latest cygcheck -c is expensive

latest cygcheck -c is expensive

[Jason Tishler]

Igor,

Would you be willing to make the status check portion of cygcheck -c
optional (i.e., another option)?  The latest version is very expensive:

    $ # 1.5.3 on P4 2.4 GHz
    $ time cygcheck -c >/dev/null

    real    1m49.646s
    user    0m0.010s
    sys     0m0.020s

    $ # 1.3.22 on P3 500 MHz
    $ time cygcheck -c >/dev/null

    real    0m0.042s
    user    0m0.010s
    sys     0m0.020s

Thanks,
Jason

-- 
PGP/GPG Key: http://www.tishler.net/jason/pubkey.asc or key servers
Fingerprint: 7A73 1405 7F2B E669 C19D  8784 1AFD E4CC ECF4 8EF6

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: latest cygcheck -c is expensive

[Christopher Faylor]

On Fri, Sep 05, 2003 at 08:40:47AM -0400, Jason Tishler wrote:
>Would you be willing to make the status check portion of cygcheck -c
>optional (i.e., another option)?  The latest version is very expensive:
>
>    $ # 1.5.3 on P4 2.4 GHz
>    $ time cygcheck -c >/dev/null
>
>    real    1m49.646s
>    user    0m0.010s
>    sys     0m0.020s
>
>    $ # 1.3.22 on P3 500 MHz
>    $ time cygcheck -c >/dev/null
>
>    real    0m0.042s
>    user    0m0.010s
>    sys     0m0.020s

Why does it matter if cygcheck -c is expensive?  It is supposed to be doing
a sanity check on the installation.

OTOH, it could probably be speeded up considerably if it used a mingw zlib
library...

cgf

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: latest cygcheck -c is expensive

[Jason Tishler]

On Fri, Sep 05, 2003 at 11:07:29AM -0400, Christopher Faylor wrote:
> On Fri, Sep 05, 2003 at 08:40:47AM -0400, Jason Tishler wrote:
> >Would you be willing to make the status check portion of cygcheck -c
> >optional (i.e., another option)?  The latest version is very expensive:
> >
> >[snip]
> 
> Why does it matter if cygcheck -c is expensive?  It is supposed to be
> doing a sanity check on the installation.

In the past, it just displayed the package versions.  Since most of the
time I'm only interested in the package version, I would prefer that the
sanity check be optional.

FWIW, rpm seems to separate the query and verify operations:

    $ rpm -q textutils
    textutils-2.0.11-7
    $ rpm -V textutils
    $

Jason

-- 
PGP/GPG Key: http://www.tishler.net/jason/pubkey.asc or key servers
Fingerprint: 7A73 1405 7F2B E669 C19D  8784 1AFD E4CC ECF4 8EF6

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: latest cygcheck -c is expensive

[Wes Szumera]

> > On Fri, Sep 05, 2003 at 08:40:47AM -0400, Jason Tishler wrote:
> > >Would you be willing to make the status check portion of cygcheck -c
> > >optional (i.e., another option)?  The latest version is very expensive:
> > >

A 1200 mhz athelon with just about all the cygwin stuff (some x 
omited ) I got to a non full screen session:


real    0m53.705s
user    0m0.020s
sys     0m0.040s


***
Using null 

$ time cygcheck -c > /dev/null

gzip: stdout: Bad file descriptor

real    0m27.682s
user    0m0.010s
sys     0m0.050s

Not sure why the gzip message

***
Full screen

real    0m25.996s
user    0m0.010s
sys     0m0.040s

*** 

Shut down my etrust av program and ran full screen

real    0m16.917s
user    0m0.010s
sys     0m0.020s

Just data for the grist mill.

Thanks for CYGWIN!

Wes







--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: latest cygcheck -c is expensive

[Wes Szumera]

Date sent:      	Fri, 5 Sep 2003 11:07:29 -0400
From:           	Christopher Faylor <cgf-rcm@cygwin.com>
To:             	cygwin@cygwin.com
Subject:        	Re: latest cygcheck -c is expensive
Send reply to:  	cygwin@cygwin.com

> On Fri, Sep 05, 2003 at 08:40:47AM -0400, Jason Tishler wrote:
> >Would you be willing to make the status check portion of cygcheck -c
> >optional (i.e., another option)?  The latest version is very expensive:
> >
> >    $ # 1.5.3 on P4 2.4 GHz
> >    $ time cygcheck -c >/dev/null
> >
> >    real    1m49.646s
> >    user    0m0.010s
> >    sys     0m0.020s
> >
> >    $ # 1.3.22 on P3 500 MHz
> >    $ time cygcheck -c >/dev/null
> >
> >    real    0m0.042s
> >    user    0m0.010s
> >    sys     0m0.020s
> 
> Why does it matter if cygcheck -c is expensive?  It is supposed to be doing
> a sanity check on the installation.
> 
> OTOH, it could probably be speeded up considerably if it used a mingw zlib
> library...
> 
> cgf
> 
> --
> Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
> Problem reports:       http://cygwin.com/problems.html
> Documentation:         http://cygwin.com/docs.html
> FAQ:                   http://cygwin.com/faq/
> 



--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: latest cygcheck -c is expensive

[Igor Pechtchanski]

On Fri, 5 Sep 2003, Jason Tishler wrote:

> Igor,
>
> Would you be willing to make the status check portion of cygcheck -c
> optional (i.e., another option)?  The latest version is very expensive:
>
>     $ # 1.5.3 on P4 2.4 GHz
>     $ time cygcheck -c >/dev/null
>
>     real    1m49.646s
>     user    0m0.010s
>     sys     0m0.020s
>
>     $ # 1.3.22 on P3 500 MHz
>     $ time cygcheck -c >/dev/null
>
>     real    0m0.042s
>     user    0m0.010s
>     sys     0m0.020s
>
> Thanks,
> Jason

Jason,

What CGF said.  I assume since cygcheck can also be used to find out the
version of one installed package at a time, it might be a good idea to add
another flag.  However, I'd suggest restructuring the flags while we're at
it, i.e., only print package info if "-c" is specified (just like registry
info is only printed if "-r" is set).  Add a "-e" option that prints
environment; "-n" that checks security stuff (runs "id", checks
/etc/passwd and /etc/group, and so on), and "-i" that checks package
integrity (only valid with "-c").  Then the "-s" flag will only print the
versions of DLLs and the programs found (and the mounts).  Problem reports
will then have to include the output of "cygcheck -scnver", but, IMO, this
is worth the fine-grained control over what's printed by cygcheck.
Opinions?
	Igor
-- 
				http://cs.nyu.edu/~pechtcha/
      |\      _,,,---,,_		pechtcha@cs.nyu.edu
ZZZzz /,`.-'`'    -.  ;-;;,_		igor@watson.ibm.com
     |,4-  ) )-,_. ,\ (  `'-'		Igor Pechtchanski, Ph.D.
    '---''(_/--'  `-'\_) fL	a.k.a JaguaR-R-R-r-r-r-.-.-.  Meow!

"I have since come to realize that being between your mentor and his route
to the bathroom is a major career booster."  -- Patrick Naughton

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: latest cygcheck -c is expensive

[Jason Tishler]

Igor,

On Fri, Sep 05, 2003 at 11:44:31AM -0400, Igor Pechtchanski wrote:
> Problem reports will then have to include the output of "cygcheck
> -scnver", but, IMO, this is worth the fine-grained control over what's
> printed by cygcheck.
> Opinions?

The above sounds good to me.

Thanks,
Jason

-- 
PGP/GPG Key: http://www.tishler.net/jason/pubkey.asc or key servers
Fingerprint: 7A73 1405 7F2B E669 C19D  8784 1AFD E4CC ECF4 8EF6

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: latest cygcheck -c is expensive

[Christopher Faylor]

On Fri, Sep 05, 2003 at 11:44:31AM -0400, Igor Pechtchanski wrote:
>On Fri, 5 Sep 2003, Jason Tishler wrote:
>> Would you be willing to make the status check portion of cygcheck -c
>> optional (i.e., another option)?  The latest version is very expensive:
>>
>>     $ # 1.5.3 on P4 2.4 GHz
>>     $ time cygcheck -c >/dev/null
>>
>>     real    1m49.646s
>>     user    0m0.010s
>>     sys     0m0.020s
>>
>>     $ # 1.3.22 on P3 500 MHz
>>     $ time cygcheck -c >/dev/null
>>
>>     real    0m0.042s
>>     user    0m0.010s
>>     sys     0m0.020s
>
>What CGF said.  I assume since cygcheck can also be used to find out the
>version of one installed package at a time, it might be a good idea to add
>another flag.  However, I'd suggest restructuring the flags while we're at
>it, i.e., only print package info if "-c" is specified (just like registry
>info is only printed if "-r" is set).  Add a "-e" option that prints
>environment; "-n" that checks security stuff (runs "id", checks
>/etc/passwd and /etc/group, and so on), and "-i" that checks package
>integrity (only valid with "-c").  Then the "-s" flag will only print the
>versions of DLLs and the programs found (and the mounts).  Problem reports
>will then have to include the output of "cygcheck -scnver", but, IMO, this
>is worth the fine-grained control over what's printed by cygcheck.
>Opinions?

This is specifically what I was trying to avoid.  The current behavior is
now encoded in the DNA of the cygwin community.  If we change things so
that -rsv doesn't do the "right" thing anymore, we'll be foreever asking
people to do '-scnver'.

I realize that 'cygcheck -c' is not precisely equivalent to 'rpm -q' but
'cygcheck -c' was never intended to be an exact replacement for 'rpm -q'.
If the only concern is that cygcheck takes a long time, now, then, like
I said, that is something that can be rectified.

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: latest cygcheck -c is expensive

[Jason Tishler]

On Fri, Sep 05, 2003 at 03:01:27PM -0400, Christopher Faylor wrote:
> If the only concern is that cygcheck takes a long time, now, then,
> like I said, that is something that can be rectified.

Yes, the above is my real concern.

Jason

-- 
PGP/GPG Key: http://www.tishler.net/jason/pubkey.asc or key servers
Fingerprint: 7A73 1405 7F2B E669 C19D  8784 1AFD E4CC ECF4 8EF6

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: latest cygcheck -c is expensive

[Christopher Faylor]

On Fri, Sep 05, 2003 at 08:42:49PM -0400, Jason Tishler wrote:
>On Fri, Sep 05, 2003 at 03:01:27PM -0400, Christopher Faylor wrote:
>> If the only concern is that cygcheck takes a long time, now, then,
>> like I said, that is something that can be rectified.
>
>Yes, the above is my real concern.

I just hacked cygcheck to avoid calling gzip and use a mingw libz.a
(courtesy of Chuck Wilson) and it takes 26 seconds to complete on my
dual PIII 733MHZ WinXP system.  I have most packages installed.

Is that still too slow?

cgf

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: latest cygcheck -c is expensive

[Christopher Faylor]

On Sun, Sep 07, 2003 at 12:28:05AM -0400, Christopher Faylor wrote:
>On Fri, Sep 05, 2003 at 08:42:49PM -0400, Jason Tishler wrote:
>>On Fri, Sep 05, 2003 at 03:01:27PM -0400, Christopher Faylor wrote:
>>> If the only concern is that cygcheck takes a long time, now, then,
>>> like I said, that is something that can be rectified.
>>
>>Yes, the above is my real concern.
>
>I just hacked cygcheck to avoid calling gzip and use a mingw libz.a
>(courtesy of Chuck Wilson) and it takes 26 seconds to complete on my
>dual PIII 733MHZ WinXP system.  I have most packages installed.
>
>Is that still too slow?

I should add that the previous version took 1 minute, 9 seconds.

cgf

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

RE: latest cygcheck -c is expensive

[Hannu E K Nevalainen (garbage mail)]

> From: cygwin-owner@cygwin.com [mailto:cygwin-owner@cygwin.com]On Behalf
> Of Christopher Faylor

> >On Fri, Sep 05, 2003 at 08:42:49PM -0400, Jason Tishler wrote:
> >>On Fri, Sep 05, 2003 at 03:01:27PM -0400, Christopher Faylor wrote:
> >>> If the only concern is that cygcheck takes a long time, now, then,
> >>> like I said, that is something that can be rectified.
> >>
> >>Yes, the above is my real concern.
> >
> >I just hacked cygcheck to avoid calling gzip and use a mingw libz.a
> >(courtesy of Chuck Wilson) and it takes 26 seconds to complete on my
> >dual PIII 733MHZ WinXP system.  I have most packages installed.
> >
> >Is that still too slow?
>
> I should add that the previous version took 1 minute, 9 seconds.
>
> cgf

 A good enhancement IMO... though;

 I can imagine a situation where someone is trying to automate a thing or
two in cygwin (likely to be postinstall/preremove only?), using cygcheck...
[Workaround: Create output of cygcheck and 'grep' it instead?]

Already two calls, an extra minute (or two, read below), might tend to
"enfuriate" an impatient user (e.g. me! ;-).

NOTE:
 This would undoubtedly lead to questions like "cygwin install froze! Why?"
on this list.

[ cygcheck -c   on W2K 450MHz PII -> about 1 minute]

/Hannu E K Nevalainen, B.Sc. EE - 59?16.37'N, 17?12.60'E
--END OF MESSAGE--


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: latest cygcheck -c is expensive

[Jason Tishler]

Chris,

On Sun, Sep 07, 2003 at 12:30:21AM -0400, Christopher Faylor wrote:
> On Sun, Sep 07, 2003 at 12:28:05AM -0400, Christopher Faylor wrote:
> >On Fri, Sep 05, 2003 at 08:42:49PM -0400, Jason Tishler wrote:
> >>On Fri, Sep 05, 2003 at 03:01:27PM -0400, Christopher Faylor wrote:
> >>> If the only concern is that cygcheck takes a long time, now, then,
> >>> like I said, that is something that can be rectified.
> >>
> >>Yes, the above is my real concern.
> >
> >I just hacked cygcheck to avoid calling gzip and use a mingw libz.a
> >(courtesy of Chuck Wilson) and it takes 26 seconds to complete on my
> >dual PIII 733MHZ WinXP system.  I have most packages installed.
> >
> >Is that still too slow?
> 
> I should add that the previous version took 1 minute, 9 seconds.

Approximately 3 time faster is a significant improvement, but IMO,
anything more than a few seconds is still too "expensive."

What about an option to disable the sanity checking?  In this way,
Cygwin users do not need to learn new tricks, but powers user still get
the control they desire.

BTW, I tried to give cygcheck from the 2003-Sep-08 snapshot a ride.  The
ChangeLog and diff seem to indicate the required changes have been
completed.  Unfortunately, my timing showed no difference and objdump
showed no dependency on Mingw zlib:

    $ objdump -p cygcheck.exe | fgrep 'DLL Name:'
            DLL Name: msvcrt.dll
            DLL Name: msvcrt.dll
            DLL Name: ADVAPI32.DLL
            DLL Name: KERNEL32.dll

Is the above cockpit error on my part?

Thanks,
Jason

-- 
PGP/GPG Key: http://www.tishler.net/jason/pubkey.asc or key servers
Fingerprint: 7A73 1405 7F2B E669 C19D  8784 1AFD E4CC ECF4 8EF6

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: latest cygcheck -c is expensive

[Christopher Faylor]

On Mon, Sep 08, 2003 at 07:50:16AM -0400, Jason Tishler wrote:
>Chris,
>
>On Sun, Sep 07, 2003 at 12:30:21AM -0400, Christopher Faylor wrote:
>> On Sun, Sep 07, 2003 at 12:28:05AM -0400, Christopher Faylor wrote:
>> >On Fri, Sep 05, 2003 at 08:42:49PM -0400, Jason Tishler wrote:
>> >>On Fri, Sep 05, 2003 at 03:01:27PM -0400, Christopher Faylor wrote:
>> >>> If the only concern is that cygcheck takes a long time, now, then,
>> >>> like I said, that is something that can be rectified.
>> >>
>> >>Yes, the above is my real concern.
>> >
>> >I just hacked cygcheck to avoid calling gzip and use a mingw libz.a
>> >(courtesy of Chuck Wilson) and it takes 26 seconds to complete on my
>> >dual PIII 733MHZ WinXP system.  I have most packages installed.
>> >
>> >Is that still too slow?
>> 
>> I should add that the previous version took 1 minute, 9 seconds.
>
>Approximately 3 time faster is a significant improvement, but IMO,
>anything more than a few seconds is still too "expensive."
>
>What about an option to disable the sanity checking?  In this way,
>Cygwin users do not need to learn new tricks, but powers user still get
>the control they desire.
>
>BTW, I tried to give cygcheck from the 2003-Sep-08 snapshot a ride.  The
>ChangeLog and diff seem to indicate the required changes have been
>completed.  Unfortunately, my timing showed no difference and objdump
>showed no dependency on Mingw zlib:
>
>    $ objdump -p cygcheck.exe | fgrep 'DLL Name:'
>            DLL Name: msvcrt.dll
>            DLL Name: msvcrt.dll
>            DLL Name: ADVAPI32.DLL
>            DLL Name: KERNEL32.dll
>
>Is the above cockpit error on my part?

Well, sort of, if you are expecting cygcheck to be linked with a mingw
zlib DLL.  zlib is linked statically.

I don't know why you aren't seeing a speedup.  It's substantial on my
system.  On my PII 500 at work, the difference is 55 seconds for the new
cygcheck and 1:29 on the old.  That's not as good as my previously
posted figures but it should still be noticeable.  In fact, if you are
running from a console window you can even see the difference in that
the title bar doesn't flicker.

cgf

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: latest cygcheck -c is expensive

[Jason Tishler]

Chris,

On Mon, Sep 08, 2003 at 11:36:21AM -0400, Christopher Faylor wrote:
> On Mon, Sep 08, 2003 at 07:50:16AM -0400, Jason Tishler wrote:
> >BTW, I tried to give cygcheck from the 2003-Sep-08 snapshot a ride.
> >The ChangeLog and diff seem to indicate the required changes have
> >been completed.  Unfortunately, my timing showed no difference and
> >objdump showed no dependency on Mingw zlib:
> >
> >    $ objdump -p cygcheck.exe | fgrep 'DLL Name:'
> >            DLL Name: msvcrt.dll
> >            DLL Name: msvcrt.dll
> >            DLL Name: ADVAPI32.DLL
> >            DLL Name: KERNEL32.dll
> >
> >Is the above cockpit error on my part?
> 
> Well, sort of, if you are expecting cygcheck to be linked with a mingw
> zlib DLL.  zlib is linked statically.

Doh!

> I don't know why you aren't seeing a speedup.  It's substantial on my
> system.  On my PII 500 at work, the difference is 55 seconds for the
> new cygcheck and 1:29 on the old.  That's not as good as my previously
> posted figures but it should still be noticeable.  In fact, if you are
> running from a console window you can even see the difference in that
> the title bar doesn't flicker.

I see a minor speed up (now) -- 1:24 versus 1:18.  Note this is on my
laptop.  Maybe the disk subsystem is so slow as to dominate the
throughput?

Jason

-- 
PGP/GPG Key: http://www.tishler.net/jason/pubkey.asc or key servers
Fingerprint: 7A73 1405 7F2B E669 C19D  8784 1AFD E4CC ECF4 8EF6

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: latest cygcheck -c is expensive

[Christopher Faylor]

On Mon, Sep 08, 2003 at 04:32:04PM -0400, Jason Tishler wrote:
>Chris,
>
>On Mon, Sep 08, 2003 at 11:36:21AM -0400, Christopher Faylor wrote:
>> On Mon, Sep 08, 2003 at 07:50:16AM -0400, Jason Tishler wrote:
>> >BTW, I tried to give cygcheck from the 2003-Sep-08 snapshot a ride.
>> >The ChangeLog and diff seem to indicate the required changes have
>> >been completed.  Unfortunately, my timing showed no difference and
>> >objdump showed no dependency on Mingw zlib:
>> >
>> >    $ objdump -p cygcheck.exe | fgrep 'DLL Name:'
>> >            DLL Name: msvcrt.dll
>> >            DLL Name: msvcrt.dll
>> >            DLL Name: ADVAPI32.DLL
>> >            DLL Name: KERNEL32.dll
>> >
>> >Is the above cockpit error on my part?
>> 
>> Well, sort of, if you are expecting cygcheck to be linked with a mingw
>> zlib DLL.  zlib is linked statically.
>
>Doh!
>
>> I don't know why you aren't seeing a speedup.  It's substantial on my
>> system.  On my PII 500 at work, the difference is 55 seconds for the
>> new cygcheck and 1:29 on the old.  That's not as good as my previously
>> posted figures but it should still be noticeable.  In fact, if you are
>> running from a console window you can even see the difference in that
>> the title bar doesn't flicker.
>
>I see a minor speed up (now) -- 1:24 versus 1:18.  Note this is on my
>laptop.  Maybe the disk subsystem is so slow as to dominate the
>throughput?

In that case, loading gzip.exe once for every package file should have
shown a substantial penalty.

Maybe you have fewer packages loaded than I do?

cgf

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/

Re: latest cygcheck -c is expensive

[Jason Tishler]

On Mon, Sep 08, 2003 at 05:05:22PM -0400, Christopher Faylor wrote:
> On Mon, Sep 08, 2003 at 04:32:04PM -0400, Jason Tishler wrote:
> >I see a minor speed up (now) -- 1:24 versus 1:18.  Note this is on my
> >laptop.  Maybe the disk subsystem is so slow as to dominate the
> >throughput?
> 
> In that case, loading gzip.exe once for every package file should have
> shown a substantial penalty.

Due to NT's file system cache, I don't believe the above to be an issue.
I actually have to "flush the cache" between cygcheck runs, otherwise
subsequent runs are *much* faster:

    $ time cygcheck -c >/dev/null

    real    0m9.683s
    user    0m0.010s
    sys     0m0.020s

Of course, the latest version has eliminated the cost of forking gzip
once for every package file.

> Maybe you have fewer packages loaded than I do?

I don't know.  I have the following:

    $ cygcheck -c | wc -l
        319

Jason

-- 
PGP/GPG Key: http://www.tishler.net/jason/pubkey.asc or key servers
Fingerprint: 7A73 1405 7F2B E669 C19D  8784 1AFD E4CC ECF4 8EF6

--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/