Re: cygwin + windows update = lock up (W2K SP4)

Re: cygwin + windows update = lock up (W2K SP4)

[John Hackett]

I've just be told that a client is having this problem - they have a  
webserver running on win2k/sp4 with secure access via OpenSSH/Cygwin  
and the auto update is failing.

I've joined this list specifically to track this issue: is there a  
solution yet?

Any info would be very helpful. Thanks.

Regards,

John Hackett
Icon Information Systems Ltd.
t: 020 8764 2663
f: 08700 525759
m: 07801 231118
e: john.hackett@icon-is.co.uk
w: www.icon-is.co.uk

Re: cygwin + windows update = lock up (W2K SP4)

[Danilo Turina]

Probably this is due to a misconfigured hippo.
Some time ago it happened to me to add:

	fall=true

to hippo.cfg and I had similar symptoms (except that packages different 
from hippo would succeed the update from time to time).

Ciao,

	Danilo

John Hackett wrote:
> I've just be told that a client is having this problem - they have a 
> webserver running on win2k/sp4 with secure access via OpenSSH/Cygwin and 
> the auto update is failing.
> 
> I've joined this list specifically to track this issue: is there a 
> solution yet?
> 
> Any info would be very helpful. Thanks.
> 
> Regards,
> 
> John Hackett
> Icon Information Systems Ltd.
> t: 020 8764 2663
> f: 08700 525759
> m: 07801 231118
> e: john.hackett@icon-is.co.uk
> w: www.icon-is.co.uk
> 
> 

RE: cygwin + windows update = lock up (W2K SP4)

[Dave Korn]

On 19 May 2006 13:03, John Hackett wrote:

> I've just be told that a client is having this problem - they have a
> webserver running on win2k/sp4 with secure access via OpenSSH/Cygwin
> and the auto update is failing.
> 
> I've joined this list specifically to track this issue: is there a
> solution yet?

  Boy, did you ever choose the wrong list!

> Any info would be very helpful. Thanks.

  *Any* info?  Ok....


Hippopotamus  	 

 

Common hippopotamus: Hippopotamus amphibius

Pygmy hippopotamus: Hexaprotodon liberiensis

Distribution: once numerous in rivers throughout Africa but now extinct north
of Khartoum and south of the Zambezi river, except in a few protected areas
such as the Kruger National Park. The Pygmy hippopotamus lives only in
Liberia, Guinea, Ivory Coast, Sierra Leone and possibly part of southern
Nigeria (though it may be extinct here).

Habitat: in and around lakes. The pygmy hippo lives in forest streams.

Description: huge body supported on short, pillar-like legs, each with four
toes ending in hoof-like nails, placed well apart. Eyes are raised on top of
the large, flattish head. Small ears and nostrils slit-like and high up on the
muzzle. Hairless body, except for sparse bristles on the muzzle, inside the
ears and on the tip of the short tail. Large mouth armed with canine tusks;
these average 0.75m but may be over 1.5m long including the long root embedded
in the gums. Pygmy hippo is similar in appearance but its head is smaller in
proportion to the body and it is more pig-like in shape.

Size: up to 4m in length; 1.4m at the shoulder. Weighs up to 4 tonnes. Pygmy
hippo is 1.5m long, 0.8m at the shoulder and weighs up to 270kg.

Food: mainly grass and some water plants.

The common hippopotamus is the second largest living land animal, only
rivalled by the great Indian rhinoceros (the elephant is the largest). Its
name means 'river horse' and the hippo spends most of its time in water.


Hippo Habits

Territory; common hippos live in groups, sometimes known as schools (except
for the pygmy hippo which is usually found singly or in pairs). There are 20 -
100 in each group. In the centre of each group's territory there is a creche
occupied by females and youngsters; the adult males each have a separate area,
known as a refuge, around the creche. The creche is on a sandbar in midstream
or on a raised bank of the river or lake. Special paths lead from the males'
refuges to the feeding grounds.

The female hippos are the leaders of the schools. When the young males leave
the creche, they have to take up a refuge beyond the ring of adult males'
refuges lying around the edge of the creche. Each then has to try and win his
way to an inner refuge by fighting - and this entitles him to mate with one of
the females.

Fights can be very fierce. The hippo's familiar yawn is actually an aggressive
gesture, a challenge to fight. The two contestants rear up out of the water,
huge mouths wide open, trying to slash each other with their long tusks.
Terrible gashes are inflicted but they quickly heal. The aim of a fight is to
break a foreleg of an opponent - this is fatal because the animal can no
longer walk on land to feed. There are strict rules concerning the behaviour
of the males. Outside the breeding season a female may pay a social call on a
male and he may visit her - but he must enter the creche with no sign of
aggression. If a female gets up on her feet, the male must lie down - and he
may only get up when she lies down again! A male who does not keep to these
rules is driven out by attacking females.

Daily Life; a hippo spends most of its day basking on a sandbar, or lazing in
the water with just its ears, eyes and nostrils, and perhaps its back and top
of the head, exposed. It feeds mostly at night, coming on land to eat mainly
grass. During one night an individual may wander up to 20 miles but usually
does not venture far from water. The areas of grass, kept short by the hippos,
are known as hippo lawns.

If frightened, common hippos rush into the water but pygmy hippos prefer to
head for the undergrowth. Pygmy hippos are very shy and hard to find. They are
usually found singly or in pairs. They are not usually aggressive unless
disturbed.

Breeding; when the female hippo is ready to breed she goes out to choose a
mate and he must behave in a respectful manner as she enters his refuge.

The baby is born after a gestation period of 210 - 255 days. It is 1m long,
0.5m at the shoulder and weighs 27kg. The birth may take place in the water
but normally it is on land, the mother preparing a bed of trampled reeds. 5
minutes after its birth, the baby can walk, run or swim. The mother soon takes
her baby for walks on land, not along the usual paths taken when going to the
feeding grounds, but all over the place. She teaches it to walk level with her
neck, so she can keep an eye on it, and it must stop when she stops. In water
the baby must swim level with her shoulder so that she can quickly protect it
with her body from any aggressive males. Later, when going to the feeding
pastures, she teaches it to walk at heel.

A mother hippo is very strict with her baby and if it is disobedient she
punishes it by lashing it with her head, often rolling it over and over. She
may even slash it with her tusks. When the baby cowers in submission, its
mother stops punishing it and begins to lick and caress it.

Babysitting; if a female leaves the creche for any reason she places her baby
in charge of another female who may already be supervising several others. The
young hippos play with others of similar age, the females together playing a
form of hide-and-seek or rolling over in the water with stiff legs. Young
males play similar games but also indulge in mock fights.


Hippos and Man

Common hippo; information collected from 34 African countries suggests that
the total population of the common hippo in the whole of Africa is about
157,000 animals. They are not common in West Africa and the population is
split into a number of small groups totalling about 7,000 spread over 19
countries. East Africa has about 70,000 and Southern Africa's total is around
80,000.

Numbers are decreasing in 18 African countries and are stable in only six. In
only two countries, Congo and Zambia, are hippo numbers increasing.
Populations most at risk are those in West Africa. Many of the groups there
contain less than 50 animals each and in order to be relatively free from the
risk of extinction, each population should probably number around 500.

The most serious threat to the future of the hippo is loss of habitat. There
is usually no shortage of water for daytime retreats, except in times of
serious drought, but there are reports of loss of grazing habitat to
cultivation.

Hunting for meat, skins or for the trophy trade, occurs in a few countries but
is probably not a serious threat.

Hippos do come into conflict with people quite often. Crops are damaged and
fishermen kill the hippo because of its attacks on them. As the human
population increases, this conflict is likely to become more serious. There is
also the possibility that the canine teeth could be used in trade as a
substitute for elephant ivory - the species would be at serious risk if this
were to happen.

At the moment, the common hippo does not appear to be in danger of extinction
but it is important to make sure that all hippo groups are protected to ensure
their future survival. It must be remembered that the hippo can be a dangerous
animal and any conservation plans have to take into consideration the safety
of human populations.

Pygmy hippo; the pygmy hippo occurs mainly in Liberia, where it is widely
distributed although it does not occur in large populations anywhere. The
numbers are unknown but probably total a few thousand at the most. It is
described as 'vulnerable' amongst the threatened animals of the world.
Although numbers have probably declined recently there is, as yet, not enough
evidence to class it as actually 'endangered'.

Although it is fully protected legally in all countries, the pygmy hippo is
hunted for meat throughout its range except, apparently, in Guinea - perhaps
it is too hard to find. It is not known for sure whether hunting is a threat
to its survival, but with such small, scattered populations any killing is
bound to have an effect. The main threat, however, is loss of habitat through
forest clearance, particularly in Liberia, and, if the pygmy hippo is to
survive, the forests in which it lives should be protected. Fortunately, it
does breed well in captivity and there are 350 animals in collections around
the world. One day it may be possible to reintroduce captive-bred animals into
suitable, protected forests.

The data used in this last section, 'Hippos and Man', was compiled and
provided by The World Conservation Monitoring Centre under sponsorship by BT. 

[ http://www.yptenc.org.uk/docs/factsheets/animal_facts/hippopotamus.html ]

    cheers,
      DaveK
-- 
Can't think of a witty .sigline today....

Re: cygwin + windows update = lock up (W2K SP4)

[John Hackett]

Funny but not terribly useful - esp. as you seem to have posted to  
this list about this problem earlier in the month and I was following  
up on this post from cfg to see whether there was a solution posted  
anywhere yet

"I'm not exactly sure where to send this, since it is a unique  
situation,
but someone has contacted me off-list to offer to pay me to look into
this problem.  Money is always a great motivator for me, so, I've let
Dave Korn know about this and have spent some time today looking into
the problem.

I hope that I'll have a fix sometime next week.

So, I guess this means that I've retracted my retraction.  I do
appreciate the fact that Dave was looking at this very much, though.  It
really is nice to have other people besides Corinna and me looking at
the code.

If anyone is looking for something to do, it would be nice to have
another set of eyes on the snapshot problems that I mentioned in the
cygwin mailing list.

cgf"

If you have some *relevant* help then it would be appreciated.  
Sarcasm is not!

On 19 May 2006, at 13:15, Dave Korn wrote:

> On 19 May 2006 13:03, John Hackett wrote:
>
>> I've just be told that a client is having this problem - they have a
>> webserver running on win2k/sp4 with secure access via OpenSSH/Cygwin
>> and the auto update is failing.
>>
>> I've joined this list specifically to track this issue: is there a
>> solution yet?
>
>   Boy, did you ever choose the wrong list!
>
>> Any info would be very helpful. Thanks.
>
>   *Any* info?  Ok....
>
> ...

RE: cygwin + windows update = lock up (W2K SP4)

[Dave Korn]

On 19 May 2006 13:30, John Hackett wrote:

> Funny but not terribly useful

  Yep, that's the one-line description of this mailing list!

> - esp. as you seem to have posted to
> this list about this problem earlier in the month 

  That doesn't constitute any kind of binding obligation on me to be serious
at any given time in the future.  False extrapolation.

> If you have some *relevant* help then it would be appreciated.
> Sarcasm is not!

  Well, there you go.  Sarcasm and irrelevant information are two entirely
different things, and "getting a serious answer on a silly mailing list" is
not listed in the universal declaration of human rights.

    cheers,
      DaveK
-- 
Can't think of a witty .sigline today....

Re: cygwin + windows update = lock up (W2K SP4)

[mwoehlke]

John Hackett left some http://cygwin.com/acronyms/#TOFU :
>> On 19 May 2006 13:03, John Hackett wrote:
>>> I've joined this list specifically to track this issue: is there a
>>> solution yet?
>>
>> Boy, did you ever choose the wrong list!
>>
>>> Any info would be very helpful. Thanks.
>>
>> *Any* info?  Ok....
>> [snip]
> 
> Funny but not terribly useful
> [snip]
> If you have some *relevant* help then it would be appreciated. Sarcasm 
> is not!

Quote: "Boy, did you ever choose the wrong list!"

As Dave Korn pointed out, you are on the wrong list for that. Please 
review the descriptions at <http://cygwin.com/lists.html>. You want 
<cygwin AT cygwin DOT com>.

(Also, please note what I change my reply header to.)

-- 
Matthew
If you can drive a cart and horses through, will a hippo fit?

RE: cygwin + windows update = lock up (W2K SP4)

[Dave Korn]

On 19 May 2006 16:12, mwoehlke wrote:

> John Hackett left some http://cygwin.com/acronyms/#TOFU :
>>> On 19 May 2006 13:03, John Hackett wrote:
>>>> I've joined this list specifically to track this issue: is there a
>>>> solution yet?
>>> 
>>> Boy, did you ever choose the wrong list!
>>> 
>>>> Any info would be very helpful. Thanks.
>>> 
>>> *Any* info?  Ok....
>>> [snip]
>> 
>> Funny but not terribly useful
>> [snip]
>> If you have some *relevant* help then it would be appreciated. Sarcasm
>> is not!
> 
> Quote: "Boy, did you ever choose the wrong list!"
> 
> As Dave Korn pointed out, you are on the wrong list for that. Please
> review the descriptions at <http://cygwin.com/lists.html>. You want
> <cygwin AT cygwin DOT com>.
> 
> (Also, please note what I change my reply header to.)


  Also, he should please note what I changed my To: header to!

    cheers,
      DaveK
-- 
Can't think of a witty .sigline today....

Re: cygwin + windows update = lock up (W2K SP4)

[mwoehlke]

Dave Korn wrote:
>   Also, he should please note what I changed my To: header to!

Meh, you sneaky people, you! ;-)

Anyone know if there is an easy (i.e. does not involve recompiling) way 
to change what headers Thunderbird considers "normal"?

-- 
Matthew
If you can drive a cart and horses through, will a hippo fit?

Re: cygwin + windows update = lock up (W2K SP4)

[Christopher Faylor]

On Fri, May 19, 2006 at 01:29:58PM +0100, John Hackett wrote:
>Funny but not terribly useful - esp. as you seem to have posted to  
>this list about this problem earlier in the month and I was following  
>up on this post from cfg to see whether there was a solution posted  
>anywhere yet
>
>"I'm not exactly sure where to send this, since it is a unique  
>situation,
>but someone has contacted me off-list to offer to pay me to look into
>this problem.  Money is always a great motivator for me, so, I've let
>Dave Korn know about this and have spent some time today looking into
>the problem.
>
>I hope that I'll have a fix sometime next week.
>
>So, I guess this means that I've retracted my retraction.  I do
>appreciate the fact that Dave was looking at this very much, though.  It
>really is nice to have other people besides Corinna and me looking at
>the code.
>
>If anyone is looking for something to do, it would be nice to have
>another set of eyes on the snapshot problems that I mentioned in the
>cygwin mailing list.
>
>cgf"
>
>If you have some *relevant* help then it would be appreciated.  
>Sarcasm is not!

So, are you saying that you're going to pay me, too?  Oh boy! This
problem is going to be a gold mine!

I'm thinking of a sliding scale charging fee.  $1000 to fix infinite loops.
$500 to fix infinite/2 loops, $250 for inifinite/4 loops, etc.

Anway, yes, I believe that the problem is fixed in the latest snapshot.

And, in case it isn't clear by now, this mailing list isn't really the
best place for technical discussions or queries.  That's why it's listed
as a "mindless chatter list" at http://cygwin.com/lists.html .

cgf

RE: cygwin + windows update = lock up (W2K SP4)

[Dave Korn]

On 19 May 2006 16:27, Christopher Faylor wrote:

> 
> Anway, yes, I believe that the problem is fixed in the latest snapshot.

  I certainly can't reproduce it any more.

  Oh, hang on, I did just upgrade my ProcExp to the latest version.
<double-checks with old version>  Nope, it's really gone!
 
> And, in case it isn't clear by now, this mailing list isn't really the
> best place for technical discussions or queries.  That's why it's listed
> as a "mindless chatter list" at http://cygwin.com/lists.html .

  EWEeeeeeeglebleegie oogyboogyblap <bibbles lips with
finger>bl-bl-bl-bl-bl-bl-bl-bl-bl-bl YAK SQUAWK BLOP gazebo SPATULA beluga
BULBOUS BOUFFANT!


    cheers,
      DaveK
-- 
Can't think of a witty .sigline today....

RE: cygwin + windows update = lock up (W2K SP4)

[Gary R. Van Sickle]

[snip]
> > Quote: "Boy, did you ever choose the wrong list!"
> > 
> > As Dave Korn pointed out, you are on the wrong list for 
> that. Please 
> > review the descriptions at <http://cygwin.com/lists.html>. You want 
> > <cygwin AT cygwin DOT com>.
> > 
> > (Also, please note what I change my reply header to.)
> 
> 
>   Also, he should please note what I changed my To: header to!
> 
>     cheers,
>       DaveK

Huh?  If he did that he'd get banned!

-- 
Gary R. Van Sickle
 

RE: cygwin + windows update = lock up (W2K SP4)

[Igor Peshansky]

On Fri, 19 May 2006, Dave Korn wrote:

> On 19 May 2006 16:27, Christopher Faylor wrote:
>
> > Anway, yes, I believe that the problem is fixed in the latest snapshot.
>
>   I certainly can't reproduce it any more.
>
>   Oh, hang on, I did just upgrade my ProcExp to the latest version.
> <double-checks with old version>  Nope, it's really gone!
>
> > And, in case it isn't clear by now, this mailing list isn't really the
> > best place for technical discussions or queries.  That's why it's listed
> > as a "mindless chatter list" at http://cygwin.com/lists.html .
>
>   EWEeeeeeeglebleegie oogyboogyblap <bibbles lips with
> finger>bl-bl-bl-bl-bl-bl-bl-bl-bl-bl YAK SQUAWK BLOP gazebo SPATULA
> beluga BULBOUS BOUFFANT!

Now, see what you've done, CGF -- you made Dave froth at the mouth...
	Igor
-- 
				http://cs.nyu.edu/~pechtcha/
      |\      _,,,---,,_	    pechtcha@cs.nyu.edu | igor@watson.ibm.com
ZZZzz /,`.-'`'    -.  ;-;;,_		Igor Peshansky, Ph.D. (name changed!)
     |,4-  ) )-,_. ,\ (  `'-'		old name: Igor Pechtchanski
    '---''(_/--'  `-'\_) fL	a.k.a JaguaR-R-R-r-r-r-.-.-.  Meow!

"Las! je suis sot... -Mais non, tu ne l'es pas, puisque tu t'en rends compte."
"But no -- you are no fool; you call yourself a fool, there's proof enough in
that!" -- Rostand, "Cyrano de Bergerac"

Re: cygwin + windows update = lock up (W2K SP4)

[Christopher Faylor]

On Fri, May 19, 2006 at 09:01:17PM -0400, Igor Peshansky wrote:
>On Fri, 19 May 2006, Dave Korn wrote:
>>On 19 May 2006 16:27, Christopher Faylor wrote:
>>>Anway, yes, I believe that the problem is fixed in the latest snapshot.
>>
>>I certainly can't reproduce it any more.
>>
>>Oh, hang on, I did just upgrade my ProcExp to the latest version.
>><double-checks with old version> Nope, it's really gone!
>>
>>>And, in case it isn't clear by now, this mailing list isn't really the
>>>best place for technical discussions or queries.  That's why it's
>>>listed as a "mindless chatter list" at http://cygwin.com/lists.html .
>>
>>EWEeeeeeeglebleegie oogyboogyblap <bibbles lips with
>>finger>bl-bl-bl-bl-bl-bl-bl-bl-bl-bl YAK SQUAWK BLOP gazebo SPATULA
>>beluga BULBOUS BOUFFANT!
>
>Now, see what you've done, CGF -- you made Dave froth at the mouth...

But, really, that really wasn't very hard to do.  :-)

I sure do like pie.  It is strange how few restaurants serve it, though.

cgf

Re: cygwin + windows update = lock up (W2K SP4)

[Igor Peshansky]

On Fri, 19 May 2006, Christopher Faylor wrote:

> On Fri, May 19, 2006 at 09:01:17PM -0400, Igor Peshansky wrote:
> >On Fri, 19 May 2006, Dave Korn wrote:
> >>On 19 May 2006 16:27, Christopher Faylor wrote:
> >>>Anway, yes, I believe that the problem is fixed in the latest snapshot.
> >>
> >>I certainly can't reproduce it any more.
> >>
> >>Oh, hang on, I did just upgrade my ProcExp to the latest version.
> >><double-checks with old version> Nope, it's really gone!
> >>
> >>>And, in case it isn't clear by now, this mailing list isn't really the
> >>>best place for technical discussions or queries.  That's why it's
> >>>listed as a "mindless chatter list" at http://cygwin.com/lists.html .
> >>
> >>EWEeeeeeeglebleegie oogyboogyblap <bibbles lips with
> >>finger>bl-bl-bl-bl-bl-bl-bl-bl-bl-bl YAK SQUAWK BLOP gazebo SPATULA
> >>beluga BULBOUS BOUFFANT!
> >
> >Now, see what you've done, CGF -- you made Dave froth at the mouth...
>
> But, really, that really wasn't very hard to do.  :-)
>
> I sure do like pie.  It is strange how few restaurants serve it, though.

Because most of them don't implement RFC3091...
But here's one that serves a pretty good pi:
<http://3.141592653589793238462643383279502884197169399375105820974944592.com/index1.html>
	Igor
[*] <http://www.ietf.org/rfc/rfc3091.txt>
[*] <http://www.rfc-archive.org/getrfc.php?rfc=3091>
-- 
				http://cs.nyu.edu/~pechtcha/
      |\      _,,,---,,_	    pechtcha@cs.nyu.edu | igor@watson.ibm.com
ZZZzz /,`.-'`'    -.  ;-;;,_		Igor Peshansky, Ph.D. (name changed!)
     |,4-  ) )-,_. ,\ (  `'-'		old name: Igor Pechtchanski
    '---''(_/--'  `-'\_) fL	a.k.a JaguaR-R-R-r-r-r-.-.-.  Meow!

"Las! je suis sot... -Mais non, tu ne l'es pas, puisque tu t'en rends compte."
"But no -- you are no fool; you call yourself a fool, there's proof enough in
that!" -- Rostand, "Cyrano de Bergerac"

RE: cygwin + windows update = lock up (W2K SP4)

[Igor Peshansky]

On Fri, 19 May 2006, Dave Korn wrote:

> On 19 May 2006 16:27, Christopher Faylor wrote:
>
> > Anway, yes, I believe that the problem is fixed in the latest snapshot.
>
>   I certainly can't reproduce it any more.
>
>   Oh, hang on, I did just upgrade my ProcExp to the latest version.
> <double-checks with old version>  Nope, it's really gone!

Incidentally, this also (partly) fixed another pretty annoying problem I
had with yesterday's CVS build: starting an xterm via a Windows
(Hummingbird) inetd service popped up a DOS console titled
"c:\cygwin\bin\xterm.exe".  That console doesn't pop up anymore (all I did
was pick up CGF's patch to dtable.cc and rebuild from CVS).  However,
starting an xterm with a shortcut or cygstart still results in an extra
console window.  I should probably report it properly on the main list --
feel free to drop a hippo on me.
	Igor
-- 
				http://cs.nyu.edu/~pechtcha/
      |\      _,,,---,,_	    pechtcha@cs.nyu.edu | igor@watson.ibm.com
ZZZzz /,`.-'`'    -.  ;-;;,_		Igor Peshansky, Ph.D. (name changed!)
     |,4-  ) )-,_. ,\ (  `'-'		old name: Igor Pechtchanski
    '---''(_/--'  `-'\_) fL	a.k.a JaguaR-R-R-r-r-r-.-.-.  Meow!

"Las! je suis sot... -Mais non, tu ne l'es pas, puisque tu t'en rends compte."
"But no -- you are no fool; you call yourself a fool, there's proof enough in
that!" -- Rostand, "Cyrano de Bergerac"

RE: cygwin + windows update = lock up (W2K SP4)

[Igor Peshansky]

On Fri, 19 May 2006, Igor Peshansky wrote:

> On Fri, 19 May 2006, Dave Korn wrote:
>
> > On 19 May 2006 16:27, Christopher Faylor wrote:
> >
> > > Anway, yes, I believe that the problem is fixed in the latest snapshot.
> >
> >   I certainly can't reproduce it any more.
> >
> >   Oh, hang on, I did just upgrade my ProcExp to the latest version.
> > <double-checks with old version>  Nope, it's really gone!
>
> Incidentally, this also (partly) fixed another pretty annoying problem I
> had with yesterday's CVS build: starting an xterm via a Windows
> (Hummingbird) inetd service popped up a DOS console titled
> "c:\cygwin\bin\xterm.exe".  That console doesn't pop up anymore (all I did
> was pick up CGF's patch to dtable.cc and rebuild from CVS).  However,
                             ^^^^^^^^^
Whoops!  Make that fhandler_console.cc.  This isn't called the mindless
chatter list for nothing, I guess.

> starting an xterm with a shortcut or cygstart still results in an extra
> console window.  I should probably report it properly on the main list --
> feel free to drop a hippo on me.

Hippos away!
	Igor
-- 
				http://cs.nyu.edu/~pechtcha/
      |\      _,,,---,,_	    pechtcha@cs.nyu.edu | igor@watson.ibm.com
ZZZzz /,`.-'`'    -.  ;-;;,_		Igor Peshansky, Ph.D. (name changed!)
     |,4-  ) )-,_. ,\ (  `'-'		old name: Igor Pechtchanski
    '---''(_/--'  `-'\_) fL	a.k.a JaguaR-R-R-r-r-r-.-.-.  Meow!

"Las! je suis sot... -Mais non, tu ne l'es pas, puisque tu t'en rends compte."
"But no -- you are no fool; you call yourself a fool, there's proof enough in
that!" -- Rostand, "Cyrano de Bergerac"

RE: cygwin + windows update = lock up (W2K SP4)

[Dave Korn]

On 20 May 2006 00:22, Gary R. Van Sickle wrote:

> [snip]
>>> Quote: "Boy, did you ever choose the wrong list!"
>>> 
>>> As Dave Korn pointed out, you are on the wrong list for that. Please
>>> review the descriptions at <http://cygwin.com/lists.html>. You want
>>> <cygwin AT cygwin DOT com>. 
>>> 
>>> (Also, please note what I change my reply header to.)
>> 
>> 
>>   Also, he should please note what I changed my To: header to!
>> 
>>     cheers,
>>       DaveK
> 
> Huh?  If he did that he'd get banned!


  I'm fairly sure nobody ever got banned for trying a snapshot.


 ...  With the possible exception of papparazzi.

    cheers,
      DaveK
-- 
Can't think of a witty .sigline today....

Re: cygwin + windows update = lock up (W2K SP4)

[Christopher Faylor]

On Wed, May 03, 2006 at 01:46:39PM -0400, Christopher Faylor wrote:
>On Wed, May 03, 2006 at 06:30:15PM +0100, Dave Korn wrote:
>>On 03 May 2006 16:59, Christopher Faylor wrote:
>>>On Wed, May 03, 2006 at 11:12:02AM +0200, Ludovic Drolez wrote: Once I
>>>saw that Dave was on the case and could reproduce the problem at will,
>>>I haven't been pursuing every report of this terrifying bug because we
>>>actually do have someone investigating it.  I did try running process
>>>explorer while also running a cygwin process and did not see the
>>>problem.
>>
>>You very specifically have to bring up the properties page for the
>>process and display the "Threads" tab.  It doesn't occur (any more)
>>when you just select a process or even when you bring up the other
>>properties page tabs.
>
>I'll take your word for it.
>
>Ok.  Let me almost completely retract what I said earlier.
>
>I'm not going to actively try to fix the problem since you are currently
>working on it and I see no reason to duplicate effort.
>
>If this scenario involved Corinna and me then we would gratefully let
>the other person work on the problem.  I appreciate the fact that you
>are looking into this and have no plans on actively looking into the
>issue myself.  I may still provide a patch if an idea percolates into my
>conscious mind (and, as usual, I have been thinking about this problem),
>though.

I'm not exactly sure where to send this, since it is a unique situation,
but someone has contacted me off-list to offer to pay me to look into
this problem.  Money is always a great motivator for me, so, I've let
Dave Korn know about this and have spent some time today looking into
the problem.

I hope that I'll have a fix sometime next week.

So, I guess this means that I've retracted my retraction.  I do
appreciate the fact that Dave was looking at this very much, though.  It
really is nice to have other people besides Corinna and me looking at
the code.

If anyone is looking for something to do, it would be nice to have
another set of eyes on the snapshot problems that I mentioned in the
cygwin mailing list.

cgf

RE: cygwin + windows update = lock up (W2K SP4)

[Dave Korn]

On 03 May 2006 18:47, Christopher Faylor wrote:

> On Wed, May 03, 2006 at 06:30:15PM +0100, Dave Korn wrote:
>> On 03 May 2006 16:59, Christopher Faylor wrote:
>>> On Wed, May 03, 2006 at 11:12:02AM +0200, Ludovic Drolez wrote: Once I
>>> saw that Dave was on the case and could reproduce the problem at will,
>>> I haven't been pursuing every report of this terrifying bug because we
>>> actually do have someone investigating it.  I did try running process
>>> explorer while also running a cygwin process and did not see the
>>> problem.
>> 
>> You very specifically have to bring up the properties page for the
>> process and display the "Threads" tab.  It doesn't occur (any more)
>> when you just select a process or even when you bring up the other
>> properties page tabs.
> 
> I'll take your word for it.

  Sure, it was only FYI since you happened to mention it.  No need to try and
reproduce it.

> I'm not going to actively try to fix the problem since you are currently
> working on it and I see no reason to duplicate effort.

  Makes sense to me too.

> issue myself.  I may still provide a patch if an idea percolates into my
> conscious mind (and, as usual, I have been thinking about this problem),
> though.

  If you have any interesting thoughts that don't come in the form of a patch,
please nonetheless throw them out here for me to bat around.  Otherwise my
approach (given the problems in debugging this remote thread) is going to be
to add masses of instrumentation to all the dllmain and init stuff and see if
I can figure out where it's going wrong.


    cheers,
      DaveK
-- 
Can't think of a witty .sigline today....

Re: cygwin + windows update = lock up (W2K SP4)

[Christopher Faylor]

On Wed, May 03, 2006 at 06:30:15PM +0100, Dave Korn wrote:
>On 03 May 2006 16:59, Christopher Faylor wrote:
>>On Wed, May 03, 2006 at 11:12:02AM +0200, Ludovic Drolez wrote: Once I
>>saw that Dave was on the case and could reproduce the problem at will,
>>I haven't been pursuing every report of this terrifying bug because we
>>actually do have someone investigating it.  I did try running process
>>explorer while also running a cygwin process and did not see the
>>problem.
>
>You very specifically have to bring up the properties page for the
>process and display the "Threads" tab.  It doesn't occur (any more)
>when you just select a process or even when you bring up the other
>properties page tabs.

I'll take your word for it.

Ok.  Let me almost completely retract what I said earlier.

I'm not going to actively try to fix the problem since you are currently
working on it and I see no reason to duplicate effort.

If this scenario involved Corinna and me then we would gratefully let
the other person work on the problem.  I appreciate the fact that you
are looking into this and have no plans on actively looking into the
issue myself.  I may still provide a patch if an idea percolates into my
conscious mind (and, as usual, I have been thinking about this problem),
though.

cgf

RE: cygwin + windows update = lock up (W2K SP4)

[Dave Korn]

On 03 May 2006 16:59, Christopher Faylor wrote:

> On Wed, May 03, 2006 at 11:12:02AM +0200, Ludovic Drolez wrote:

> Once I saw that Dave was on the case and could reproduce the problem at
> will, I haven't been pursuing every report of this terrifying bug
> because we actually do have someone investigating it.  I did try running
> process explorer while also running a cygwin process and did not see the
> problem.  

  You very specifically have to bring up the properties page for the process
and display the "Threads" tab.  It doesn't occur (any more) when you just
select a process or even when you bring up the other properties page tabs.

> In other projects (and expecially in projects the size of cygwin) there
> are usually a group of people who are perfectly capable of fixing bugs
> without involving the principals.  I am working under the assumption
> that Dave Korn is one of those people.  I had, in fact, hoped that, just
> this once, someone else would track down a thorny issue in cygwin.
> 
> IMO, it is insulting to me to suggest that I am cruelly ignoring a
> problem and it is insulting to him to suggest that the only person who
> can fix the problem is me.

  It's the same old problem: the circumference of my tuits has diverged quite
a long way from 3.14159... times their diameter.  But I think I can probably
work it out given enough patience :) and a bit of luck....


    cheers,
      DaveK
-- 
Can't think of a witty .sigline today....

Re: cygwin + windows update = lock up (W2K SP4)

[Christopher Faylor]

On Wed, May 03, 2006 at 11:12:02AM +0200, Ludovic Drolez wrote:
>>If either Corinna or I could reproduce the problem it would be fixed by
>>now.
>
>Yes it's very easy to reproduce, you just need cat.exe and a W2K SP4 as
>you can read here:
>
>http://www.cygwin.com/ml/cygwin/2006-05/msg00044.html

I knew that I'd regret making that statement.  I guess I have to clarify.

Once I saw that Dave was on the case and could reproduce the problem at
will, I haven't been pursuing every report of this terrifying bug
because we actually do have someone investigating it.  I did try running
process explorer while also running a cygwin process and did not see the
problem.  I've tried other scenarios which have supposedly tickled the
bug, too.  I believe Corinna has, as well.

In other projects (and expecially in projects the size of cygwin) there
are usually a group of people who are perfectly capable of fixing bugs
without involving the principals.  I am working under the assumption
that Dave Korn is one of those people.  I had, in fact, hoped that, just
this once, someone else would track down a thorny issue in cygwin.

IMO, it is insulting to me to suggest that I am cruelly ignoring a
problem and it is insulting to him to suggest that the only person who
can fix the problem is me.

cgf

RE: cygwin + windows update = lock up (W2K SP4)

[Dave Korn]

On 03 May 2006 13:33, Gary R. Van Sickle wrote:

>> From: Igor Peshansky
>> Sent: Wednesday, May 03, 2006 7:22 AM
>> To: The Cygwin-Talk Maiming List
>> Subject: RE: cygwin + windows update = lock up (W2K SP4)
>> 
>> On Wed, 3 May 2006, Dave Korn wrote:
>> 
> [snip]
>>>     cheers,
>>>       DaveK
>> 
>> [Attachment: cygdeath.c (2KB) deleted]
>> 
>> My only question is: why isn't it using the undocumented WinAPI function
>> ProcessDropHippo() (the 48-argument version)?
>> 
>> IOW, aren't you on the wrong list for this useful technical discussion?
>> 	Igor

  If you look closely, you'll see it had the dreaded "-mno-cygwin" flag, which
makes it off-topic for the main list.  Nt native API coding is non-cygwin
related IMO.
 
> Indeed.  That's easily your third warning Korns....

  Heavens!  Unmitigated anarchy prevails!  Oh, will no-one think of teh poor
'ickul chickens?

    cheers,
      DaveK
-- 
Can't think of a witty .sigline today....

RE: cygwin + windows update = lock up (W2K SP4)

[Gary R. Van Sickle]

> From: Igor Peshansky
> Sent: Wednesday, May 03, 2006 7:22 AM
> To: The Cygwin-Talk Maiming List
> Subject: RE: cygwin + windows update = lock up (W2K SP4)
> 
> On Wed, 3 May 2006, Dave Korn wrote:
> 
[snip]
> >     cheers,
> >       DaveK
> 
> [Attachment: cygdeath.c (2KB) deleted]
> 
> My only question is: why isn't it using the undocumented 
> WinAPI function
> ProcessDropHippo() (the 48-argument version)?
> 
> IOW, aren't you on the wrong list for this useful technical 
> discussion?
> 	Igor

Indeed.  That's easily your third warning Korns....

-- 
Gary R. Van Sickle
 

RE: cygwin + windows update = lock up (W2K SP4)

[Igor Peshansky]

On Wed, 3 May 2006, Dave Korn wrote:

> On 02 May 2006 22:11, Christopher Faylor wrote:
>
> > On Tue, May 02, 2006 at 01:59:12PM -0700, Shankar Unni wrote:
> >> Brian Dessent wrote:
> >>> It's a real shame that Russinovich is apparently aware of this
> >>> defect but remains uninterested in dealing with it past calling it a
> >>> Cygwin deficiency.
> >>
> >> Time for a Russinovitch-CGF cage match!
> >>
> >> (First one who tries to debug the problem loses..)
> >
> > If either Corinna or I could reproduce the problem it would be fixed
> > by now.
> >
> > cgf
>
>   Attached please find PPAST!
>
>   Compile with "gcc -g -O0 -mno-cygwin cygdeath.c -o cygdeath -l ntdll"
>
>   Runs with one arg: windows pid of any handy cygwin process, such as
> the shell you've just used to compile it in, or an instance of cat, or
> whatever.
>
>   Press any key when it tells you.  Watch csrss and chosen cygwin
> process start to thrash.
>
>     cheers,
>       DaveK

[Attachment: cygdeath.c (2KB) deleted]

My only question is: why isn't it using the undocumented WinAPI function
ProcessDropHippo() (the 48-argument version)?

IOW, aren't you on the wrong list for this useful technical discussion?
	Igor
-- 
				http://cs.nyu.edu/~pechtcha/
      |\      _,,,---,,_	    pechtcha@cs.nyu.edu | igor@watson.ibm.com
ZZZzz /,`.-'`'    -.  ;-;;,_		Igor Peshansky, Ph.D. (name changed!)
     |,4-  ) )-,_. ,\ (  `'-'		old name: Igor Pechtchanski
    '---''(_/--'  `-'\_) fL	a.k.a JaguaR-R-R-r-r-r-.-.-.  Meow!

"Las! je suis sot... -Mais non, tu ne l'es pas, puisque tu t'en rends compte."
"But no -- you are no fool; you call yourself a fool, there's proof enough in
that!" -- Rostand, "Cyrano de Bergerac"

RE: cygwin + windows update = lock up (W2K SP4)

[Dave Korn]

[-- Attachment #1: Type: text/plain, Size: 918 bytes --]

On 02 May 2006 22:11, Christopher Faylor wrote:

> On Tue, May 02, 2006 at 01:59:12PM -0700, Shankar Unni wrote:
>> Brian Dessent wrote:
>>> It's a real shame that Russinovich is apparently aware of this defect
>>> but remains uninterested in dealing with it past calling it a Cygwin
>>> deficiency.
>> 
>> Time for a Russinovitch-CGF cage match!
>> 
>> (First one who tries to debug the problem loses..)
> 
> If either Corinna or I could reproduce the problem it would be fixed by
> now.
> 
> cgf

  Attached please find PPAST!

  Compile with "gcc -g -O0 -mno-cygwin cygdeath.c -o cygdeath -l ntdll"

  Runs with one arg: windows pid of any handy cygwin process, such as the
shell you've just used to compile it in, or an instance of cat, or whatever.

  Press any key when it tells you.  Watch csrss and chosen cygwin process
start to thrash.


    cheers,
      DaveK
-- 
Can't think of a witty .sigline today....

[-- Attachment #2: cygdeath.c --]
[-- Type: text/plain, Size: 1969 bytes --]


#include <windef.h>
#include <winnt.h>
#include <ntdef.h>
#include <ntdll.h>
#include <stdio.h>

typedef struct _DEBUG_BUFFER {
	HANDLE  SectionHandle;
	PVOID  SectionBase;
	PVOID  RemoteSectionBase;
	ULONG  SectionBaseDelta;
	HANDLE  EventPairHandle;
	ULONG  Unknown[2];
	HANDLE  RemoteThreadHandle;
	ULONG  InfoClassMask;
	ULONG  SizeOfInfo;
	ULONG  AllocatedSize;
	ULONG  SectionSize;
	PVOID  ModuleInformation;
	PVOID  BackTraceInformation;
	PVOID  HeapInformation;
	PVOID  LockInformation;
	PVOID  Reserved[8];
} DEBUG_BUFFER, *PDEBUG_BUFFER;

PDEBUG_BUFFER
NTAPI
RtlCreateQueryDebugBuffer(
	IN ULONG  Size,
	IN BOOLEAN  EventPair);

/* RtlQueryProcessDebugInformation.DebugInfoClassMask constants */
#define PDI_MODULES                       0x01
#define PDI_BACKTRACE                     0x02
#define PDI_HEAPS                         0x04
#define PDI_HEAP_TAGS                     0x08
#define PDI_HEAP_BLOCKS                   0x10
#define PDI_LOCKS                         0x20

NTSTATUS
NTAPI
RtlQueryProcessDebugInformation(
	IN ULONG  ProcessId,
	IN ULONG  DebugInfoClassMask,
	IN OUT PDEBUG_BUFFER  DebugBuffer);

NTSTATUS
NTAPI
RtlDestroyQueryDebugBuffer(
  IN PDEBUG_BUFFER  DebugBuffer);



int main (int argc, const char **argv)
{
int n, pid;
PDEBUG_BUFFER buffer;
NTSTATUS rv;

	if (argc != 2)
	{
		fprintf (stderr, "Usage: %s [winpid of cygwin process]\n");
		return -1;
	}
	n = sscanf (argv[1], "%i", &pid);
	if (n != 1)
	{
		fprintf (stderr, "Invalid pid.\n");
		return -1;
	}
	buffer = RtlCreateQueryDebugBuffer (0, 0);
	if (buffer == NULL)
	{
		fprintf (stderr, "Seriously low memory!\n");
		return -1;
	}

	fprintf (stdout, "Now press a key.....");
	getch ();
	rv = RtlQueryProcessDebugInformation (pid, PDI_BACKTRACE, buffer);
	fprintf (stdout, "RtlQueryProcessDebugInformation returned $%08x\n", rv);

	fprintf (stdout, "\nNow press another key.....");
	getch ();
	RtlDestroyQueryDebugBuffer (buffer);
	fprintf (stdout, "\n");
	return 0;
}

Re: cygwin + windows update = lock up (W2K SP4)

[Ludovic Drolez]

> If either Corinna or I could reproduce the problem it would be fixed by
> now.
>
> cgf

Yes it's very easy to reproduce, you just need cat.exe and a W2K SP4 as you can
read here:

http://www.cygwin.com/ml/cygwin/2006-05/msg00044.html

Cheers,

-- 
Ludovic DROLEZ                              Linbox / Free&ALter Soft
www.linbox.com www.linbox.org

Re: cygwin + windows update = lock up (W2K SP4)

[Christopher Faylor]

On Tue, May 02, 2006 at 01:59:12PM -0700, Shankar Unni wrote:
>Brian Dessent wrote:
>>It's a real shame that Russinovich is apparently aware of this defect
>>but remains uninterested in dealing with it past calling it a Cygwin
>>deficiency.
>
>Time for a Russinovitch-CGF cage match!
>
>(First one who tries to debug the problem loses..)

If either Corinna or I could reproduce the problem it would be fixed by
now.

cgf

Re: cygwin + windows update = lock up (W2K SP4)

[Shankar Unni]

Brian Dessent wrote:

> It's a real shame that Russinovich is apparently aware of this defect
> but remains uninterested in dealing with it past calling it a Cygwin
> deficiency.

Time for a Russinovitch-CGF cage match!

(First one who tries to debug the problem loses..)