RPM changes for 3.2rc1

I'm attaching a patch for the Red Hat spec file. The top 3 %define 
variables hold the version string components and are designed to be set 
before creating the tarball. (This is typically done by keeping a .spec.in 
file and editing it with sed or Perl in the "make tar" section.)

Could someone with Mandrake and SUSE test this to see if it can be used 
unchanged on those distros? I don't think there's anything RH-specific in 
it. I just used it to package on Fedora Core 2.

_______________________________________________
Ntop-dev mailing list
Ntop-dev <at> unipi.it
http://listgateway.unipi.it/mailman/listinfo/ntop-dev

RE: RPM changes for 3.2rc1

Timing is perfect - I'm about to cut rc2

But:

  1. Use -U3 on your diff command).  Patches w/o context are useless (-U3
would put some context on the initial addition of lines, w/o U3 it will just
blindly do whatever the line #s tell it to do, whether it makes sense or
not...)

  2. Unfortunately 7 of 10 chunks fail on the spec file for 3.2rc1...  But I
think I get the drift.  In fact, looking at the patch the only thing I
didn't already have in place was some of the constant tricks.

As to SuSE, I have SuSE VMs - the old .spec worked fine, so I doubt this
will have problems.  Mandriva - we'll I'm out of room to dedicate to VMs.

-----Burton

-----Original Message-----
From: ntop-dev-bounces <at> unipi.it [mailto:ntop-dev-bounces <at> unipi.it] On Behalf
Of Kenneth Porter
Sent: Saturday, October 01, 2005 1:13 PM
To: ntop-dev <at> Unipi.IT
Subject: [Ntop-dev] RPM changes for 3.2rc1

I'm attaching a patch for the Red Hat spec file. The top 3 %define variables
hold the version string components and are designed to be set before
creating the tarball. (This is typically done by keeping a .spec.in file and
editing it with sed or Perl in the "make tar" section.)

RE: RPM changes for 3.2rc1

--On Saturday, October 01, 2005 5:20 PM -0500 Burton Strauss 
<Burton <at> ntopSupport.com> wrote:

>   1. Use -U3 on your diff command).  Patches w/o context are useless (-U3
> would put some context on the initial addition of lines, w/o U3 it will
> just blindly do whatever the line #s tell it to do, whether it makes
> sense or not...)

Doh, sorry about that. I knew I'd forget at least one thing in posting the 
patch. ;) Let me know if it would still be useful, as I still have the 
source files.

RE: RPM changes for 3.2rc1

I think not - as we talked about off list, most of your ideas were already
incorporated in the file I used for 3.2rc1 from your earlier .diffs (I just
didn't get around to committing the version I actually used to the cvs).

But it's a general good idea for anyone who wants to contribute patches to
use -U3.  The patch Kenneth sent had lots of lines like this:

5,7c11,13
< Version: 3.1
< Release: 0
< Source: ntop-3.1.tgz
---
> Version: %{rpmversion}
> Release: %{rpmrelease}
> Source: ntop-%{version}%{versuffix}.tgz

Which work fine - there's an was/is set for patch to figure out.  But pure
inserts:

2a3,8
> # next 3 variables should be changed at tarball-creation time
> %define rpmversion 3.2
> %define versuffix rc1
> # next starts with 0 for prereleases and betas, 1 for final
> %define rpmrelease 0.%{versuffix}
>

Just have line #s and will apply regardless of whether that makes sense!

-----Burton

New ntop commit (author deri)

Update of /export/home/ntop/ntop/docs
In directory unknown:/tmp/cvs-serv15792

Modified Files:
	FAQ 
Log Message:
Added FAQ contributions

ntop 3.2 roadmap

Hi all,
I see that there are some ongoing discussions on the mailing list about
pre-3.2. I understand that some of your still have some minor issues
with ntop, but we cannot postpone indefinitively the release of 3.2. The
latest stable release 3.1 is almost one year old, and it's really time
to move ahead. In fact we cannot postpone indefinitively new features
that cannot be introduced at this stage as 3.2 is very close. Nor major
patches cannot be accepted, for the very same reason.

All I can tell you is that I would like to see 3.2 out next week. Any
comments? Burton?

Cheers, Luca

--

-- 
Luca Deri <deri <at> ntop.org>	http://luca.ntop.org/
				skype://lucaderi
Hacker: someone who loves to program and enjoys being
clever about it - Richard Stallman

RE: ntop 3.2 roadmap

Well, I take issue with the basic premise, that ntop should be released
because a period of time has elapsed.  However, we've both stopped
committing functional enhancements and focused on stability and cleanup for
the last month or two.  Which is MY definition of ready to release.

There is really only one thing I would like to change, which would be to
alter the point at which we drop root privileges.  But reworking that code
is a significant disruption, so I'm willing to postpone it.

This is not an issue where we have made ntop less secure 3.2 will be better
than 3.1 and that was better than 3.0, etc.!

  We remain committed to fixing any security issues that are discovered with
ntop.

  I am not presently aware of any security issues with ntop.

  We also suggest that people do not publicly expose ntop instances -
there's simply no reason to run even that slight risk.

That said, I will be tagging the cvs as 3.2rc2 and building .tgz and .rpm
files for upload to SourceForge today.

I suggest people test with 3.2rc2.  Baring any major problems, I plan to
release 3.2 during the week of 10-14 Oct 2005.

Based on VMs that I have built, I should be able to create additional .rpm
files this time vs. 3.2rc1.

Action items:

Fw: Core dump in 3.2rc1 netflow handling

Does anyone have any more ideas on the ntop crashes?  I can't run the
program for more than a few minutes with the latest CVS version.

Thanks,
Beirne

----- Forwarded by Beirne Konarski/CORP/FXG on 10/03/2005 09:49 AM -----

             beirne.konarski <at> f                                             
             edex.com                                                      
             Sent by:                                                   To 
             ntop-dev-bounces <at>          ntop-dev <at> unipi.it                   
             unipi.it                                                   cc 

                                                                   Subject 
             09/13/2005 11:13          RE: [Ntop-dev] Core dump in 3.2rc1  
             AM                        netflow handling                    

                                                                           
             Please respond to                                             
             ntop-dev <at> unipi.it                                             

It is a 4-processor Compaq/HP DL380 with HT in its Xeon processors.

             "Burton Strauss"
             <Burton <at> ntopSuppo
             rt.com>                                                    To
             Sent by:                  <ntop-dev <at> unipi.it>
             ntop-dev-bounces <at>                                           cc
             unipi.it

Re: ntop 3.2 roadmap

Burton Strauss wrote:

>Well, I take issue with the basic premise, that ntop should be released
>because a period of time has elapsed.  However, we've both stopped
>committing functional enhancements and focused on stability and cleanup for
>the last month or two.  Which is MY definition of ready to release.
>
>There is really only one thing I would like to change, which would be to
>alter the point at which we drop root privileges.  But reworking that code
>is a significant disruption, so I'm willing to postpone it.
>
>This is not an issue where we have made ntop less secure 3.2 will be better
>than 3.1 and that was better than 3.0, etc.!
>
>  We remain committed to fixing any security issues that are discovered with
>ntop.
>
>  I am not presently aware of any security issues with ntop.
>
>  We also suggest that people do not publicly expose ntop instances -
>there's simply no reason to run even that slight risk.
>
>
>That said, I will be tagging the cvs as 3.2rc2 and building .tgz and .rpm
>files for upload to SourceForge today.
>
>I suggest people test with 3.2rc2.  Baring any major problems, I plan to
>release 3.2 during the week of 10-14 Oct 2005.
>
>

Re: ntop 3.2 roadmap

Luca,
ntop compiled on MinGW crashes when I try to generate
an arbitrary graph. Is it OK on .NET? Regards,

Georger

--- Luca Deri <deri <at> ntop.org> escreveu:
...
> I would like to avoid this now. The code works
> perfectly under .NET and
> I don't want to break it for MinGW. We can postpone
> it to 3.2.1

	

	
		
_______________________________________________________ 
Novo Yahoo! Messenger com voz: ligações, Yahoo! Avatars, novos emoticons e muito mais. Instale agora! 
www.yahoo.com.br/messenger/