Linux-Advocacy Digest #393, Volume #30 Fri, 24 Nov 00 09:13:04 EST
Contents:
Re: Windoze 2000 - just as shitty as ever ("Tom Wilson")
Re: Windoze 2000 - just as shitty as ever ("Tom Wilson")
Re: Windoze 2000 - just as shitty as ever (Giuliano Colla)
Re: Uptime -- where is NT? (Stuart Fox)
Re: Uptime -- where is NT? (Giuliano Colla)
Re: Windoze 2000 - just as shitty as ever ("Ayende Rahien")
Re: Windoze 2000 - just as shitty as ever ("Ayende Rahien")
Re: Uptime -- where is NT? (Giuliano Colla)
Re: Linux for nitwits ([EMAIL PROTECTED])
Re: Windoze 2000 - just as shitty as ever (Curtis)
Re: Windoze 2000 - just as shitty as ever (Curtis)
Re: Windoze 2000 - just as shitty as ever (Curtis)
----------------------------------------------------------------------------
From: "Tom Wilson" <[EMAIL PROTECTED]>
Crossposted-To: alt.destroy.microsoft,comp.os.ms-windows.advocacy
Subject: Re: Windoze 2000 - just as shitty as ever
Date: Fri, 24 Nov 2000 11:12:33 GMT
"Ayende Rahien" <[EMAIL PROTECTED]> wrote in message
news:8vlgh2$4tt57$[EMAIL PROTECTED]...
>
> "Erik Funkenbusch" <[EMAIL PROTECTED]> wrote in message
> news:HHiT5.10167$[EMAIL PROTECTED]...
> > "Nigel Feltham" <[EMAIL PROTECTED]> wrote in message
> > news:8vk112$4pche$[EMAIL PROTECTED]...
> > > All of this reminds me of an old win95 bug (may also be in 98 and
> NT/2k):-
> > >
> > > Place any file on the desktop then rename this file to use all 255
> > available
> > > characters (win9x supports 255 char filenames so I would expect them
to
> > > work reliably). Then try to delete the file (or do anything with file
> from
> > > any
> > > application). This causes a GPF error in whatever application tried to
> > > access the file. Another case of MS incompetance.
> >
> > Oh, of course. Bugs don't exist in anyone elses code. Right?
>
> Not in mine, I call them spiders. :-)
Entomology doesn't enter in to MY code - Catholicism does.
Can you say "Pea Soup"? <g>
--
Tom Wilson
A Computer Programmer who wishes he'd chosen another vocation.
------------------------------
From: "Tom Wilson" <[EMAIL PROTECTED]>
Subject: Re: Windoze 2000 - just as shitty as ever
Date: Fri, 24 Nov 2000 11:16:38 GMT
===== Original Message =====
From: "Nigel Feltham" <[EMAIL PROTECTED]>
Newsgroups:
alt.destroy.microsoft,comp.os.linux.advocacy,comp.os.ms-windows.advocacy
Sent: Thursday, November 23, 2000 2:04 PM
Subject: Re: Windoze 2000 - just as shitty as ever
> All of this reminds me of an old win95 bug (may also be in 98 and NT/2k):-
>
> Place any file on the desktop then rename this file to use all 255
available
> characters (win9x supports 255 char filenames so I would expect them to
> work reliably). Then try to delete the file (or do anything with file from
> any
> application). This causes a GPF error in whatever application tried to
> access the file. Another case of MS incompetance.
I'll give MS the benefit of the doubt here. Anyone anally retentive enough
to use a filename 255 characters long deserves a good GPF.
PS: Sorry about the personal email. Brain fart on my end!
--
Tom Wilson
Linux User
Windows NT User
Windows 95/98 User
For our CARNIVOROUS FBI friends:
Kill Gore, Kill Bush, Free McVeigh, Turner Diaries, Oswald, Long Live
Saddam, Bomb, Ammonium Nitrate, Revolution, Guns, Nuke Florida, Anthrax,
Samples, The Alien at Roswell, Kennedy, Ransom, Demands, The Drop, and
Hoffa got loose again! And, per request: Free Kevin!
PS: You've been screening this shit since the eighties. Have you caught
anybody yet?
>
>
>
>
------------------------------
From: Giuliano Colla <[EMAIL PROTECTED]>
Crossposted-To: alt.destroy.microsoft,comp.os.ms-windows.advocacy
Subject: Re: Windoze 2000 - just as shitty as ever
Date: Fri, 24 Nov 2000 12:19:29 GMT
Ayende Rahien wrote:
>
> "Tom Wilson" <[EMAIL PROTECTED]> wrote in message
> news:%OqT5.2513$[EMAIL PROTECTED]...
>
> > The point is, the entire OS doesn't go down in flames as a consequence of
> > the GUI crashing... Sure, you lose whatever it might have been you were
> > working on, but core processes other systems on your net might be using
> > don't have to be interrupted by the recovery. Plus, you don't have to deal
> > with the corrupted file system and registry nonsense.
>
> Actually, you do.
> My ext2 partition gone in flames when there was a power outage.
> The registery can take quite a bit of garbage in it, and there are many
> tools to fix it.
Now you're skipping the issue, aren't you?.
The subject was Terminating and restarting X, not a power
outage, which doesn't give the OS time to clean up things
properly.
If you have an UPS (they're quite cheap now), power outage
is not an issue.
BTW even ctrl-alt-del in Linux will just trigger an "init 0"
sequence, i.e. first terminate all application, then kill
those who aren't responding.
Meaning that it will not fool up neither your file system,
nor any data files handled by applications that haven't
crashed.
The only problem you have is with text editors, and similar
utilities, which cannot know if you want to save or discard
the pending modifications, so they take the default path,
i.e. to discard.
Moreover, being a sensible system, even files left open by a
crashed application will be neatly closed.
Too bad Linux doesn't provide enough support for the
national languages you need.
You might help however, just by contributing to the
appropriate existing projects. You give something, you get
much more in exchange!
------------------------------
From: Stuart Fox <[EMAIL PROTECTED]>
Crossposted-To: comp.os.os2.advocacy,alt.destroy.microsoft
Subject: Re: Uptime -- where is NT?
Date: Fri, 24 Nov 2000 12:05:17 GMT
In article <[EMAIL PROTECTED]>,
"Aaron R. Kulkis" <[EMAIL PROTECTED]> wrote:
>
> But if you're running a database, and you want to know if a "commit"
> was carried-out without a crash in-between, clock-rollover every
> 7 weeks is a MAJOR pain in the ass...
>
It's actually a really stupid way to verify that. It doesn't matter
when the clock rolls over, because when it does (49.7 days or 497 days)
you get the same result.
Sent via Deja.com http://www.deja.com/
Before you buy.
------------------------------
From: Giuliano Colla <[EMAIL PROTECTED]>
Crossposted-To: comp.os.os2.advocacy,alt.destroy.microsoft
Subject: Re: Uptime -- where is NT?
Date: Fri, 24 Nov 2000 12:31:47 GMT
"Aaron R. Kulkis" wrote:
>
> "T. Max Devlin" wrote:
> >
> > Said sfcybear in alt.destroy.microsoft;
> > [...]
> > > Now, The statement I made was that the use of 479 days was based on a
> > >reasonable thought process at the time the OS was writen. It may be
> > >outdated But it was not an error or a flaw. There is NO reasonable
> > >thought that I know of that justifies a 49 day uptime clock at the time
> > >NT was writen. Memory was already getting MUCH cheaper and Computers
> > >were already running far longer than 49 days. This means that the 49 day
> > >uptime was an error or flaw.
> >
> > Sorry for droning on about this with late posts, but there is a
> > reasonable thought that justifies a 49 day uptime *COUNTER* (AS IN NOT A
> > CLOCK) at the time NT was given the ability to provide it. Which was
> > quite a bit after it was written (in the current form being
> > implemented). NT utilizes the standard SNMP mechanism, sysUptime, a
> > timeticks MIB counter. Timeticks MIB counters are in thousands of a
> > second. When querying a Unix server via SNMP, it also reports in
> > thousands of a second. Neither of these is a "clock", but a continuity
> > indicator, so it really doesn't matter. (Other than that weird bug in
> > the first version of NT which had the SNMP integrated with the registry,
> > in about 1997, IIRC, which caused the system to reboot when the counter
> > rolled over.)
> >
> > Think about it. All you need to do is query the server once every 49
> > days. And as long as it says it's been running for 49 days, the only
> > way you could be missing any uptime is if they rebooted at precisely the
> > rollover. And you'd still get a measure of the amount time it took
> > before it was up again, plus or minus the accuracy of the clocks
> > involved as an offset. The only time the number is at all "inaccurate"
> > (in the way you mean, which I would call 'inconsistent') is the very
> > first time you poll it.
>
> But if you're running a database, and you want to know if a "commit"
> was carried-out without a crash in-between, clock-rollover every
> 7 weeks is a MAJOR pain in the ass...
>
By a strictly logical point of view, Max is correct: the
information provided is sufficient.
By a practical standpoints, I stick with Aaron.
You have a 32 bit counter which rolls over before reaching
maximum count, and does it at a value which isn't related to
maximum count by a power of two, but a by power of ten.
Sort of "how can we implement it in such a way as to annoy
client software?"
------------------------------
From: "Ayende Rahien" <[EMAIL PROTECTED]>
Crossposted-To: alt.destroy.microsoft,comp.os.ms-windows.advocacy
Subject: Re: Windoze 2000 - just as shitty as ever
Date: Fri, 24 Nov 2000 14:28:32 +0200
"Tom Wilson" <[EMAIL PROTECTED]> wrote in message
news:w_rT5.14$[EMAIL PROTECTED]...
Talking about the registery.
>. I think MS could have done
> a far better job by taking the dedicated partition approach. It'd be a lot
> safer kept from the OS's file system.
You are probably correct, the problem is that it's *much* more convenient
and easier to handle files than partitions.
------------------------------
From: "Ayende Rahien" <[EMAIL PROTECTED]>
Crossposted-To: alt.destroy.microsoft,comp.os.ms-windows.advocacy
Subject: Re: Windoze 2000 - just as shitty as ever
Date: Fri, 24 Nov 2000 14:31:37 +0200
"Giuliano Colla" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> Ayende Rahien wrote:
> >
> > "Tom Wilson" <[EMAIL PROTECTED]> wrote in message
> > news:%OqT5.2513$[EMAIL PROTECTED]...
> >
> > > The point is, the entire OS doesn't go down in flames as a consequence
of
> > > the GUI crashing... Sure, you lose whatever it might have been you
were
> > > working on, but core processes other systems on your net might be
using
> > > don't have to be interrupted by the recovery. Plus, you don't have to
deal
> > > with the corrupted file system and registry nonsense.
> >
> > Actually, you do.
> > My ext2 partition gone in flames when there was a power outage.
> > The registery can take quite a bit of garbage in it, and there are many
> > tools to fix it.
>
> Now you're skipping the issue, aren't you?.
>
> The subject was Terminating and restarting X, not a power
> outage, which doesn't give the OS time to clean up things
> properly.
> If you have an UPS (they're quite cheap now), power outage
> is not an issue.
>
> BTW even ctrl-alt-del in Linux will just trigger an "init 0"
> sequence, i.e. first terminate all application, then kill
> those who aren't responding.
>
> Meaning that it will not fool up neither your file system,
> nor any data files handled by applications that haven't
> crashed.
> The only problem you have is with text editors, and similar
> utilities, which cannot know if you want to save or discard
> the pending modifications, so they take the default path,
> i.e. to discard.
> Moreover, being a sensible system, even files left open by a
> crashed application will be neatly closed.
Actually, I'd a *lot* of experiance with failed win9x systems.
I can't recall a time when the FS itself died on me.
Some files did, yes, but never the FS itself.
Even power down doesn't do it.
------------------------------
From: Giuliano Colla <[EMAIL PROTECTED]>
Crossposted-To: comp.os.os2.advocacy,alt.destroy.microsoft
Subject: Re: Uptime -- where is NT?
Date: Fri, 24 Nov 2000 12:56:06 GMT
Erik Funkenbusch wrote:
>
> "Giuliano Colla" <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]...
> > > I have used LockFile extensively, but suppose for second you're right
> that
> > > LockFile doesn't work well (it does). That doesn't change the fact you
> > > claimed NT can't do it at all.
> >
> > Maybe you didn't grasp the context. Or maybe I was too
> > concise. The context was that of an RPC implementation.
> > You have a server process, and a number of clients. Well the
> > server process doesn't start a different process for each
> > user request. It just spawns a new thread.
>
> It could very well spawn a new process, or it might spawn a new thread.
> That's something the implementer can choose to do.
>
No, he's forced to choose as a function of available API's.
But if it's performing the same transactions, on the same
database for different clients, spawning a new process is
the last thing it'd do.
> > As LockFile is
> > process related, and not thread related, it's totally
> > inexistant, as far as this sort of applications is
> > concerned. You must resort to a specific device which turns
> > out to be a File locking mechanism.
>
> What are you talking about? It's painfully easy to write multi-threaded
> safe locking within a process. Simply maintain a map of file position
> ranges and thread ID's. If the file position range is in the list, it's
> locked and don't use it until it is, fire an event when you unlock the
> range. Big deal.
>
Why should an application implement a typical OS mechanism?
Isn't OS intended to provide services which are common
requirement for all the applications? I heard that it's both
a matter of efficiency and reliability.
> > I thought I was thick,
> > but DB designers have confirmed to me from different sources
> > that for multiple user database access the FileLock API is
> > totally useless.
>
> That depends entirely on the design of your server. Use the right API for
> the right job. FileLocking is designed to arbitrate file access between
> processes because processes don't know about each other in most cases.
> Threads should very well know about each other and can arbitrate access
> themselves.
>
If the OS doesn't provide the appropriate mechanism, they're
forced to. With an approach which is much less efficient and
much more error prone than if implemented by a well designed
OS.
> > It's an API designed for classical point-and-click users.
>
> Bullshit. It has nothing to do with "classical point-and-click users".
> It's an API designed for blind arbitration. Threads are not blind.
>
Blind arbitration is all that's required for
point-and-click. Database applications require more than
that.
> > If you've used it for such applications, then you've been
> > happy, of course. If you'd tried to use it for the sort of
> > application I was mentioning, then you'd have been utterly
> > disappointed.
>
> And which API does something similar under any other OS and works with
> threads?
Well, Unix processes are what most nearly resembles to MS
threads (they can know about each other, they are spawned
one from the other, they may share the same code and just
have different data, they've special IPC mechanism, etc.).
Unix provides process related file/record lock, which
translates into thread related file/record lock in MS terms.
Intel's RMX OS, (a real-time multitasking, used for
industrial application) provides Task level locking, with
tasks being the closest thing to MS threads, Jobs being the
closest thing to MS process.
Those are the last two I've been using, besides Windows, so
I clearly remember. Of other systems I should dig into
manuals.
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Linux for nitwits
Reply-To: [EMAIL PROTECTED]
Date: Fri, 24 Nov 2000 13:38:09 GMT
I misread the thread.
claire
On Fri, 24 Nov 2000 09:47:35 GMT, "Tom Wilson" <[EMAIL PROTECTED]>
wrote:
>You mean, YOUR cable guys will actually come to your house, link your
>systems via the parallel ports and set up a box for you?!?
>
>
>Wow!
>
>
>I mean, OUR cable guys can't even crimp cable ends properly!
Neither can mine.
>
>I'm both impressed and envious!
>
>
>(Yeeesh!)
>
>
>Tom Wilson
>
>
>
>>
>>
>> claire
>>
>>
>> On Tue, 21 Nov 2000 18:37:49 -0500, "Clifford W. Racz"
>> <[EMAIL PROTECTED]> wrote:
>>
>> >> First, I too am a Linux nitwit, but unfortunately
>> >> your solution to install Linux on your laptop is
>> >> already supported by Windows 9x. "Direct Cable
>> >> Connection" uses the parallel port to directly
>> >> connect computers together in Windows, thereby
>> >> allowing the same procedure to work in Windows as
>> >> it did in Linux.
>> >
>> >A "direct cable connection" is the biggest PITA I have ever dealt with
>wrt
>> >Windows. I have yet to get 1 to work. I would rather use Linux than
>rely
>> >on that.
>> >
>> >
>> >
>>
>
------------------------------
From: Curtis <[EMAIL PROTECTED]>
Crossposted-To: alt.destroy.microsoft,comp.os.ms-windows.advocacy
Subject: Re: Windoze 2000 - just as shitty as ever
Date: Fri, 24 Nov 2000 08:50:07 -0500
T. Max Devlin wrote...
> >Press shift when you click the OK button on the shut down screen, this would
> >give you quick shutdown.
>
> Wow, that's a new one. Is it just 98SE? And of course it begs the
> question, "why isn't the shutdown always quick?"
The longer shutdown method takes the time to save applications as well as
explorers recent configuration changes.
--
| ,__o
!ACM _-\_<, A thing is not necessarily true because
<(*)>--(*)/'(*)______________________ a man dies for it.
mailto:martian*at*cwjamaica*dot*com
------------------------------
From: Curtis <[EMAIL PROTECTED]>
Crossposted-To: alt.destroy.microsoft,comp.os.ms-windows.advocacy
Subject: Re: Windoze 2000 - just as shitty as ever
Date: Fri, 24 Nov 2000 08:54:50 -0500
Ayende Rahien wrote...
>
> "T. Max Devlin" <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]...
> > Said Ayende Rahien in alt.destroy.microsoft on Thu, 23 Nov 2000 03:44:12
>
>
> > >Why *should* I care about language that I'm neither using nor likely to
> use.
> >
> > Because you want others to care about languages that they neither use
> > nor are likely to use, but that you use. Get it? Its called being
> > civilized.
>
> I don't want them to care, *I* care.
> Linux just can't supply my needs, period.
> That was the point that started this arguement.
>
>
> > >Linux has *bad* support for the launguages that *I* need. I don't give a
> > >horse's ass for those that I don't need
> > >Guess who has the best support for those languages that I *do* need?
> >
> > Monopoly crapware, of course. Windoze 2000 - just as shitty as ever,
> > but you're too dumb to know it.
>
> You know, it's exactly this kind of statements that make you look like an
> idiot.
An ignorant, pontificating poo he is. :=) He has the same blinkered
approach as the ignorant Windows users out there. His blinkers are of a
different type and therefore fit differently. It's a pity though. I'd
have expected him to have a better/more pragmatic perspective on things.
--
| ,__o
!ACM _-\_<, A thing is not necessarily true because
<(*)>--(*)/'(*)______________________ a man dies for it.
mailto:martian*at*cwjamaica*dot*com
------------------------------
From: Curtis <[EMAIL PROTECTED]>
Crossposted-To: alt.destroy.microsoft,comp.os.ms-windows.advocacy
Subject: Re: Windoze 2000 - just as shitty as ever
Date: Fri, 24 Nov 2000 09:04:17 -0500
Ayende Rahien wrote...
> > And so now suddenly you're suggesting that others should give a shit
> > about languages they don't use, even though you just said previously
> > that you see no reason to do that yourself.
>
> They shouldn't, they give shit about what *they* are using.
> Got that?
That's the practical approach. Why should I care that Windows supports
Japanese? If it does, good for the Japanese, but I don't really care.
It's for MS to care since they wish to market their OS to the Japanese.
The same goes for the OS features. I only care about the features I wish
to use and how they're implemented. I couldn't care less about the
features a sysadmin would need. That's for MS and the sysadmins to care
about.
etc. etc. There's nothing uncivilized about that approach. If I offer a
service, then I care about the target customers and their needs or
preferences. If I'm the customer, then it's a different perspective.
> > Of *course* Microsoft supports as many languages as they possibly can.
> > When you're trying to ensure that nobody who uses a computer can avoid
> > using your product, its worth putting quite a bit of (non-efficient,
> > from a competitive free market production standpoint) investment in
> > removing any excuses they may have for avoiding it.
>
> Nonesense.
> MS translate much of its products to many other languages, because there is
> *profit* in it.
> Most people in the world *don't* know english, therefor, they need an OS in
> their own language, and they'll pay for it.
>
> Take a look at what happened when Iceland wanted windows in their own
> language, btw.
I think you missed what T. Max was really saying but that's OK. It's not
worth it. :=)
--
| ,__o
!ACM _-\_<, A thing is not necessarily true because
<(*)>--(*)/'(*)______________________ a man dies for it.
mailto:martian*at*cwjamaica*dot*com
------------------------------
** 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 (and comp.os.linux.advocacy) via:
Internet: [EMAIL PROTECTED]
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
******************************
