of the women of the defeated
>and mass sacrifices of the defeated soldiers.
>
Well that's not true at all. Show us an example in 3000BC where
massive rape of the defeated women occurred. Since you throw
out such blatantly false statements as fact, everything you
say is questionable.
------------------------------
From: [EMAIL PROTECTED] (Chad Everett)
Crossposted-To: alt.destroy.microsoft,us.military.army,soc.singles
Subject: Re: OT: Treason (was Re: Communism)
Reply-To: [EMAIL PROTECTED]
Date: 14 Apr 2001 13:49:10 -0500
On 14 Apr 2001 17:35:48 GMT, Roberto Alsina <[EMAIL PROTECTED]> wrote:
>On 12 Apr 2001 23:05:57 GMT, Joseph T. Adams <[EMAIL PROTECTED]> wrote:
>>In comp.os.linux.advocacy T. Max Devlin <[EMAIL PROTECTED]> wrote:
>>: Said Joseph T. Adams in alt.destroy.microsoft on 12 Apr 2001 09:44:09
>>:>In comp.os.linux.advocacy Russianbear <[EMAIL PROTECTED]> wrote:
>>:>
>>:>: Bah - If there is a God he is no better than a common dictator and there is
>>:>: NO reason at all to worship him. Anyone who says live by my rules or be
>>:>: punished with death or eternal damnation is an asshole.
>>:>
>>:>
>>:>First of all, God has only two main rules, according to Jesus. First
>>:>is to love Him. Second is to love your neighbor.
>>:>
>>:>I don't think those are unreasonable requests.
>>
>>: Says who? According to God and Jesus (according to those who say
>>: 'according to God and Jesus'), there were plenty of other rules, and
>>: less reasonable requests.
>>
>>
>>Those are details. They're there to explain exactly what is meant by
>>loving God and loving your neighbor.
>
>How does eating fish on fridays connect to loving my neighbour?
>
You really have not idea what you're talking about, do you?
------------------------------
From: [EMAIL PROTECTED] (Chad Everett)
Subject: Re: Microsoft: Closed source is more secure
Reply-To: [EMAIL PROTECTED]
Date: 14 Apr 2001 13:54:44 -0500
On Sat, 14 Apr 2001 20:38:42 +1200, Adam Warner <[EMAIL PROTECTED]> wrote:
>Hi all,
>
>Security Focus is running this story of comments by the Head of
>Microsoft's security response team at the RSA Conference:
>
>http://www.securityfocus.com/frames/?content=/templates/article.html%3Fid%3D191
>
>I think some of the arguments are self-serving and utterly disingenuous.
>For example:
>
>`By contrast, Microsoft does extensive testing on every product, and on
>every patch, said Lipner. "People ask us why our security patches take so
>long. One of the reasons they take so long is because we test them."'
>
Microsoft has millions of testers, they're called end users.
>`Lipner, who oversees Microsoft's response to newly-reported security
>holes in its products, took the opportunity to point out "the repeated and
>recurring vulnerabilities in the Unix utilities BIND, WU-FTP, and so on.
>The repeated theme is people use this stuff, but they don't spend time
>security reviewing."'
>
So let's see, he is saying: Our security patches take so long because we
test them so well and we still have lots of security holes in Microsoft
products because Unix has bugs in BIND and WU-FTP?
------------------------------
From: [EMAIL PROTECTED] (Chad Everett)
Subject: Re: Microsoft: Closed source is more secure
Reply-To: [EMAIL PROTECTED]
Date: 14 Apr 2001 13:56:58 -0500
On Sat, 14 Apr 2001 20:38:42 +1200, Adam Warner <[EMAIL PROTECTED]> wrote:
>Hi all,
>
>Security Focus is running this story of comments by the Head of
>Microsoft's security response team at the RSA Conference:
>
>http://www.securityfocus.com/frames/?content=/templates/article.html%3Fid%3D191
>
>I think some of the arguments are self-serving and utterly disingenuous.
>For example:
>
>`By contrast, Microsoft does extensive testing on every product, and on
>every patch, said Lipner. "People ask us why our security patches take so
>long. One of the reasons they take so long is because we test them."'
>
TCP/IP sequence number guessing was laughably easy on NT for YEARS.
Microsoft knew it because it was reported to them very early on.
They NEVER fixed it until W2K came out.
------------------------------
From: GreyCloud <[EMAIL PROTECTED]>
Subject: Re: To Eric FunkenBush
Date: Sat, 14 Apr 2001 12:10:46 -0700
Erik Funkenbusch wrote:
>
> Actually, it has nothing to do with templates. This is a bug with friend
> functions combined with the using declaration, which have had some known
> issues.
>
> In any event, the issue here has nothing to do with MI, as you claimed. If
> you knew much about C++, you'd know this, and you could have recreated it in
> a much smaller sample.
>
> You can recreate it in the following simple code that doesn't use MI at all.
> To make it work, comment out the #include and the using namespace std:
>
> #include <iostream>
> using namespace std;
> class S
> {
> public:
> private:
> int a;
> friend bool operator>(const S& s1, const S& s2);
> };
>
> bool operator>(const S& s1, const S& s2){return s1.a > s2.a;}
>
> "Erik Funkenbusch" <[EMAIL PROTECTED]> wrote in message
> news:OOTB6.3188$[EMAIL PROTECTED]...
> > Your example uses templates. That is most likely the issue, not the MI.
> VC
> > is a pre-standard compiler, and as such cannot be expected to conform to a
> > standard that did not exist when it was written.
> >
> > I'll compile them and see.
> >
> > "GreyCloud" <[EMAIL PROTECTED]> wrote in message
> > news:[EMAIL PROTECTED]...
> > > I've found the programs in the book "C++ Primer Plus" by Stephen Prata.
> > >
> > > See attached source code.
> >
> >
> >
Then why does it compile and work correctly under g++ ??
The compiler errors are pointing to an MI problem. If you declare a
main class, then create two more classes that rely on the main class,
then create a third level class below all of these that require methods
from both upper classes, you are needing access to both classes private
members as declared as such. If MS C++ can't do it but g++ can, I see a
problem in MS compiler.... also in the book it provided MetroWerks
compiler as well and it too works on this example. In the book it
explains that these work under a lot of different UNIX compilers as well
and gives gcc or g++ commands and options.
Again, MS C++ compiler has a problem.
--
V
------------------------------
From: GreyCloud <[EMAIL PROTECTED]>
Subject: Re: To Eric FunkenBush
Date: Sat, 14 Apr 2001 12:12:45 -0700
Ian Pulsford wrote:
>
> GreyCloud wrote:
> >
> > I've found the programs in the book "C++ Primer Plus" by Stephen Prata.
> >
> > See attached source code.
> >
>
> This is almost as long as Kulkis' sig, surely you could have used email
> instead.
>
> IanP
What for... I was providing this to a pubic forum to hear from others
own experiences.
You do have g++ don't you?? How about VC++6.0??
Let the others decide how this works for them so that a determination
can be made.
Maybe VC++ has a switch that needs to be set.
------------------------------
From: [EMAIL PROTECTED] (Chad Everett)
Subject: Re: To Eric FunkenBush
Reply-To: [EMAIL PROTECTED]
Date: 14 Apr 2001 14:04:33 -0500
On Sat, 14 Apr 2001 03:24:16 -0500, Erik Funkenbusch <[EMAIL PROTECTED]> wrote:
>Your example uses templates. That is most likely the issue, not the MI. VC
>is a pre-standard compiler, and as such cannot be expected to conform to a
>standard that did not exist when it was written.
>
>
And what is amazing is that Microsoft does not have plans to ever release
a standards compliant C/C++ compiler.
------------------------------
From: [EMAIL PROTECTED] (Chad Everett)
Subject: Re: hmm getting tired of this!
Reply-To: [EMAIL PROTECTED]
Date: 14 Apr 2001 14:07:15 -0500
On Sat, 14 Apr 2001 15:22:47 +0000, Karel Jansens <[EMAIL PROTECTED]> wrote:
>Darin Johnson wrote:
>>
>> "Aaron R. Kulkis" <[EMAIL PROTECTED]> writes:
>>
>> > Anybody who claims to have "proved" Global warming is a fraud.
>>
>> And anyone who claims it doesn't exist is also a fraud! Both sides
>> should have lots of fraudness in common.
>
>As far as I can tell, global warming is indeed a fact.
>
>As to what _causes_ it, a whole lot of BS-sing is going on. [*]
>
>--
>Regards,
>
>Karel Jansens
Really? How can you tell? Cause it's warmer outside today than it
was yesterday?
------------------------------
From: "Joseph T. Adams" <[EMAIL PROTECTED]>
Crossposted-To: comp.theory,comp.arch,comp.object
Subject: Re: Blame it all on Microsoft
Date: 14 Apr 2001 19:21:22 GMT
In comp.os.linux.advocacy Goldhammer <[EMAIL PROTECTED]> wrote:
:>
:>Name one feature in MS Office that isn't offered in Star Office or Applix.
: Clippy ?
He said "feature," not "misfeature." :)
Joe
------------------------------
From: [EMAIL PROTECTED] (The Ghost In The Machine)
Crossposted-To: comp.os.ms-windows.nt.advocacy
Subject: Re: Undeniable proof that Aaron R. Kulkis is a hypocrite, and a
Date: Sat, 14 Apr 2001 19:23:33 GMT
In comp.os.linux.advocacy, Aaron R. Kulkis
<[EMAIL PROTECTED]>
wrote
on Sat, 14 Apr 2001 00:42:28 -0400
<[EMAIL PROTECTED]>:
>The Ghost In The Machine wrote:
>>
>> In comp.os.linux.advocacy, Aaron R. Kulkis
>> <[EMAIL PROTECTED]>
>> wrote
>> on Thu, 12 Apr 2001 18:25:08 -0400
>> <[EMAIL PROTECTED]>:
>> >The Ghost In The Machine wrote:
>> >>
>> >> In comp.os.linux.advocacy, Aaron R. Kulkis
>> >> <[EMAIL PROTECTED]>
>> >> wrote
>> >> on Wed, 11 Apr 2001 13:59:21 -0400
>> >> <[EMAIL PROTECTED]>:
>> >> >WesTralia wrote:
>> >> >>
>> >> >> "Aaron R. Kulkis" wrote:
>> >> >> >
>> >> >>
>> >> >> > One which has the full overhead of a function call, including
>> >> >> > the overhead of pushing the CPU state onto the stack, and
>> >> >> > recovering it at the end.
>> >> >> >
>> >> >>
>> >> >> "pushing the CPU state onto the stack"
>> >> >>
>> >> >> OK genius, explain what you mean!
>> >> >
>> >> >
>> >> >Pushing all data registers, pointers, and the program counter
>> >> >onto the stack.
>> >> >
>> >> >What did YOU think it means?
>> >>
>> >> Pointers? What pointers?
>> >
>> >
>> >Stack pointer
>>
>> Pushing the stack pointer on the stack seems to me rather pointless... :-)
>>
>> >Frame pointer
>>
>> Yes, that's a valid one. Of course, that's usually in a register.
>> (Side point: Oh gosh the VAX was elegant, from a machine language
>> point of view. Am I spoiled rotten? :-) )
>
>Yes, all of us who programmed VAXes were spoiled.
>
>I'm thinking of writing a VAX emulator.
Would be nice; I suspect someone's done it, but would be nice... :-)
>
>
>
>>
>> >Heap Pointer.
>>
>> A heap pointer needn't be saved; it's part of process address space.
>> Even if it is, it's just another register.
>>
>> >Data Segment pointer
>>
>> Ditto. (Maybe on a pre-VMpage system, though, like System7 on a
>> PDP 11/70. But that's positively *ancient*. :-) )
>
>If it's Intel, it's part of the CPU state.
Hm...you might be referring to CS, SS, DS, and ES, then (GS was introduced
later). I'd forgotten about those; AFAIAC they're more registers.
(Yuck, what an ugly memory model.)
>
>
>>
>> >Text Segment Pointer
>>
>> Ditto.
>>
>> Oh, and you forgot block segment start.... :-)
>>
>> >
>> >etc. etc.
>> >
>> >MORON.
>>
>> Well, excuse me for pedantifying. :-P
>
>Sorry, Ghost...I got you confused with someone else.
Ah, well, that's OK. :-)
>
>
>
>>
>> >
>> >>
>> >> That's tangential to storing the state of the micro during a context
>> >> switch or function call for later restore. A pointer is merely a
>> >> novel method of interpreting a number within a general register or
>> >> memory location as a memory address. (For that matter, a number is
>> >> merely a novel method of interpreting the bits within a general
>> >> register or memory address, and the bits within a general register
>> >> are a novel method of interpreting logic voltages, which are merely
>> >> novel methods of interpreting largish collections of electrons [*]:
>> >> 1 pF = about 6,250,000 electrons per volt.)
>> >>
>> >> Of course, one of the pieces of CPU state is the program counter.... :-)
>> >>
>> >> OK, so I quibble.
>> >
>> >You also just demonstrated that you don't know fuck about CPUs
>>
>> Oh really? :-P
>>
>> I at least know that pointers are merely a human way of relating
>> to bags o' bits. That was my point; as far as the micro is concerned,
>> it's pushing bits from a data register to the address buss.
>> A bit's a bit.
>>
>> Many CPUs for example don't care what's stored in their registers,
>> which means that one can do stupid things like
>>
>> MOV #1,R1
>> MOV 4(R1),R1
>>
>> and of course cause an access violation, but why should the
>> first instruction really care?
>>
>> I will note that the 68000, however, went the split-register route
>> (8 D registers, 8 A registers, two of which are dedicated to
>> SP (A6) and PC (A7), respectively.) In cases such as that, one
>> might be able to cause an exception (although why even bother?)
>> on an immediate load to an A register. However, HP PA uses
>> 32 general-purpose registers which can be used for pretty much
>> anything (although HP does have a calling convention, dedicating 5 registers
>> thereto (4 parameters, 1 return; extra parameters are pushed on the
>> stack as always), and register 0 is *always* zero). I don't know
>> what SPARC uses, but I suspect it's similar. Makes compiler writing
>> very simple; one doesn't have to worry about whether LEA 10(%ax)
>> makes sense when allocating %ax.
>>
>> And then there's ... errrm .... the x86 ... . I don't know whether
>> the 386 fixed all of the weidies regarding register usage or not.
>> (It probably did, but I do wonder.)
>
>x86 has the Data Segment, Text Segment, etc. registers.
>
>So also the IBM-370 programmer's model (who knows what the
>ACTUAL hardware looked lik...probably even changed over time,
>..but that was the model presented to the programmer by the
>microcode.
Can't say I know; never worked on that machine.
>
>
>
>>
>> >
>> >
>> >>
>> >> Note also that a smart compiler won't push all registers onto the stack;
>> >
>> >Depends on the CPU. Some CPU's don't give you any choice. Some do.
>>
>> I've yet to run into one at this point which doesn't
>> allow specification of which registers to push and pop.
>> To be fair, though, I really haven't coded in that many:
>>
>> 68000 (Motorola)
>> 8086 (Intel et al)
>> 1802 (Harris)
>> VAX 11/7xx (DEC)
>> 6502 (?)
>> HP PA RISC (HP)
>
>
>If you ever get a chance to look at IBM 370, you'll see the
>80x86 is derived from it.
I thought the 80x86 was derived from the 8080; it's source code
compatible and all. (It's been awhile.)
Maybe both participated; obviously, someone thought it a good idea
at the time, and it did give them an advantage over their competitors
(68000 came later, with a much better architecture IMO).
[rest snipped]
--
[EMAIL PROTECTED] -- insert random misquote here
EAC code #191 8d:12h:54m actually running Linux.
Use the source, Luke.
------------------------------
From: [EMAIL PROTECTED] (Chad Everett)
Crossposted-To: comp.theory,comp.arch,comp.object
Subject: Re: Blame it all on Microsoft
Reply-To: [EMAIL PROTECTED]
Date: 14 Apr 2001 14:20:04 -0500
On Sat, 14 Apr 2001 17:11:20 GMT, The Ghost In The Machine
<[EMAIL PROTECTED]> wrote:
>In comp.os.linux.advocacy, Aaron R. Kulkis
><[EMAIL PROTECTED]>
> wrote
>on Sat, 14 Apr 2001 01:28:26 -0400
><[EMAIL PROTECTED]>:
>>Chad Everett wrote:
>>>
>>> On Fri, 13 Apr 2001 17:09:49 GMT, The Ghost In The Machine
>>> <[EMAIL PROTECTED]> wrote:
>>> >In comp.os.linux.advocacy, Chad Everett
>>> ><[EMAIL PROTECTED]>
>>> > wrote
>>> >on 12 Apr 2001 17:09:51 -0500
>>> ><[EMAIL PROTECTED]>:
>>> >>On Thu, 12 Apr 2001 15:23:02 -0600, Jerry Coffin
>>> >> <[EMAIL PROTECTED]> wrote:
>>> >>>In article <[EMAIL PROTECTED]>, [EMAIL PROTECTED]
>>> >>>says...
>>> >>>
>>> >>>[ ... ]
>>> >>>
>>> >>>> Yes, and you can construct bad trees that require
>>> >>>> exponential time to search -- in reality it rarely
>>> >>>> happens [/me sticks his neck out on that assertation].
>>> >>>
>>> >>>You're not really sticking it out very far -- testing with
>>> >>>generational scavengers seems to agree quite closely.
>>> >>>
>>> >>>> If you view a program as a tree of objects,
>>> >>>> stemming from the root object, you would end up with a
>>> >>>> tree and not a list, albeit with circular references.
>>> >>>
>>> >>>Which is to say that it's a graph, not a tree. A tree would be an
>>> >>>acyclic graph, where this is a more general graph that may contain
>>> >>>cycles.
>>> >>
>>> >>A tree is an acyclic, connected graph, not just acyclic.
>>> >
>>> >Pedant point: so is a DAG. :-)
>>> >
>>> >A tree needs to include the requirement that nodes aren't entered
>>> >more than once (they can be exited as many times as required, however).
>>> >
>>>
>>> This is not a requirement of a tree. It is a property of a tree when
>>> it is explored. A tree is an acyclic, connected graph...period.
>>>
>>> Hey this is fun! :)
>>
>>Chad is correct.
>
>This is getting very off topic (so what else is new? :-) ) but
>I fail to see how
>
>A ----------> B ----------> D
> \ /
> ----------> C ------>---
>
>can be a tree....? Granted, one can spit the node D into two
>images: D1 and D2, but I'm obviously missing something very obvious
>here (this sort of thing also comes up in inheritance diagrams in C++
>and can get ugly fast).
>
You've introduced directed edges. And it is not really correct to have
a DAG where some edges have directions and others don't.
A tree is not a DAG, but some DAGs can be a tree.
------------------------------
From: "Joseph T. Adams" <[EMAIL PROTECTED]>
Crossposted-To: alt.destroy.microsoft,us.military.army,soc.singles
Subject: Re: OT: Treason (was Re: Communism)
Date: 14 Apr 2001 19:34:04 GMT
In comp.os.linux.advocacy Roberto Alsina <[EMAIL PROTECTED]> wrote:
:>:>First of all, God has only two main rules, according to Jesus. First
:>:>is to love Him. Second is to love your neighbor.
:>:>
:>:>I don't think those are unreasonable requests.
:>
:>: Says who? According to God and Jesus (according to those who say
:>: 'according to God and Jesus'), there were plenty of other rules, and
:>: less reasonable requests.
:>
:>
:>Those are details. They're there to explain exactly what is meant by
:>loving God and loving your neighbor.
: How does eating fish on fridays connect to loving my neighbour?
You may be confusing the teachings of a particular sect (the Roman
Catholic Church) with those of Christ. They're not necessarily the
same thing.
: Roberto Alsina (who is not going to ask about how not fucking women
: who want to do it implies love of God)
You may be confusing lust with love. Those are almost never the same
thing.
Joe
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list by posting to comp.os.linux.advocacy.
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Advocacy Digest
******************************
