[Bug 197198] Review Request: ntop

[bugzilla at redhat.com]

Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug report.

Summary: Review Request: ntop


https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=197198





------- Additional Comments From michael at knox.net.nz  2006-07-07 02:18 EST -------
(In reply to comment #29)
> More issues:
> 
> * plugins loaded twice
> 
> After some look at the code and some testing I have found that all
> the plugins in the plugins directory are loaded, so they are loaded
> twice since they appear twice in the package... In my opinion 
> either the *.so without version number should be removed, or they
> should be moved to *.so without version numbers.

Ok, I will remove the plugins dangling in the /usr/lib directory. 

> * a plugin is not functionnal
> **WARNING** Unable to load plugin /usr/lib/ntop/plugins/libxmldumpPlugin-3.2.so'
> **WARNING** Message is '/usr/lib/ntop/plugins/libxmldumpPlugin-3.2.so: undefined
> symbol: gdome_str_mkref'

hrmm.. good spotting. 

> And there is also another error, but it may only happen on the first run
> **ERROR** LASTSEEN: Unable to open LsWatch database (/var/ntop/LsWatch.db)- the
> plugin will be disabled

Did you test to see if it re-occurs ? The file is present on my system. 

> * many unneeded files in the docs: CONTENTS FILES ntop.txt README.Suse
> INSTALL BUILD-MinGW.txt BUILD-NTOP.txt ntop-autotools.* PORTING 
> RedHat-rpmbuild-HOWTO.txt, and certainly DAG

I will remove the %doc doc/* line

> * should depend on logrotate

Done

> * less rpmbuild warnings with the following line in %setup
> chmod -x docs/ntop-autotools.pdf *.c *.h plugins/*.c plugins/*.xml

Done
 
> * you should remove  /usr/lib/libntop.so, /usr/lib/libmyrrd.so and 
> /usr/lib/libntopreport.so since there are no associated headers
> and also since they use -release, see below.

Removed
 
> * myrrd seems to be an old included version of rrdtools. It shouldn't
> be used but instead linked against the system rrd, unless there is
> a very good reason not to do so.

myrrd is indeed an older rrdtool, however, according to Luca, it has been
modified enough to warrant using it instead of the system provided rrdtool. I
will leave it with myrrd.
 
> * upstream uses the -release for libtool using the package version,
> this is wrong in general, since it trigggers a soname change even
> when the abi don't change, however those libraries are not meant to
> ne linked against, so if the *.so that don't have the release within
> their names are not distributed it is right.

I am not sure I follow you here. Could you clarify? 

> * The mechanisme described under PRIVACY NOTICE should be disabled in 
> the default case (and reenabled with, for example --version-check)
> or completly disabled.

I have disabled the version check in the init script. 

> * ntop doesn't seems to be interruptible by a control-C. It doesn't seems 
> right to me but I may be wrong.
> Maybe for a similar reason, when killing ntop on the first run when ntop 
> asks for a password, the console is broken.

A control-C does shut down ntop here. Its not instant, it performs a graceful
shutdown. 

> This last issue is not blocking the inclusion in extras in my opinion,
> it is more for upstream.

I am working on Ralf's glib concern at the moment, so I wont post the updated
srpm and spec. I have mostly corrected it, however, it still looks for
libglib.so incorrectly when building the xmldump plugin.

Thanks!

-- 
Configure bugmail: https://bugzilla.redhat.com/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug, or are watching the QA contact.