RE: Files appear to be modified even if they are not

[Leeuw, Guus (G.)]

Aldo,

my no 1 suggestion is: Wait for Mitch to come up with his MD5
solution and get rid of the timestamp calculations.

Mitch, you're wanted :-)))

Cheers,
Guus

> -Original Message-
> From: Borrione Aldo [mailto:[EMAIL PROTECTED]]
> Sent: 13 April 2000 18:06
> To: [EMAIL PROTECTED]
> Subject: Files appear to be modified even if they are not
> 
> 
> We have the repository on an HP workstation and access it as a local
>  repository by means of a samba share, which is mounted on the
>  developer PCs (95 or NT). Yes, I understand that pserver 
> would be better...
> 
> On an NT PC, since a few days, all the files in the developer sandbox
> appear in the morning as modified, they show in fact the red icon. 
> 
> Really, nobody has modified either the repository or the local file
> themselves. 
> Other sandboxes sharing the same repository, but located on other PCs,
> do not suffer from this odd behaviour.
> 
> (This is similar to the situation we got when local time 
> changed a few weeks
> ago, 
> but in that case all the the sandboxes on all PCs were 
> showing the same
> behaviour;
> a simple update re-established the correct status of the *apparently*
> modified files.)
> 
> Any suggestions for solving this problem ?
> 
> Thanks in advance
> 
> Aldo Borrione
> 
> 
> Aldo BORRIONE
> Centro Ricerche FIAT
> Direzione Sistemi Elettronici
> DSS-Progettazione Software
> 
> Tel: +39.011.9083.945
> Fax:+39.011.9083.083
> Email:  [EMAIL PROTECTED]
> 
> 
> 
> 

RE: Could timestamps be replaced with MD5?

[Leeuw, Guus (G.)]

Heya Mitch,

Fair enough.

I'ld suggest going ahead because as it stands, CVS is quite often
used in a remote manner where the server is a zillion (or so) miles
away. If you could make CVS to not everytime go thru the Net and
find the server serving it the same code, that'ld be great!

Maybe (for a second phase) we might want to add a flag in the
config file that says whether CVS should calculate with
timestamps or with MD5 (or CRC for that matter).

Oh, and for that matter, I'ld like to be on the Beta test group :-)
Got a fairly large code base here which is far away from my dev-box,
so I'm more than happy if the comms between client and server would
become faster :-)) (I'm using 1.10.8 now).

Just my two cents,
Guus

> -Original Message-
> From: Mitch Davis [mailto:[EMAIL PROTECTED]]
> Sent: 14 April 2000 07:08
> To: [EMAIL PROTECTED]
> Cc: [EMAIL PROTECTED]; [EMAIL PROTECTED]
> Subject: Re: Could timestamps be replaced with MD5?
> 
> 
> Dear Gerhard,
> 
> > Also there is the problem of doing
> > cp -p to copy a file from somewhere else back into the working copy
> 
> Or running ALL your files through a perl or sed script.  Most may be
> unchanged, but ALL will be updated by CVS as it currently stands.
> 
> > Given the size of our code base I'd prefer a CRC over an MD5 
> > (at 1 second per file overhead, an md5 would cost over an 
> hour extra)
> 
> I think the algorithm is not the problem, since md5 can processes
> many many megabytes per second.
> 
> Like all computing time problems, the real impact is only
> apparent when you run it in real life and profile it.  Let's try
> it and see.  While the algorithm is surprisingly fast, it WILL
> require that every file be read into memory, something the
> timestamp approach doesn't require.  So I think the impact will
> not be from the checksum (by whichever method) but from reading
> the files.
> 
> Anyway, the proof of the pudding is in the eating, and I shan't ask
> others to put up with my wonderings until I have some code to
> show for it!! :-)
> 
> Regards,
> 
> Mitch.
> --
> | mailto:[EMAIL PROTECTED]   | Not the official view of: |
> | mailto:[EMAIL PROTECTED] | Australian Calculator Opn |
> | Certified Linux Evangelist! | Hewlett Packard Australia |
> 

Re: Disable *.prog

[Tanaka Akira]

In article <[EMAIL PROTECTED]>,
  Tanaka Akira <[EMAIL PROTECTED]> writes:

> Is there a way to disable Checkin.prog, Update.prog, etc?  I don't
> want to provide general access to a server for commiters.
> 
> At least, commiters can run an arbitrary binary by Checkin.prg and I
> think *.prog is potentially dangerous even in chrooted environment.

I found a description about this in cvs.texinfo:

| Note also that the commit and update programs work ONLY when using local
| repository access -- the files simply aren't created when sources are
| checked out from a pserver or other remote CVS.

But *.prog seems to work with remote repository.  So, I think it is
a bug.  Which is a bug, the document or the source?
-- 
Tanaka Akira

Re: Company mode dev. (one disk server) opposed to Open Source mo de (each has his own disk server) ?

[Mitch Davis]

Gilles-Eric Descamps wrote:
> 
> When you're in company mode, you don't use
> cheap disk on each PC because it becomes
> a mess. You use a network file server
> - with expensive certified fiber disk -
> and a tape backup, because you need to have
> a garantee that your downtime will be less
> than a day per year. With that kind of
> tools, the GB costs you something like $750/GB
> 
> Now, if your company is doing chip design,
> some of the files are 600MB EACH. The whole
> repository is several GB...
> You definitively don't want to duplicate this.
> 
> A commercial version control/problem tracking tool
> like synchronicity offers that kind of cache feature,
> but it's expensive (even more than the file server).

CVS does many jobs extremely well.  One attractive thing
about CVS is that it is very light-weight.  We like it
that way.  It has limitations, but it's simplicity makes
up for it.

CVS does not do the things you ask, for a reason.
If your company has "fiber-channel"-type money, I think trying
to save a few dollars by using CVS is false economy
to the extreme.

As an example, here in HP we use Rational ClearCase.  ClearCase
solves the "wasted diskspace from everyone's unchanged files"
problem very nicely.  It also handles large files and binary
files much better than CVS.  And it handles structural
changes to the repository very well, which is a known
weakness of CVS.

Sure it's hideously expensive, but in this case you're
getting what you pay for.

I am not advocating actually using ClearCase (hell, it's a
constant pain to use, and from an admin pov I have cause to
curse it daily), but I do want to point out that if your
company wants an industrial strength large-scale solution, it
should be prepared to pay for it.

Regards,

Mitch.

PS: Not intended as a flame.  More of a setting-of-expectations.
--
| mailto:[EMAIL PROTECTED]   | Not the official view of: |
| mailto:[EMAIL PROTECTED] | Australian Calculator Opn |
| Certified Linux Evangelist! | Hewlett Packard Australia |

Re: Company mode dev. (one disk server) opposed to Open Source mode (each has his own disk server) ?

[Michael Gersten]

> (I can hear the "but we only check good stuff in" comments coming
> already! :) )

Commit early, commit often, commit to branches!

.cvsignore problem

[Todd A. Lyons]

I too was experiencing the /root/.cvsignore problem and joined this list
specifically to solve it.  But before I could post the question, I found
it in the archives already answered.  Restarted the computer and init
started inet (instead of /etc/rc.d/init.d/inet restart as root) and it
works fine now .

To answer one question that was posed, YES it should be in the FAQ
and/or the FAQ-o-matic.  I too am running on a RedHat 6.x system, so you
might get a few people running the cvs server on other Linux OS's to see
if they experience the same problem.

Rebooting a machine is a Windows solution.  It should not be a Linux
solution because rebooting is for adding new hardware and kernel
upgrades.  Is the more permanent solution to add the following line in
/etc/rc.d/init.d/inet:
HOME=
then have it call 'daemon inetd'?   Or is it more complicated than that?
-- 
Blue skies...   Cannonball  http://www.mrball.net
*   One GUI to rule them all, One GUI to find them* 
*One GUI to bring them all, and with the blue screen bind them*
*   In the land of Redmond, where the shadows lie.*

Re: Company mode dev. (one disk server) opposed to Open Source mode (each has his own disk server) ?

[Derek Scherger]

There's other problems with this idea of cacheing for CVS in general.

Think about :pserver: repositories for a minute. They aren't available
for symbolic links to access, so you pretty much have to copy the files
to the working directories. This might not be an issue in "company 
mode" but it certainly is in "open source mode".

BTW, I agree with your estimation of open source mode verses company
mode but this might be another reason that open source is getting so
popular! $6/GB verses $600/GB is a bit of a no brainer. Certainly your
repository needs to be backed up but maybe your working files don't, so
long as you check stuff in relatively frequently. 

(I can hear the "but we only check good stuff in" comments coming
already! :) )

-- 
Cheers,
Derek
_
Derek Scherger Echologic Software Corporation
mailto:[EMAIL PROTECTED]   http://www.echologic.com

RE: Company mode dev. (one disk server) opposed to Open Source mode (each has his own disk server) ?

[Gilles-Eric Descamps]

  Hi Alan,

  That's exactly what I was trying to say
by "company mode" <-> "open source mode".

When you're in Open source mode, you have
a copy of the sources on your local cheap hard disk.
You don't backup it because you know that
hundreds of people all around the world
also have a copy on their cheap disk,
and only a world class disaster could be
dangerous.

When you're in company mode, you don't use
cheap disk on each PC because it becomes
a mess. You use a network file server
- with expensive certified fiber disk -
and a tape backup, because you need to have
a garantee that your downtime will be less
than a day per year. With that kind of
tools, the GB costs you something like $750/GB

Now, if your company is doing chip design,
some of the files are 600MB EACH. The whole
repository is several GB...
You definitively don't want to duplicate this.

A commercial version control/problem tracking tool
like synchronicity offers that kind of cache feature,
but it's expensive (even more than the file server).

That's why I was looking to have CVS work
in a mode of default read-only check-outs
so that we could enhance it later with a disk cache,
or switch to a different cheap tool.

  Thanks,


> -Original Message-
> From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]
> Sent: Thursday, April 13, 2000 3:55 PM
> To: Gilles-Eric Descamps
> Cc: '[EMAIL PROTECTED]'
> Subject: Re: Company mode dev. (one disk server) opposed to 
> Open Source
> mode ( each has his own disk server) ?
> 
> 
> 
> 
> To my knowledge, CVS does not support your caching idea.  Is 
> this really a
> problem?  Perhaps you could break up your codebase, if a 
> given developer
> required access to only 10% of the 600 MB.  The other answer 
> is to just ignore
> the problem.  I bought a 13.6GB disk for $150 last week, and 
> saw it on sale
> later for $120.  This is about $10/Gb, or about $6 for each 
> developer's working
> copy of the sources.  At these prices, I wouldn't worry about 
> trying too hard to
> find a tricky caching mechanism.
> 
> I know this may not be what you want to hear, but it may be 
> the most economical
> answer.
> 
> Alan Thompson
> 
> 
> 
> 
> 
> 
> Gilles-Eric Descamps <[EMAIL PROTECTED]> 
> on 2000/04/13
> 02:05:53 PM
> 
> To:   "'[EMAIL PROTECTED]'" <[EMAIL PROTECTED]>
> cc:(bcc: Alan Thompson/Orincon)
> Subject:  Company mode dev. (one disk server) opposed to Open 
> Source mode ( each
>   has his own disk server) ?
> 
> 
> 
> Hi all,
> 
> Is there a way to configure CVS to work by default in
> a read-only workspaces with concurrent edit:
> 
> We're a company with forty developers checking out
> files (can be big: 600MB text/binary files).
> All these developers share the same disk server.
> With standard CVS, when they build their workspace,
> they end up having a copy of the files & therefore
> we use almost 40x the space.
> Could CVS be configured with some kind of a cache? :
> By default each file is checked-out as read-only. The data
> of the file goes into some kind of shared cache, and
> the programmer's workspace is only populated with
> links to that file. By that way we would use only 1x the disk space.
> Of course, upon editing one file, the link is suppressed
> and replaced by a copy of the file, and several developers
> can work concurrently on the same file.
> 
> Is there a way to setup some kind of disk cache for CVS ?
> 
> Is there any other version control tool that offers this mechanism
> of disk caching allowing read-only workspaces with concurrent edit ?
> 
> Thanks,
> 
> --
> Gilles-Eric DESCAMPS, Fax (419) 844 7467
> Silicon Access < Enabling the Terabit Internet >
> 2801A Orchard Parkway - San Jose, CA, 95134-2013
> mailto:[EMAIL PROTECTED]
> 
> "Sed quis custodiet ipsos custodes ?"
> 
> 
> 
> 
> 
> 
> 

RE: Linking CVS repositories to other CVS repositories

[Chris Cameron]

On Friday, April 14, 2000 8:10 AM, William K. Gibson 
[SMTP:[EMAIL PROTECTED]] wrote:
>
> I need to know the best way to create links in CVS repositories.
>
> Specifically, can I construct links such that a module can be linked 
within
> another module and be checked out as a subdirectory.
>
> Here is an example of my situation. Say I have two modules, A and B. Both
> modules reside in their own seperate repositories in the CVSROOT. I have
> another module C containing a subdirectory C'. Within C' is source code 
that
> I wish to share amongst A and B. Can I simply create a link to C' in A 
and B
> and expect C' to be checked out as a subdirectory of A or B?
>
> If so, where do I create the link? Do I checkout A, and create a link to
> $CVSROOT/C/C' then add and commit the link? Will this work? Am I off my 
nut
> for wanting to do this?
>
We use the '&' module feature to do this type of thing, along with the -d 
option in the modules file.

If you do something like
_moduleA-d srcA ModuleA
_moduleB-d srcB ModuleB
_moduleC-d srcC ModuleC

moduleA -d moduleA  &_moduleC &_moduleA
moduleB -d moduleB  &_moduleC &_moduleB

a cvs co moduleA will give a directory structure
moduleA/srcA
moduleA/srcC

a cvs co moduleB will give
moduleB/srcB
moduleB/srcC

You cannot checkout two modules into one directory, and you can't create an 
empty directory (e.g. srcC) and then checkout a & module that inserts a 
file into the directory, you need to do things the other way around.  If 
you had another line
_moduleAmakemakefiles/moduleA/makefile
you would have to make the moduleA definition
moduleA -d moduleA  &_moduleAmake &_moduleC &_moduleA
Any other order and the moduleA repository will be set as 'Emptydir' and it 
won't work correctly.

***
Chris CameronOpen Telecommunications NZ Ltd
Software Development Team Leader
[EMAIL PROTECTED]   P.O.Box 10-388
  +64 4 495 8403 (DDI)  The Terrace
fax:  +64 4 495 8419 Wellington
cell: +64 21 650 680New Zealand
Life, don't talk to me about life (Marvin - HHGTTG)

Re: Company mode dev. (one disk server) opposed to Open Source mode ( each has his own disk server) ?

[alant]

To my knowledge, CVS does not support your caching idea.  Is this really a
problem?  Perhaps you could break up your codebase, if a given developer
required access to only 10% of the 600 MB.  The other answer is to just ignore
the problem.  I bought a 13.6GB disk for $150 last week, and saw it on sale
later for $120.  This is about $10/Gb, or about $6 for each developer's working
copy of the sources.  At these prices, I wouldn't worry about trying too hard to
find a tricky caching mechanism.

I know this may not be what you want to hear, but it may be the most economical
answer.

Alan Thompson






Gilles-Eric Descamps <[EMAIL PROTECTED]> on 2000/04/13
02:05:53 PM

To:   "'[EMAIL PROTECTED]'" <[EMAIL PROTECTED]>
cc:(bcc: Alan Thompson/Orincon)
Subject:  Company mode dev. (one disk server) opposed to Open Source mode ( each
  has his own disk server) ?



Hi all,

Is there a way to configure CVS to work by default in
a read-only workspaces with concurrent edit:

We're a company with forty developers checking out
files (can be big: 600MB text/binary files).
All these developers share the same disk server.
With standard CVS, when they build their workspace,
they end up having a copy of the files & therefore
we use almost 40x the space.
Could CVS be configured with some kind of a cache? :
By default each file is checked-out as read-only. The data
of the file goes into some kind of shared cache, and
the programmer's workspace is only populated with
links to that file. By that way we would use only 1x the disk space.
Of course, upon editing one file, the link is suppressed
and replaced by a copy of the file, and several developers
can work concurrently on the same file.

Is there a way to setup some kind of disk cache for CVS ?

Is there any other version control tool that offers this mechanism
of disk caching allowing read-only workspaces with concurrent edit ?

Thanks,

--
Gilles-Eric DESCAMPS, Fax (419) 844 7467
Silicon Access < Enabling the Terabit Internet >
2801A Orchard Parkway - San Jose, CA, 95134-2013
mailto:[EMAIL PROTECTED]

"Sed quis custodiet ipsos custodes ?"

Re: Linking CVS repositories to other CVS repositories

[alant]

Check out the documentation for modules in the Cederqvist manual.  It comes with
each CVS distribution and is named something like "cvs-1.10/doc/cvs.ps".  You
can also get it online from:

 http://www.loria.fr/~molli/cvs-index.html

along with a CVS FAQ.

Alan Thompson





"William K. Gibson" <[EMAIL PROTECTED]> on 2000/04/13 01:09:56 PM

To:   [EMAIL PROTECTED]
cc:(bcc: Alan Thompson/Orincon)
Subject:  Linking CVS repositories to other CVS repositories




I need to know the best way to create links in CVS repositories.

Specifically, can I construct links such that a module can be linked within
another module and be checked out as a subdirectory.

Here is an example of my situation. Say I have two modules, A and B. Both
modules reside in their own seperate repositories in the CVSROOT. I have
another module C containing a subdirectory C'. Within C' is source code that
I wish to share amongst A and B. Can I simply create a link to C' in A and B
and expect C' to be checked out as a subdirectory of A or B?

If so, where do I create the link? Do I checkout A, and create a link to
$CVSROOT/C/C' then add and commit the link? Will this work? Am I off my nut
for wanting to do this?

Thanks in advance!

--William K. Gibson
1stDesk Systems
[EMAIL PROTECTED]

Re: CVS security

[Tobias Weingartner]

On Thursday, April 13, "Brian Huddleston" wrote:
> 
> I think the tunneled pserver approach follows the principle of least
> privelege best.  Although certainly, the pure ssh approach  is easier to
> setup, easier on the clients,  and if you've got a clueful sysadmin that
> stays on top of security maybe having login ports open is not too big a
> deal.  Anyway, that's my $.02.  Take it for what it's worth.

Hmm, I tend to run a quite "minimal" setup for CVS, having only the ssh
port "open" to the outside world.

--Toby.

Re: CVS pserver connection

[Tobias Weingartner]

On Thursday, April 13, Pavel Roskin wrote:
> 
> PS. Nowadays when anonymous CVS is so widespread, it's yet to be seen how
> many security holes are still in CVS. Maybe we should ask OpenBSD guys to
> examine CVS for buffer overflows and other security issues.

Ouch!  Ouch!  I've looked at CVS some, and it is not "easy", nor
particularly "pleasant" to audit.

--Toby.

Company mode dev. (one disk server) opposed to Open Source mode (each has his own disk server) ?

[Gilles-Eric Descamps]

Hi all,
 
Is there a way to configure CVS to work by default in
a read-only workspaces with concurrent edit:
 
We're a company with forty developers checking out
files (can be big: 600MB text/binary files).
All these developers share the same disk server.
With standard CVS, when they build their workspace,
they end up having a copy of the files & therefore
we use almost 40x the space.
Could CVS be configured with some kind of a cache? :
By default each file is checked-out as read-only. The data
of the file goes into some kind of shared cache, and
the programmer's workspace is only populated with
links to that file. By that way we would use only 1x the disk space.
Of course, upon editing one file, the link is suppressed
and replaced by a copy of the file, and several developers
can work concurrently on the same file.
 
Is there a way to setup some kind of disk cache for CVS ?
 
Is there any other version control tool that offers this mechanism
of disk caching allowing read-only workspaces with concurrent edit ?
 
Thanks,

--
Gilles-Eric DESCAMPS, Fax (419) 844 7467
Silicon Access < Enabling the Terabit Internet >
2801A Orchard Parkway - San Jose, CA, 95134-2013
mailto:[EMAIL PROTECTED]
 
"Sed quis custodiet ipsos custodes ?"

 

Re: libxnet

[Noel L Yap]

I'm not so sure if copying over the file is the best way to go since (at least
in my case) libxnet.so was made for Solaris 2.6, not Solaris 2.5 -- you never
know what kind of side effects you may encounter.  I would suggest either
building separate copies of CVS for each platform or using the older-OS CVS on
the newer OS platform (assuming backwards compatibility).

Noel




[EMAIL PROTECTED] on 04/13/2000 03:32:19 PM

To:   [EMAIL PROTECTED]
cc:   (bcc: Noel L Yap)
Subject:  Re: libxnet




never mind...  I discovered that I was looking for the file in my local library
path not on the server...   I had the file, but the server was missing it..   I
had the sys-admin guy copy it over and all works well!!!   Thanks anyways!!!

GaRy

Gary Pinkham wrote:
>
> I just tried to install 1.10.8 and now when I run cvs login I get libxnet.so.1
> not foundI know libxnet.so.1 is there in /lib... I probably have
> something messed up in the config   (I'm not the sys-admin so I had a hell of
a
> time getting this installed..) Any suggestions of what I did wrong???  I
> didn't notice any errors during the compile...
>
> GaRy
>
> --
> Gary Pinkham
> Reasoning, Inc. Voice: 781-359-3132
> One New England Executive Park  Fax:   781-229-2770
> Burlington, MA 01803Email: [EMAIL PROTECTED]

--
Gary Pinkham
Reasoning, Inc. Voice: 781-359-3132
One New England Executive Park  Fax:   781-229-2770
Burlington, MA 01803Email: [EMAIL PROTECTED]

Re: CVS security

[Brian Huddleston]

> On Thursday, April 13, "Brian Huddleston" wrote:
> > > The other possibility is to tunnel the pserver over ssh, which
most
> > > people would consider outright silly, since you can just 'run the
server'
> > > native in ssh mode.
> >
> > The advantage of this approach is that you can set developers up as CVS
> > users as opposed to real users of your system.  While you might trust
> > people to have repository access you might not want to trust them shell
> > access to your machine.
>
> With an "approriate" shell, this is not a problem.  Basically, you can
> take the anoncvssh, and turn it into a slightly less restrictive cvssh,
> and then give each cvs user who you don't wish to have shell access this
> shell...
>
> --Toby.

Ahh...interesting.  I hadn't heard of either one of those.  Of course, if
you just tunnel with pserver,
you don't have to leave port 22 open on your firewall.  If an 3vi1 Hac3r was
able to access your file
system somehow, they've now got a doorway in.

One of our customers had this problem, they hadn't kept up with the security
patches and some script kiddie had managed to add entries to the passwd file
using a BIND exploit.  But, since there weren't any ports open that let them
use their newly created accounts so no harm was done.

I think the tunneled pserver approach follows the principle of least
privelege best.  Although certainly, the pure ssh approach  is easier to
setup, easier on the clients,  and if you've got a clueful sysadmin that
stays on top of security maybe having login ports open is not too big a
deal.  Anyway, that's my $.02.  Take it for what it's worth.

Brian Huddleston
Huddleston Consulting

Re: libxnet

[Noel L Yap]

Does your LD_LIBRARY_PATH set correctly?

Are you using the same binary on the server?  Is the server of an older OS
revision?

Noel




[EMAIL PROTECTED] on 04/13/2000 02:18:14 PM

To:   [EMAIL PROTECTED]
cc:   (bcc: Noel L Yap)
Subject:  libxnet




I just tried to install 1.10.8 and now when I run cvs login I get libxnet.so.1
not foundI know libxnet.so.1 is there in /lib... I probably have
something messed up in the config   (I'm not the sys-admin so I had a hell of a
time getting this installed..) Any suggestions of what I did wrong???  I
didn't notice any errors during the compile...

GaRy

--
Gary Pinkham
Reasoning, Inc. Voice: 781-359-3132
One New England Executive Park  Fax:   781-229-2770
Burlington, MA 01803Email: [EMAIL PROTECTED]

Linking CVS repositories to other CVS repositories

[William K. Gibson]

I need to know the best way to create links in CVS repositories.

Specifically, can I construct links such that a module can be linked within
another module and be checked out as a subdirectory.

Here is an example of my situation. Say I have two modules, A and B. Both
modules reside in their own seperate repositories in the CVSROOT. I have
another module C containing a subdirectory C'. Within C' is source code that
I wish to share amongst A and B. Can I simply create a link to C' in A and B
and expect C' to be checked out as a subdirectory of A or B?

If so, where do I create the link? Do I checkout A, and create a link to
$CVSROOT/C/C' then add and commit the link? Will this work? Am I off my nut
for wanting to do this?

Thanks in advance!

--William K. Gibson
1stDesk Systems
[EMAIL PROTECTED]

Re: CVS pserver connection

[Pavel Roskin]

Hello, Hans!

>   CVS pserver seems to have no timeout for broken connections. After a
>   few time I have some dead cvs-pservers-processes. What can I do?

Perhaps it's time to raise an alert on CERT.
Somebody is misusing a hole that I discovered in "cvs log" a while ago :-(
It's fixed in 1.10.8

Upgrading to cvs-1.10.8 will hopefully solve the problem.

PS. Nowadays when anonymous CVS is so widespread, it's yet to be seen how
many security holes are still in CVS. Maybe we should ask OpenBSD guys to
examine CVS for buffer overflows and other security issues.

Regards,
Pavel Roskin

Re: status feature

[Rex_Jolliff]

cvs co -c will give a list of module names. But you must define the module in
CVSROOT/modules in order for it to show up on this list.

Rex.




Dave Sherohman <[EMAIL PROTECTED]> on 04/13/2000 10:01:10 AM

Please respond to [EMAIL PROTECTED]

To:   Michael Gersten <[EMAIL PROTECTED]>
cc:   [EMAIL PROTECTED], Mark Harrison <[EMAIL PROTECTED]>, Win32 M$
  <[EMAIL PROTECTED]>, [EMAIL PROTECTED] (bcc: Rex Jolliff/YM/RWDOE)

Subject:  Re: status feature




On Thu, Apr 13, 2000 at 11:42:08AM -0500, Michael Gersten wrote:
> List of projects:
> mkdir will-be-big
> cd will-be-big
> cvs get .

*shudder*  I was hoping for a way of just getting all the project _names_,
not all their files...  (I'm a fledgeling CVS admin at a company that's never
used any sort of document control consistently before.  We've got a flock of
~40 small programs/projects and I don't want to have to manually maintain a
list of them all.)

--
The Shortest Windows Manual:  "Turn off the power switch."
Geek Code 3.1:  GCS d- s+: a- C++ UL++$ P+>+++ L++> E- W--(++) N+ o+ !K
w---$ O M- !V PS+ PE Y+ PGP t 5++ X+ R++ tv- b++ DI D G e* h+ r++ y+

Re: libxnet

[Gary Pinkham]

never mind...  I discovered that I was looking for the file in my local library
path not on the server...   I had the file, but the server was missing it..   I
had the sys-admin guy copy it over and all works well!!!   Thanks anyways!!!

GaRy

Gary Pinkham wrote:
> 
> I just tried to install 1.10.8 and now when I run cvs login I get libxnet.so.1
> not foundI know libxnet.so.1 is there in /lib... I probably have
> something messed up in the config   (I'm not the sys-admin so I had a hell of a
> time getting this installed..) Any suggestions of what I did wrong???  I
> didn't notice any errors during the compile...
> 
> GaRy
> 
> --
> Gary Pinkham
> Reasoning, Inc. Voice: 781-359-3132
> One New England Executive Park  Fax:   781-229-2770
> Burlington, MA 01803Email: [EMAIL PROTECTED]

-- 
Gary Pinkham   
Reasoning, Inc. Voice: 781-359-3132
One New England Executive Park  Fax:   781-229-2770
Burlington, MA 01803Email: [EMAIL PROTECTED]

Re: CVS question

[Donald Sharp]

Read the manual.  Specifically section 12 keyword substitution.
Starting on page 73.  The manual should come with your version 
of cvs.

donald
On Thu, Apr 13, 2000 at 11:40:34AM -0600, David A. Hite wrote:
> Hi,
> 
> I'm using CVS and I can't find any info on making CVS automatically
> update my source files. I have seen a few examples but would like a full
> list of available options.
> 
> Here's the examples I've seen:
> 
> // Version: $Revision: $
> // Version Date:$Date: $
> 
> If you can help me or send me to some web pages that have this info I
> would appreciate it.
> 
> Thanks,
> -Dave-
> 

libxnet

[Gary Pinkham]

I just tried to install 1.10.8 and now when I run cvs login I get libxnet.so.1
not foundI know libxnet.so.1 is there in /lib... I probably have
something messed up in the config   (I'm not the sys-admin so I had a hell of a
time getting this installed..) Any suggestions of what I did wrong???  I
didn't notice any errors during the compile...

GaRy

-- 
Gary Pinkham   
Reasoning, Inc. Voice: 781-359-3132
One New England Executive Park  Fax:   781-229-2770
Burlington, MA 01803Email: [EMAIL PROTECTED]

Re: CVS security

[Tobias Weingartner]

On Thursday, April 13, "Brian Huddleston" wrote:
> > The other possibility is to tunnel the pserver over ssh, which most
> > people would consider outright silly, since you can just 'run the server'
> > native in ssh mode.
> 
> The advantage of this approach is that you can set developers up as CVS
> users as opposed to real users of your system.  While you might trust
> people to have repository access you might not want to trust them shell
> access to your machine.

With an "approriate" shell, this is not a problem.  Basically, you can
take the anoncvssh, and turn it into a slightly less restrictive cvssh,
and then give each cvs user who you don't wish to have shell access this
shell...

--Toby.

RE: status feature

[Jerry Nairn]

Hi, Dave,
To list modules:
cvs co -c

But this depends on the contents of CVSROOT/modules

To list the contents of a repository directory:
cvs rdiff -s -r 0 directoryname

or the contents of a module:
cvs rdiff -s -r 0 modulename

To list every file in your entire repository:
cvs rdiff -s -r 0 .

To just dump the entire CVSROOT modules file:
cvs -q co -p CVSROOT/modules

I like to do that sometimes because "cvs co -c" gives only the modules in
sorted order, but "cvs co -p CVSROOT/modules" spits out the modules file
exactly as it is, with your own formatting and comments.

Cheers,
Jerry

> From: Dave Sherohman [mailto:[EMAIL PROTECTED]]
> Sent: Thursday, April 13, 2000 8:49 AM

> 
> So where does one get cvsls?  Searches on Freshmeat and Google come up
> empty...  Or any other (non-web-based) way of getting a list 
> of projects in a
> CVS repository and/or a list of tags on a project?

CVS question

[David A. Hite]

Hi,

I'm using CVS and I can't find any info on making CVS automatically
update my source files. I have seen a few examples but would like a full
list of available options.

Here's the examples I've seen:

// Version: $Revision: $
// Version Date:$Date: $

If you can help me or send me to some web pages that have this info I
would appreciate it.

Thanks,
-Dave-

RE: WinCVS Compatibility : Windows 2000

[Hamid Ghassemi]

Title: WinCVS Compatibility : Windows 2000



We 
are running WinCVS1.1b12 under Windows 2000 and it is working 
fine.
 
-Original Message-From: [EMAIL PROTECTED] 
[mailto:[EMAIL PROTECTED]]Sent: Thursday, April 13, 2000 7:09 
AMTo: [EMAIL PROTECTED]Subject: WinCVS Compatibility : 
Windows 2000
Hi All, 
I was simply curious if anyone has run WinCVS 1.0.6 under 
Windows 2000.  Is that particular build compatible with Win2K?
Thanks! 
 -Jay 

CVS pserver connection

[Hans Fuchs]

Hello,

  CVS pserver seems to have no timeout for broken connections. After a
  few time I have some dead cvs-pservers-processes. What can I do?

  

Gruss,
   Hans

-- 
Gesendet an [EMAIL PROTECTED] am 13.04.2000 mit The Bat 1.38e
Homepage: http://www.enova.ch - EMail: [EMAIL PROTECTED]
Public Key: http://www.enova.ch/gucky/gucky.htm

"Mobile Computing?  Isn't that where you put a long extension cable on your T-1 so you 
can move around?"

Anhang: 

Re: CVS security

[Brian Huddleston]

> The other possibility is to tunnel the pserver over ssh, which most
> people would consider outright silly, since you can just 'run the server'
> native in ssh mode.

The advantage of this approach is that you can set developers up as CVS
users as opposed
to real users of your system.  While you might trust people to have
repository access you might
not want to trust them shell access to your machine.

(Of course, you'll want to restrict permissions on CVSROOT/ or giving them
repository access
is practically the same as giving them a shell account. :-)

Brian Huddleston
Huddleston Consulting

Re: status feature

[Dave Sherohman]

On Thu, Apr 13, 2000 at 11:42:08AM -0500, Michael Gersten wrote:
> List of projects:
> mkdir will-be-big
> cd will-be-big
> cvs get .

*shudder*  I was hoping for a way of just getting all the project _names_,
not all their files...  (I'm a fledgeling CVS admin at a company that's never
used any sort of document control consistently before.  We've got a flock of
~40 small programs/projects and I don't want to have to manually maintain a
list of them all.)

-- 
The Shortest Windows Manual:  "Turn off the power switch."
Geek Code 3.1:  GCS d- s+: a- C++ UL++$ P+>+++ L++> E- W--(++) N+ o+ !K
w---$ O M- !V PS+ PE Y+ PGP t 5++ X+ R++ tv- b++ DI D G e* h+ r++ y+

Re: status feature

[Michael Gersten]

List of projects:
mkdir will-be-big
cd will-be-big
cvs get .

List of tags on a project:
(not debugged yet)
cvs log * |  |  | uniq

I'm sure someone else can come up with a better version of that last one,
doing a log on every file is probably excessive. Would history work better?

Dave Sherohman wrote:
> 
> On Thu, Apr 13, 2000 at 03:06:27PM +0800, Mark Harrison wrote:
> > CVSLS=/aitools/bin/cvsls
> 
> So where does one get cvsls?  Searches on Freshmeat and Google come up
> empty...  Or any other (non-web-based) way of getting a list of projects in a
> CVS repository and/or a list of tags on a project?
> 
> --
> The Shortest Windows Manual:  "Turn off the power switch."
> Geek Code 3.1:  GCS d- s+: a- C++ UL++$ P+>+++ L++> E- W--(++) N+ o+ !K
> w---$ O M- !V PS+ PE Y+ PGP t 5++ X+ R++ tv- b++ DI D G e* h+ r++ y+

Files appear to be modified even if they are not

[Borrione Aldo]

We have the repository on an HP workstation and access it as a local
 repository by means of a samba share, which is mounted on the
 developer PCs (95 or NT). Yes, I understand that pserver would be better...

On an NT PC, since a few days, all the files in the developer sandbox
appear in the morning as modified, they show in fact the red icon. 

Really, nobody has modified either the repository or the local file
themselves. 
Other sandboxes sharing the same repository, but located on other PCs,
do not suffer from this odd behaviour.

(This is similar to the situation we got when local time changed a few weeks
ago, 
but in that case all the the sandboxes on all PCs were showing the same
behaviour;
a simple update re-established the correct status of the *apparently*
modified files.)

Any suggestions for solving this problem ?

Thanks in advance

Aldo Borrione


Aldo BORRIONE
Centro Ricerche FIAT
Direzione Sistemi Elettronici
DSS-Progettazione Software

Tel: +39.011.9083.945
Fax:+39.011.9083.083
Email:  [EMAIL PROTECTED]

Re: status feature

[Dave Sherohman]

On Thu, Apr 13, 2000 at 03:06:27PM +0800, Mark Harrison wrote:
> CVSLS=/aitools/bin/cvsls

So where does one get cvsls?  Searches on Freshmeat and Google come up
empty...  Or any other (non-web-based) way of getting a list of projects in a
CVS repository and/or a list of tags on a project?

-- 
The Shortest Windows Manual:  "Turn off the power switch."
Geek Code 3.1:  GCS d- s+: a- C++ UL++$ P+>+++ L++> E- W--(++) N+ o+ !K
w---$ O M- !V PS+ PE Y+ PGP t 5++ X+ R++ tv- b++ DI D G e* h+ r++ y+

RE: Windows interface for CVS

[Leeuw, Guus (G.)]

Look for winCVS thru http://www.sourcegear.com

> -Original Message-
> From: Antonio Sindona [mailto:[EMAIL PROTECTED]]
> Sent: 13 April 2000 14:09
> To: [EMAIL PROTECTED]
> Subject: Windows interface for CVS
> 
> 
> Hi all,
> I'm just subscribed to this mailing list and i'd like to know 
> if there is a
> Windows (sorry ;-) ) interface for CVS. I've found a windows 
> client, but it
> works from command line. What i'm looking for is the 
> equivalent of Visual
> Source Safe, i.e. something  which could be used directly 
> inside VB or VC++.
> 
> Thanks in advance
> Antonio
> 
> -
> Italtel S.p.a.
> PA-SI-PROG
> Dr. Antonio Sindona
> Tel. 091/8615684 (uff.)
> Tel. 0347/6160212 (cell.)
> Fax. 091/8615445
> EMail [EMAIL PROTECTED] (Ufficio)
>   [EMAIL PROTECTED] (Privata)
> Internet Mailbox Omnitel: [EMAIL PROTECTED]
> 

WinCVS Compatibility : Windows 2000

[jay . millar]

Title: WinCVS Compatibility : Windows 2000





Hi All,


I was simply curious if anyone has run WinCVS 1.0.6 under Windows 2000.  Is that particular build compatible with Win2K?

Thanks!


 -Jay

Re: Could timestamps be replaced with MD5?

[Mitch Davis]

Dear Gerhard,

> Also there is the problem of doing
> cp -p to copy a file from somewhere else back into the working copy

Or running ALL your files through a perl or sed script.  Most may be
unchanged, but ALL will be updated by CVS as it currently stands.

> Given the size of our code base I'd prefer a CRC over an MD5 
> (at 1 second per file overhead, an md5 would cost over an hour extra)

I think the algorithm is not the problem, since md5 can processes
many many megabytes per second.

Like all computing time problems, the real impact is only
apparent when you run it in real life and profile it.  Let's try
it and see.  While the algorithm is surprisingly fast, it WILL
require that every file be read into memory, something the
timestamp approach doesn't require.  So I think the impact will
not be from the checksum (by whichever method) but from reading
the files.

Anyway, the proof of the pudding is in the eating, and I shan't ask
others to put up with my wonderings until I have some code to
show for it!! :-)

Regards,

Mitch.
--
| mailto:[EMAIL PROTECTED]   | Not the official view of: |
| mailto:[EMAIL PROTECTED] | Australian Calculator Opn |
| Certified Linux Evangelist! | Hewlett Packard Australia |

Re: Windows interface for CVS

[Donald Sharp]

www.wincvs.org

donald
On Thu, Apr 13, 2000 at 02:08:33PM +0200, Antonio Sindona wrote:
> Hi all,
> I'm just subscribed to this mailing list and i'd like to know if there is a
> Windows (sorry ;-) ) interface for CVS. I've found a windows client, but it
> works from command line. What i'm looking for is the equivalent of Visual
> Source Safe, i.e. something  which could be used directly inside VB or VC++.
> 
> Thanks in advance
> Antonio
> 
> -
> Italtel S.p.a.
> PA-SI-PROG
> Dr. Antonio Sindona
> Tel. 091/8615684 (uff.)
> Tel. 0347/6160212 (cell.)
> Fax. 091/8615445
> EMail [EMAIL PROTECTED] (Ufficio)
>   [EMAIL PROTECTED] (Privata)
> Internet Mailbox Omnitel: [EMAIL PROTECTED]
> 

Re: Windows interface for CVS

[Gary Pinkham]

try 


How to integrate cvc into MS Developer Studio (Visual C++)

You can integrate cvc into the MS Developer Studio (Visual C++) and use cvc
commands from inside programming environment to update or commit changes
made. 

URL: http://depc14.gsi.de/Hades/cvc-msdev.htm 


Antonio Sindona wrote:
> 
> Hi all,
> I'm just subscribed to this mailing list and i'd like to know if there is a
> Windows (sorry ;-) ) interface for CVS. I've found a windows client, but it
> works from command line. What i'm looking for is the equivalent of Visual
> Source Safe, i.e. something  which could be used directly inside VB or VC++.
> 
> Thanks in advance
> Antonio
> 
> -
> Italtel S.p.a.
> PA-SI-PROG
> Dr. Antonio Sindona
> Tel. 091/8615684 (uff.)
> Tel. 0347/6160212 (cell.)
> Fax. 091/8615445
> EMail [EMAIL PROTECTED] (Ufficio)
>   [EMAIL PROTECTED] (Privata)
> Internet Mailbox Omnitel: [EMAIL PROTECTED]

-- 
Gary Pinkham   
Reasoning, Inc. Voice: 781-359-3132
One New England Executive Park  Fax:   781-229-2770
Burlington, MA 01803Email: [EMAIL PROTECTED]

Re: Test: can anyone see this

[William K. Gibson]

on 4/12/00 2:34 PM, William K. Gibson at [EMAIL PROTECTED] wrote:

> I sent a message earlier and it did not show up in this list. Can anyone see
> this message?

Ok, Ok, after recieving about ten replies I now know that my messages are
seen. Hooray! I'm Somebody! :-)

For some reason my first post did not get emailed directly to me- although
the second one did. Just a glitch I guess...

Anyway, I guess that means my intial cvs question about directories with
spaces has no answer :-(

Thanks to all.

--William K. Gibson
1stDesk Systems
[EMAIL PROTECTED]

How to add a new module to the branch

[Jake Colman]

I added and committed a new module to the HEAD and now need to add the same
module to the branch.  I don't need to add it but how do I commit so that the
branch sees it too?

Thanks!

-- 
Jake Colman 

Principia Partners LLC  Phone: (201) 946-0300
Harborside Financial Center   Fax: (201) 946-0320
902 Plaza II   Beeper: (800) 928-4640
Jersey City, NJ 07311  E-mail: [EMAIL PROTECTED]
   E-mail: [EMAIL PROTECTED]
  web: http://www.ppllc.com

microsoft: "where do you want to go today?"
linux: "where do you want to go tomorrow?"
BSD:   "are you guys coming, or what?"

Windows interface for CVS

[Antonio Sindona]

Hi all,
I'm just subscribed to this mailing list and i'd like to know if there is a
Windows (sorry ;-) ) interface for CVS. I've found a windows client, but it
works from command line. What i'm looking for is the equivalent of Visual
Source Safe, i.e. something  which could be used directly inside VB or VC++.

Thanks in advance
Antonio

-
Italtel S.p.a.
PA-SI-PROG
Dr. Antonio Sindona
Tel. 091/8615684 (uff.)
Tel. 0347/6160212 (cell.)
Fax. 091/8615445
EMail [EMAIL PROTECTED] (Ufficio)
  [EMAIL PROTECTED] (Privata)
Internet Mailbox Omnitel: [EMAIL PROTECTED]

Re: Bug in rcs.c ?

[Win32 M$]

Hi David, Jesper,

Do you think it is OK that somebody checks in the changes without any 
revision comments?

Well, I think that every checkin should have one. If your developers can not 
come up with short description of what are the changes done to the file 
doing, maybe he/she should not check it it at all, since they apparently 
don't know what they doing...?

I also see that WinCvs is allowing the same 'solution' to the problem by 
allowing the empty message to be entered on commit and replacing it with 'no 
  message' when issue the commit. Alex, sorry to say that but I don't like 
it :( In the long run you will notice that is makes more troubles than it 
solves. For example try to create change log from such a checkins...

Just my two cents :)
Take care.
Jerzy

>Jesper,
>
>>During the move from my MKS/SI repository to
>>CVS I found this thing:
>>
>>In src/rcs.c line 4068 (CVS-1.10.8) there is a
>>strncmp which fails, due to 'log' is NULL.
>>
>>I added a check for log != NULL and the program
>>run fine again.
>>
>>Is this a bug or has something in the conversion
>>gone wrong ?
>>
>>Cheers,
>>
>>Jesper Pedersen
>
>This could be considered a bug, but I think it's really more of a
>"functional characteristic" of RCS.  RCS does not allow empty revision
>comments in non-interactive check-ins.  I discovered this also a few years
>ago when writing a PVCS to RCS conversion script.  This is the way I dealt
>with in (in perl script form):
>
># If an empty comment is specified, RCS will not check in the file;
># check for this case.  (but an empty -t- description is fine - go figure!)
># Since RCS will pause and ask for a comment if one is not given,
># substitute a dummy comment "no comment".
>if ( $comment[$i] eq "\"\"" )
>{
>  $ci_command =
>"ci -f -r$rcs_rev_num[$i] -d$checked_in[$i] -w$author[$i]
>-t-$description -m\"no comment\" $workfile";
>}
>else
>{
>  $ci_command =
>"ci -f -r$rcs_rev_num[$i] -d$checked_in[$i] -w$author[$i]
>-t-$description -m$comment[$i] $workfile";
>}
>
>
>Regards,
>David
>
>
>

BR,
Jerzy

The first thing they don't teach you at school: "Never say never".
All the issues not related to the list please send to me in private, thanks.

__
Get Your Private, Free Email at http://www.hotmail.com

Re: Could timestamps be replaced with MD5?

[Gerhard den Hollander]

fROm Sean Cavanaugh (on Wed, Apr 12, 2000 at 09:31:51PM -0700):
> I'd like to see both here, if the timestamps are exactly the same assume
> its the same,

Ehm,
that's not always true (as I've found out the hardway)
touch-ing and retouching (touch woth a datestamp) does happen in makefiles,
and hile The GooD Thing(tm) would probably be to fix the makefiles to
nolonger rely on (re)touching it would be much more convenient to have cvs
not rely on the timestamps but on Something Else(tm).

Also there is the problem of doing
cp -p to copy a file from somewhere else back into the working copy

Whether the Something Else is a CRC check or an MD5 sum.

Given the size of our code base I'd prefer a CRC over an MD5 
(at 1 second per file overhead, an md5 would cost over an hour extra)

> otherwise take the time to open the local file and run it
> through MD5 and compare it to the checksum in the local CVS/Entries to make
> sure.  We might need an option to force one of three possible modes
> (timestamp only, MD5 only, and hybrid)

timestamp only, MD5 only, crc-only, MD5iftimestampdiffers,
crciftimestampdiffers ?

How does that sound ?

Kind regards,
 --
Gerhard den Hollander   Phone +31-10.280.1515
Technical Support Jason Geosystems BV   Fax   +31-10.280.1511
Adres:  POBox 1573
3000 BN Rotterdam
[EMAIL PROTECTED]  The Netherlands
visit us at http://www.jasongeo.com
"TeamWare" for finding Oil & Gas...The Smart Way

Re: status feature

[Mark Harrison]

>>How often do you type the following cvs commands?
>>
>>cvs status | grep Status:
>>cvs status | grep Status: | grep -v Up-to-date

Here's the solution that I use.  It gives the appearance
of adding 3 commands:

cvs qstatus   quick status (one line per file)
cvs oodateout of date files
cvs lsdirectory listing of repository

FWIW,
Mark.


#!/bin/sh
# cvs -- add a few commands to cvs, the hacky way
# feedback to: [EMAIL PROTECTED]

CVS=/aitools/bin/cvs
CVSLS=/aitools/bin/cvsls

if test "$1" = "qstatus"; then
$CVS status | egrep '^(\?|File:)'
elif test "$1" = "oodate"; then
($CVS status | egrep "^(File:)|([?])" | grep -v Up-to-date) 2>&1
|grep -v '^cvs server: Examini'
elif test "$1" = "ls"; then
shift
$CVSLS $*
else
$CVS "$@"
fi



--
Mark Harrison http://usai.asiainfo.com:8080/
AsiaInfo Computer Networks[EMAIL PROTECTED]
Beijing, China / Santa Clara, CA  http://www.markharrison.net/

Re: Double byte characters and CVS

[Mark Harrison]

Sara Koehler <[EMAIL PROTECTED]> wrote:
> Is there a CVS version for the Japanese language? If not, would checking
>all japanese text in as binary work? I couldn't find anything in the manual
>for double byte characters of the Japanese language.


Since CVS is 8-bit clean, it will work in normal text mode.  The japanese
double-byte characters are just treated as two hi-bit characters.
We use it for chinese text files.

HTH,
Mark.

--
Mark Harrison http://usai.asiainfo.com:8080/
AsiaInfo Computer Networks[EMAIL PROTECTED]
Beijing, China / Santa Clara, CA  http://www.markharrison.net/