Linux-Misc Digest #816, Volume #24               Wed, 14 Jun 00 18:13:03 EDT

Contents:
  Re: mounting VFAT for all users..HELP! ("David ..")
  Re: Detaching users from /etc/passwd (Floyd Davidson)
  linux
  Re: Read Linux partition from Win95 ? (Rootman)
  Re: Video editing w/ linux ([EMAIL PROTECTED])
  Re: compaq dl380 & rh62 scsi woes ("David ..")
  Mandrake 7.1 ISO too big! (Broccoli Puffs)
  Re: linux (Henjo)
  Re: linux (Robert J Carter)
  Boot sound (&-ed progs in rc.local) (Tijmen Stam)
  kernel Makefile bugs? (Jeff Lacki)
  Re: Detaching users from /etc/passwd ("Andrew N. McGuire ")
  Give me a general advice on X please. (Tijmen Stam)
  Re: environment variable ("Andrew N. McGuire ")
  Re: linux ("David ..")
  Re: Detaching users from /etc/passwd (Dirk Freese)
  Day planner for linux? (kj0)
  ansiTime library: mktime() inconsistencies across platforms? (Mark Law)
  ansiTime library: mktime() inconsistencies across platforms? (Mark Law)
  Re: Opening files (Jonathan M Hill)
  Re: Time sync problem in Cron (Toni)
  Re: Give me a general advice on X please. ("David ..")
  Re: Day planner for linux? ("David ..")
  Re: kernel Makefile bugs? ("David ..")

----------------------------------------------------------------------------

From: "David .." <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux,comp.os.linux.redhat
Subject: Re: mounting VFAT for all users..HELP!
Date: Wed, 14 Jun 2000 14:05:22 -0500

"Kent A. Signorini" wrote:
> 
> I want to mount a VFAT drive (my win98 drive) at /mnt/win98 for all users to
> have write access. 

Since permissions do not work on vfat it is possible for a user to
destroy your windoz partition.  You have been warned!

This line should give all users access to the /mnt/win98 partition.

/dev/hda1       /mnt/win98      vfat
      user,owner,exec,dev,suid,rw,noauto 0 0

-- 
Registered with the Linux Counter.  http://counter.li.org
ID # 123538

------------------------------

From: Floyd Davidson <[EMAIL PROTECTED]>
Subject: Re: Detaching users from /etc/passwd
Date: 14 Jun 2000 10:22:30 -0800

Vilmos Soti <[EMAIL PROTECTED]> wrote:
>
>I think his problem is that each time someone logs in to check
>email (which any users do every few minutes) then the parsing
>factor of /etc/passwd becomes cumbersome. Imagine if ten people
>want to check her/his email in the same time on a machine with
>1000 users. Then the 1000 lines long passwd file would be
>parsed ten times simultaneously.  One solution is to use
>something like NIS. In that case the user info is in a database
>which scales better while /etc/passwd not.

That doesn't remove the need to parse the /etc/passwd file though.

Note that on any reasonable system, Linux included, the whole thing
will basically be kept in memory by disk caching.  It just really
isn't an intensive task.

>Once I read an article where the admins had *EXACTLY* the same
>problem.  They were using FreeBSD. Originally the system was
>fast, but as more users were added, the time to parse the
>passwd file became longer and it happened more often. Their
>solution was to change to YP. They were talking about thousands
>of users though.

I've admin'ed a small ISP using a BSDI machine with 5-6000 users
where several hundreds would normally be connected at any given
time.  Not only was POP doing authentication, but we were
serving RADIUS authentication to the terminals servers too.
That was being done on a 200 Mhz K6 cpu if I remember right.  It
wasn't taxed at all.  A modern SMP system with PIII's wouldn't
even know it was running.

However, that does bring up the point that if there is a problem
with cpu load due to other processes running, it would be a very
simple and inexpensive solution to acquire an older cpu and set
up a networked authentication server which does nothing else but
authentication.  Clearly a PII 350, for example, would be able
to handle a rather large user base.

  Floyd

-- 
Floyd L. Davidson                          [EMAIL PROTECTED]
Ukpeagvik (Barrow, Alaska)

------------------------------

From: <[EMAIL PROTECTED]>
Subject: linux
Date: Wed, 14 Jun 2000 19:30:23 GMT

how do you upgrade linux to win95

--
Posted via CNET Help.com
http://www.help.com/

------------------------------

From: Rootman <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux,comp.os.linux.setup,linux.redhat.misc
Subject: Re: Read Linux partition from Win95 ?
Date: Wed, 14 Jun 2000 19:32:07 GMT

So is any single disk Linux distro too.  Once physical access of the
machine is had then anything is up for grabs.

Note that this utility cannot look at notworked Linux paritions just
LOCAL partitions.

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

  "Andrew E. Schulman" <[EMAIL PROTECTED]> wrote:
> > Check out Explore2fs  Windows 95 file browser like interface for
Linux
> > parititions , write enable is "experimental" but seems to work
alright
> > (cross my fingers).
>
> Sounds to me like a massive security hole.
>


Sent via Deja.com http://www.deja.com/
Before you buy.

------------------------------

From: [EMAIL PROTECTED]
Subject: Re: Video editing w/ linux
Date: 14 Jun 2000 12:55:50 -0700

Yan Seiner <[EMAIL PROTECTED]> writes:

> My parents bought a new digital cam corder.  They are now looking into
> getting some software for digital editing.
> 
> They are now using a turkey computer I put together a few years ago out
> of left over parts - a P5/133, 16 MB ram, win95.  This won't cut it for
> video editing.
> 
> I am looking for suggestions for software that runs under linux that
> allows video editing.  The hardware will probably be something like a
> 20~30 GB scsi HD, 128 MB RAM, pIII/500 (or maybe AMD?) and a good video
> card that's supported by XFree 4.0.
> 
> I am actually looking for suggestions for hardware as well that fits the
> above.  I'm not really familiar with new video cards, though Matox has
> long been my favorite.
> 
> --Yan
> 
> -- 
> 
> Think different
>       ride a recumbent
>               use Linux.
You might check the following links:
http://heroine.linuxave.net/
http://penguin.lvcm.com/

The latter is more to do with video capture.  The first link is for for
Broadcast2000.  They say try the binary to see if you want it then get
the source and compile it.  I skipped the binary for various reasons and
struggled to compile the source.  The last 5 % is tough with me having
to go around making a lot of symbolic links and so on on my slackware
system and I finally abandoned it for awhile to do something else.
        The bottom line is, last, time I looked, there wasn't much in the
way of video editing SW for linux.  But that may change.

   ---- Remove "UhUh" and "Spam" to get my real email address----

------------------------------

From: "David .." <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware.
Subject: Re: compaq dl380 & rh62 scsi woes
Date: Wed, 14 Jun 2000 14:57:07 -0500

[EMAIL PROTECTED] wrote:
> 
> Compaq - Linux folks,
> 
> Has anybody had any luck using non-disk array scsi devices such as dat
> tape drives with a compaq dl380 and redhat 6.2? 


If it is an SCSI tape drive you might check arkeia they have a free
download that I use and find it works great.

http://www.arkeia.com/downloadfree.html

-- 
Registered with the Linux Counter.  http://counter.li.org
ID # 123538

------------------------------

From: Broccoli Puffs <[EMAIL PROTECTED]>
Subject: Mandrake 7.1 ISO too big!
Date: Wed, 14 Jun 2000 20:26:05 GMT

Damn,

Tell me, why can't Mandrake keep the size of their iso down to the 
"regular" limit of about 650M?

The newest one is already two isos anyhow, why not make the 667.4M 
install iso just a little smaller??!!

grrrr-

james
[EMAIL PROTECTED]

------------------------------

From: Henjo <[EMAIL PROTECTED]>
Subject: Re: linux
Date: 14 Jun 2000 20:27:31 GMT

[EMAIL PROTECTED] wrote:
> how do you upgrade linux to win95

> --
> Posted via CNET Help.com
> http://www.help.com/

:):)

   Henjo
 /-----------------------------------\
|  http://glas.its.tudelft.nl/~henjo  |
|  http://www.sweetdesign.nl          |
|  ICQ#  18616700                     |
 \-----------------------------------/

------------------------------

From: [EMAIL PROTECTED] (Robert J Carter)
Subject: Re: linux
Date: Wed, 14 Jun 2000 20:33:14 GMT

In article <8i8prj$s3o$[EMAIL PROTECTED]>,
        Henjo <[EMAIL PROTECTED]> writes:
> [EMAIL PROTECTED] wrote:
>> how do you upgrade linux to win95
> 
You don't. It's a downgrade :-)

-- 
Robert J Carter at Oghma dot on dot ca
Use My initials to reach me via e-mail

------------------------------

From: Tijmen Stam <[EMAIL PROTECTED]>
Subject: Boot sound (&-ed progs in rc.local)
Date: Wed, 14 Jun 2000 22:40:28 +0200

This is a multi-part message in MIME format.
==============4B6B389AB87CC3E57B665D68
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit

<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
Hi.
<p>I have a very cool mp3 and I want to set it as boot sound while logging
in on my non-X machine. So in /etc/rc.d/rc.local, after esd startup, I
place:
<p>mpg123 /mp3/demo/Jetta.mp3
<p>Then I hear the sound ok, but the local dosn't go on and i can't login
until Jetta has finished. So instead of the previous, I do a
<p>mpg123 /mp3/demo/Jetta.mp3 &amp;
<p>But I hear no sound at all. When logged in (and Jetta can't be finished
yet), a ps -aux |&nbsp;grep mpg gives me no hits (exept for the grep mpg).
<br>When i'm logged in, I run rc.local, I get a load of "prog alredy loaded"
messages, but the sound does play... So why not when I'm not logged in
yet?
<p>I use redhat 6.0
<p>By the way: what I&nbsp;mean is to start playing the sound before the
getty, so It'll finish after i'm logged in (if you didn't get that)
<p>Tijmen
<p>plaese Mailto [EMAIL PROTECTED]
<pre>--&nbsp;
>From Tijmen Stam - "I believe in Linux" - [EMAIL PROTECTED]
counter.li.org reg#178552-54654, Machine#78930 &amp; #78931
---
I'm a signature virus.&nbsp;
Please help me spread and set me as your signature. ;-)</pre>
&nbsp;</html>

==============4B6B389AB87CC3E57B665D68
Content-Type: text/x-vcard; charset=us-ascii;
 name="tijmen.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Tijmen Stam
Content-Disposition: attachment;
 filename="tijmen.vcf"

begin:vcard 
n:Stam;Tijmen
tel;fax:++31 (0)50 5425400
tel;home:++31 (0)50 5415448
x-mozilla-html:TRUE
adr:;;Middelberterweg 48;Groningen;Groningen;9723 EW;The Netherlands
version:2.1
email;internet:[EMAIL PROTECTED]
x-mozilla-cpt:;30624
fn:Tijmen Stam
end:vcard

==============4B6B389AB87CC3E57B665D68==


------------------------------

From: Jeff Lacki <[EMAIL PROTECTED]>
Subject: kernel Makefile bugs?
Date: 14 Jun 2000 20:36:05 GMT

Ok, Im new enough to Linux to be dangerous, but smart enough
to realize something is going on ;)  

Installed a Sound Blaster 512 Card 2 weeks ago just fine
(RedHat 6.2).  Then, I bought a NetGear Ethernet card and found
an article on how to get it to work, so I read the article, followed
the instructions on make xconfig (ONLY setting the appropriate
ether module etc) in this exact order (via the instructions):

make clean
make xconfig
make dep
make clean
make bzImage
make modules modules_install bzlilo
shutdown -r 0

NetGear card now works, Sound Blaster Module emu10k1.o is now
toast.  I get a lot of undefined symbols.

I saw this problem in the past as well.  It appears that the
Makefile or some derivation of it all is not compiling or loading
in the same stuff consistantly after a make clean is done.

Anyone know how I can correct this?  Is Kernel 2.2-16 the answer?

Thanks




------------------------------

From: "Andrew N. McGuire " <[EMAIL PROTECTED]>
Subject: Re: Detaching users from /etc/passwd
Date: Wed, 14 Jun 2000 15:43:57 -0500

On 14 Jun 2000, Floyd Davidson wrote:

+ Vilmos Soti <[EMAIL PROTECTED]> wrote:
+ >
+ >I think his problem is that each time someone logs in to check
+ >email (which any users do every few minutes) then the parsing
+ >factor of /etc/passwd becomes cumbersome. Imagine if ten people
+ >want to check her/his email in the same time on a machine with
+ >1000 users. Then the 1000 lines long passwd file would be
+ >parsed ten times simultaneously.  One solution is to use
+ >something like NIS. In that case the user info is in a database
+ >which scales better while /etc/passwd not.
+ 
+ That doesn't remove the need to parse the /etc/passwd file though.
+ 
+ Note that on any reasonable system, Linux included, the whole thing
+ will basically be kept in memory by disk caching.  It just really
+ isn't an intensive task.
+ 
+ >Once I read an article where the admins had *EXACTLY* the same
+ >problem.  They were using FreeBSD. Originally the system was
+ >fast, but as more users were added, the time to parse the
+ >passwd file became longer and it happened more often. Their
+ >solution was to change to YP. They were talking about thousands
+ >of users though.
+ 
+ I've admin'ed a small ISP using a BSDI machine with 5-6000 users
+ where several hundreds would normally be connected at any given
+ time.  Not only was POP doing authentication, but we were
+ serving RADIUS authentication to the terminals servers too.
+ That was being done on a 200 Mhz K6 cpu if I remember right.  It
+ wasn't taxed at all.  A modern SMP system with PIII's wouldn't
+ even know it was running.

I have admin'ed a larger ISP, 30,000 users or so, we used /etc/passwd
with no problem.  We were using SPARC Ultra 2's, and had RADIUS on
a different machine.

[ snip ]

Regards,

anm
-- 
/*-------------------------------------------------------.
| Andrew N. McGuire                                      |
| [EMAIL PROTECTED]                              |
`-------------------------------------------------------*/


------------------------------

From: Tijmen Stam <[EMAIL PROTECTED]>
Subject: Give me a general advice on X please.
Date: Wed, 14 Jun 2000 22:53:53 +0200

This is a multi-part message in MIME format.
==============2D1399B9E302D3C94BDCFA3D
Content-Type: text/html; charset=us-ascii
Content-Transfer-Encoding: 7bit

<!doctype html public "-//w3c//dtd html 4.0 transitional//en">
<html>
Hi.
<p>I have a pc I use sometimes in console, mostly in X. But I'm growing
more and more tired of X. It gives random crashes nobody can resolve (vertical
bars on the screen, mostly (but not always) when scrolling with the mouse)
and netscape (my main X reason) crashes quite often too. (about once per
mail I reply, that's often!)
<br>I have a 2.2.5-15 kernel, Redhat 6.0, Gnome, Enligtenement, netscape
4.51, a S3 trio 3d/2x video card, svga server, 1024x7?? display and 64M
of ram (wich is enough) and a AMD k6-2-450.
<p>Reinstalling Redhat doesn't resolve these problems.
<p>What do you advice me: downloading X 4, or&nbsp; X3.5 and installing
it myself, or maybe netscape 6 (which is said to be unstable too)???, another
WM?
<br>What should you all do in my place?
<p>Tijmen
<p>mailto [EMAIL PROTECTED]
<br>&nbsp;
<pre>--&nbsp;
>From Tijmen Stam - "I believe in Linux" - [EMAIL PROTECTED]
counter.li.org reg#178552-54654, Machine#78930 &amp; #78931
---
I'm a signature virus.&nbsp;
Please help me spread and set me as your signature. ;-)</pre>
&nbsp;</html>

==============2D1399B9E302D3C94BDCFA3D
Content-Type: text/x-vcard; charset=us-ascii;
 name="tijmen.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Tijmen Stam
Content-Disposition: attachment;
 filename="tijmen.vcf"

begin:vcard 
n:Stam;Tijmen
tel;fax:++31 (0)50 5425400
tel;home:++31 (0)50 5415448
x-mozilla-html:TRUE
adr:;;Middelberterweg 48;Groningen;Groningen;9723 EW;The Netherlands
version:2.1
email;internet:[EMAIL PROTECTED]
x-mozilla-cpt:;30624
fn:Tijmen Stam
end:vcard

==============2D1399B9E302D3C94BDCFA3D==


------------------------------

From: "Andrew N. McGuire " <[EMAIL PROTECTED]>
Subject: Re: environment variable
Date: Wed, 14 Jun 2000 15:50:35 -0500

On 14 Jun 2000, Dances With Crows wrote:

+ On Wed, 14 Jun 2000 09:30:23 -0500, Mark Guzzo 
+ <<[EMAIL PROTECTED]>> shouted forth into the ether:
+ >What is an environment variable, how do I find out what they are set to
+ >and how would I change them ?
+ 
+ An environment variable is a variable that your shell knows about.  

Not necessarily, as *nix is multi-user.

+ Examples include EDITOR, PS1, HOME, PATH, and a host of others.  You find
+ out what they are set to by entering "printenv" or "set", and you change
+ them by doing:
+ 
+ export VARIABLE=value      (sh, bash)
+ set VARIABLE value         (tcsh, ksh)

that is the wrong syntax for set...

set var=value

or 

set var = value

set is *primarily* used for setting shell variables in tcsh.
although setting certain shell variables set the corresponding
environmental variables..

set path = ( /bin /usr/bin ... )

also sets $PATH.

If you are concerned with setting enviromental variables in tcsh
look into setenv.

Best Wishes,

anm
-- 
/*-------------------------------------------------------.
| Andrew N. McGuire                                      |
| [EMAIL PROTECTED]                              |
`-------------------------------------------------------*/


------------------------------

From: "David .." <[EMAIL PROTECTED]>
Subject: Re: linux
Date: Wed, 14 Jun 2000 15:45:01 -0500

[EMAIL PROTECTED] wrote:
> 
> how do you upgrade linux to win95
> 
> --
> Posted via CNET Help.com
> http://www.help.com/


Figure a way to make it crash a couple of times a day.
Why, So you know when it's time for bed?

-- 
Registered with the Linux Counter.  http://counter.li.org
ID # 123538

------------------------------

From: Dirk Freese <[EMAIL PROTECTED]>
Subject: Re: Detaching users from /etc/passwd
Date: Wed, 14 Jun 2000 23:01:25 +0200

Hi again,

first of all: Thanks for your numerous repsonses.

I think we can move the topic slightly meanwhile: I'd like to authenticate
my POP/FTP users using a DBMS (preferably mySQL).
So if anyone knows about this, please let me know.

Thanks,

Dirk


------------------------------

From: kj0 <[EMAIL PROTECTED]>
Subject: Day planner for linux?
Date: 14 Jun 2000 17:06:54 -0400





I very much need software that will alert me (by e-mail?) of upcoming
appointments.  A web-based reminder system is out of the question.  I
need something I can run on my linux workstation.

Can anyone recommend a public domain day planner/appointment
book/calendar for linux?

THANKS!

kj


------------------------------

From: Mark Law <[EMAIL PROTECTED]>
Crossposted-To: comp.lang.c,comp.os.linux.development.system,comp.os.linux.portable
Subject: ansiTime library: mktime() inconsistencies across platforms?
Date: Wed, 14 Jun 2000 21:14:19 GMT


#### If replying, please cc "[EMAIL PROTECTED]" ### 

I'm a software developer whose trying to resolve some issues with
ansiTime library implementations.  I've found differences between test
results produced on SunOS 5.7 (using SUN's libs), Linux (using GNU
2.95.2 libs), and Microsoft Win NT 4.0 (using MS Visual C++ 5.0, Dev.
Studio 97).

I'm running some test code that calls mktime() with varying inputs,
where the inputs are predominantly around daylight savings start and end
transitions; ie. the hour before DST starts, the hour DST starts, the
hour after DST started, etc.  On all but Win NT 4.0, I'm setting the TZ
environment variable with the DST start and end times so my test results
should be consistent across systems.  I'm using strftime() to format
then print the results.

The SunOS and Win NT results are very similar.  I also believe they
are the most accurate.  The linux results differ widely from the SunOS &
Win NT results.  

My first question is more general.  Where is a complete description of
how mktime() is supposed to behave?  I've read man pages, etc.  There's
much documentation that describes mktime(), some of it, even clearly
defines the meaning of the DST flag input parameter.  Some of the
mktime() documentation states that it can change members of the input tm
structure.  However, what I'm noticing is that different ansiTime lib
implementations on different systems produce different results given the
same inputs.  Is there some definitive description that can determine
which implementation is correct?

My second question is about 1 particular difference of the many that
I've discovered.
Between the SunOS and Win NT systems, the only difference between the
results is the following:

Given the following:

   TZ=PST8PDT7,92,302   /* SunOS only.  I don't know if it's possible
                         * to set DST start, end in Win NT 4.0 */

   time_t time;
   struct tm timeptr;

   timeptr.tm_sec   = 0;     /* seconds after the minute [0, 61]  */
   timeptr.tm_min   = 0;     /* minutes after the hour [0, 59] */
   timeptr.tm_hour  = 1;    /* hour since midnight [0, 23] */
   timeptr.tm_mday  = 29;    /* day of the month [1, 31] */
   timeptr.tm_mon   = 9;     /* months since January [0, 11] */
   timeptr.tm_year  = 100;    /* years since 1900 */
   /* int  tm_wday;  NOT SET */   /* days since Sunday [0, 6] */
   /* int  tm_yday;  NOT SET */   /* days since January 1 [0, 365] */
   timeptr.tm_isdst = -1;   /* flag for daylight savings time */

   time = mktime(&timeptr);

SunOS produces the following results:

   After mktime: day of week: 0  day of year: 302  DST: 1
   The time: 972806400  0x39fbd900  is 0 10 29 01:00:00 2000 AM 01 PDT
   datesTest: expected: 972810000 got: 972806400

SunOS determined this input time was in daylight savings time.

Win NT 4.0 produces the following results:

   After mktime: day of week: 0  day of year: 302  DST: 0
   The time: 972810000  0x39fbe710  is 0 10 29 01:00:00 2000 AM 01 PST

Win NT 4.0 determined this input time was in standard time.

Note: strftime is used with the format: "%w %m %d %I:%M:%S %Y %p %H %Z"

There are other test result diffs that follow the same pattern, where
the pattern is for times given on the boundary of transitioning from
daylight savings to standard time, SunOS treats tm_isdst = -1 as if its
a "1" and the resulting time is a daylight savings time and Win NT
treats tm_isdst = -1 as if its a "0" and the resulting time is NOT a
daylight savings time.

Which behavior is correct?

mlaw

-- 
Mark A. Law                  Phone:  (650) 701-4096
Liberate Technologies        Fax:    (650) 701-4096
2 Circle Star Way            Home:   (408) 255-3844
San Carlos, CA  94070-6200   Email:  [EMAIL PROTECTED]

------------------------------

From: Mark Law <[EMAIL PROTECTED]>
Crossposted-To: comp.lang.c,comp.os.linux.development.system,comp.os.linux.portable
Subject: ansiTime library: mktime() inconsistencies across platforms?
Date: Wed, 14 Jun 2000 21:04:54 GMT


#### If replying, please cc "[EMAIL PROTECTED]" ### 

I'm a software developer whose trying to resolve some issues with
ansiTime library implementations.  I've found differences between test
results produced on SunOS 5.7 (using SUN's libs), Linux (using GNU
2.95.2 libs), and Microsoft Win NT 4.0 (using MS Visual C++ 5.0, Dev.
Studio 97).

I'm running some test code that calls mktime() with varying inputs,
where the inputs are predominantly around daylight savings start and end
transitions; ie. the hour before DST starts, the hour DST starts, the
hour after DST started, etc.  On all but Win NT 4.0, I'm setting the TZ
environment variable with the DST start and end times so my test results
should be consistent across systems.  I'm using strftime() to format
then print the results.

The SunOS and Win NT results are very similar.  I also believe they
are the most accurate.  The linux results differ widely from the SunOS &
Win NT results.  

My first question is more general.  Where is a complete description of
how mktime() is supposed to behave?  I've read man pages, etc.  There's
much documentation that describes mktime(), some of it, even clearly
defines the meaning of the DST flag input parameter.  Some of the
mktime() documentation states that it can change members of the input tm
structure.  However, what I'm noticing is that different ansiTime lib
implementations on different systems produce different results given the
same inputs.  Is there some definitive description that can determine
which implementation is correct?

My second question is about 1 particular difference of the many that
I've discovered.
Between the SunOS and Win NT systems, the only difference between the
results is the following:

Given the following:

   TZ=PST8PDT7,92,302   /* SunOS only.  I don't know if it's possible
                         * to set DST start, end in Win NT 4.0 */

   time_t time;
   struct tm timeptr;

   timeptr.tm_sec   = 0;     /* seconds after the minute [0, 61]  */
   timeptr.tm_min   = 0;     /* minutes after the hour [0, 59] */
   timeptr.tm_hour  = 1;    /* hour since midnight [0, 23] */
   timeptr.tm_mday  = 29;    /* day of the month [1, 31] */
   timeptr.tm_mon   = 9;     /* months since January [0, 11] */
   timeptr.tm_year  = 100;    /* years since 1900 */
   /* int  tm_wday;  NOT SET */   /* days since Sunday [0, 6] */
   /* int  tm_yday;  NOT SET */   /* days since January 1 [0, 365] */
   timeptr.tm_isdst = -1;   /* flag for daylight savings time */

   time = mktime(&timeptr);

SunOS produces the following results:

   After mktime: day of week: 0  day of year: 302  DST: 1
   The time: 972806400  0x39fbd900  is 0 10 29 01:00:00 2000 AM 01 PDT
   datesTest: expected: 972810000 got: 972806400

SunOS determined this input time was in daylight savings time.

Win NT 4.0 produces the following results:

   After mktime: day of week: 0  day of year: 302  DST: 0
   The time: 972810000  0x39fbe710  is 0 10 29 01:00:00 2000 AM 01 PST

Win NT 4.0 determined this input time was in standard time.

Note: strftime is used with the format: "%w %m %d %I:%M:%S %Y %p %H %Z"

There are other test result diffs that follow the same pattern, where
the pattern is for times given on the boundary of transitioning from
daylight savings to standard time, SunOS treats tm_isdst = -1 as if its
a "1" and the resulting time is a daylight savings time and Win NT
treats tm_isdst = -1 as if its a "0" and the resulting time is NOT a
daylight savings time.

Which behavior is correct?

mlaw

-- 
Mark A. Law                  Phone:  (650) 701-4096
Liberate Technologies        Fax:    (650) 701-4096
2 Circle Star Way            Home:   (408) 255-3844
San Carlos, CA  94070-6200   Email:  [EMAIL PROTECTED]

------------------------------

From: [EMAIL PROTECTED] (Jonathan M Hill)
Subject: Re: Opening files
Date: 14 Jun 2000 21:03:32 GMT

Dances With Crows ([EMAIL PROTECTED]) wrote:
: Since you seem very new to Unix in general, I reccommend reading a book
: like _Running Linux_, which has a lot of useful information and tutorials
: for new users.

Hello;

     I agree, here are some topics to look up.

o mounting a drive - since CD-ROM is a removable media, it should not be
  automatically mounted when your system boots.  It is important to know
  how to mount and unmount a drive.

o drive access - to avoid having to become super-user every time, you can
  specify which users have permission to mount a drive.  The linuxconf
  program and other similar programs provide a simple way to configure
  your system.

o file permissions - once the drive is mounted, permissions provide a
  means to control access to the files.

      Good luck, happy reading;
                                             Jonathan



------------------------------

From: Toni <[EMAIL PROTECTED]>
Subject: Re: Time sync problem in Cron
Date: Wed, 14 Jun 2000 23:26:15 +0200


> What?  (Which version of the NTP daemon did you try?)  That doesn't happen
> for me.

Hmm, I can't determine what version xntpd is, but xntpdc is 4.0.99f

Greetz,
Toni.

------------------------------

From: "David .." <[EMAIL PROTECTED]>
Subject: Re: Give me a general advice on X please.
Date: Wed, 14 Jun 2000 16:21:44 -0500

Tijmen Stam wrote:
> 
> Hi.
> 
> I have a pc I use sometimes in console, mostly in X. But I'm growing
> more and more tired of X. It gives random crashes nobody can resolve
> (vertical bars on the screen, mostly (but not always) when scrolling
> with the mouse) and netscape (my main X reason) crashes quite often
> too. (about once per mail I reply, that's often!)
> I have a 2.2.5-15 kernel, Redhat 6.0, Gnome, Enligtenement, netscape
> 4.51, a S3 trio 3d/2x video card, svga server, 1024x7?? display and
> 64M of ram (wich is enough) and a AMD k6-2-450.
> 
> Reinstalling Redhat doesn't resolve these problems.
> 
> What do you advice me: downloading X 4, or  X3.5 and installing it
> myself, or maybe netscape 6 (which is said to be unstable too)???,
> another WM?
> What should you all do in my place?
> 

Netscape 4.51 has bugs. Upgrade to 4.73 

Upgrade the kernel to 2.2.16 a bug has been found in previous kernels.
http://www.lwn.net/2000/0608/a/2.2.16-security.php3

Netscape 6 is still a preview version and not fully functional.

As for X it could be just a wrong configuration for your monitor or
video card in /etc/X11/XF86Config 

You might want to consider upgrading to RH-6.1 or 6.2

-- 
Registered with the Linux Counter.  http://counter.li.org
ID # 123538

------------------------------

From: "David .." <[EMAIL PROTECTED]>
Subject: Re: Day planner for linux?
Date: Wed, 14 Jun 2000 16:29:08 -0500

kj0 wrote:
> 
[snip]
> Can anyone recommend a public domain day planner/appointment
> book/calendar for linux?

Don't know if this is what you are looking for but here is a link.

http://www.bitrot.de/plan.html

-- 
Registered with the Linux Counter.  http://counter.li.org
ID # 123538

------------------------------

From: "David .." <[EMAIL PROTECTED]>
Subject: Re: kernel Makefile bugs?
Date: Wed, 14 Jun 2000 16:32:38 -0500

Jeff Lacki wrote:
> 
[snip]
> Anyone know how I can correct this?  Is Kernel 2.2-16 the answer?

The 2.2.16 kernel fixes some security issues.

http://www.lwn.net/2000/0608/a/2.2.16-security.php3

-- 
Registered with the Linux Counter.  http://counter.li.org
ID # 123538

------------------------------


** 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.misc) 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-Misc Digest
******************************

Reply via email to