Hmmm... forgot some:
1. Make sure that no packages that were installed from dotdeb, and from other
back-ported sources are left, and that all the packages are from vanilla
Sarge. [1]
2. Due to the cp -i -f thing, I wanted to run chkrootkit to see if we were
rooted. So I had to install gcc, make and other build tools. I suppose I can
remove it now, but it might be useful to build Perl modules out of CPAN using
dh-perl-make (assuming one cannot find them in Sarge as not all of them are
there).
Do some people here still insist that a compiler present is a security issue?
Regards,
Shlomi Fish
[1] - I had to install them because some PHP applications required them. Feel
free to send assassins to kill me, but make sure to sort out the packages
before or afterwards.
On Friday 05 August 2005 20:36, Shlomi Fish wrote:
> Hi!
>
> <<<
> First of all a note to Lior: are you subscribed to IGLU-Web? If not, please
> subscribe, because it's not a high volume list and we discuss everything
> related to eskimo there. Note that iglu-web blocks posts from
> non-subscribers, to prevent spam. (and because manual moderation proved to
> be too annoying a task, due to that).
>
>
> Here are the tasks that now have to be done:
>
> 1. Perl Modules
> ---------------
>
> Due to the fact the Perl modules in Woody were horribly out of date, I had
> to take the Perl .pm files out of their equivalent CPAN packages, and put
> them in "." with appropriate Perl directives to give precedence to modules
> found there.
>
> Now that we are using Sarge, then the CPAN modules found there, may already
> be adequate for our needs, and we wouldn't need the local ad-hoc
> installation from source. So I'm going to go over them and check that.
>
> 1.1 Investigate the Jobs Database RSS module.
>
> The Jobs data at the moment uses a Perl RSS-generation and parsing module
> (called XML::RSS, IIRC) to generate the RSS. The module that shipped with
> Woody was very old and did not support RSS 2.0 or possibly not even RSS
> 1.0. However, the instructions given to it were that for RSS 2.0. [1]
> (which worked fine on my local machine, which runs an up-to-date version of
> Mandriva).
>
> Some people reported problems with it with rss2email, in which it caused
> the HTML body of the item, to appear as quoted strings.
>
> Now Sarge should have a more up-to-date module from CPAN, which may work
> better. But we need to verify what happens there.
>
> 2. Including the URL of the job entry in the jobs list.
> -------------------------------------------------------
>
> Also regarding the jobs database: there should be a link from each entry to
> its /show-record URL (which displays only it and which can be used to
> quickly identify it or linke to it). It should not be too hard to program
> it in.
>
> 3. Moving the jobs/consultants database to BerliOS or OpenSVN.
> --------------------------------------------------------------
>
> At the moment, the code for the jobs list back-end is version controlled in
> my personal repository on stalker. (a Co-Op server where Nadav Har'El gave
> me an account). It's not bad being there, but it still occupies more and
> more space (as version control repositories go) and I have a limited quota
> there. I'd like to move it to BerliOS or to OpenSVN.
>
> BerliOS is supposed to be a hub for open-source development. At the moment,
> the jobs list is technically open-source, (MIT X11 license) but was not yet
> packaged into a CPAN module, documentation is scarce, and there may be
> other issues that may still label it as "software for internal use". I can
> put it in my web-cpan.berlios.de repositories without too many problems,
> but I may be slightly violating their social contract.
>
> OpenSVN.csie.org as far as I know doesn't care if what you put there is
> open-source or not, but connectivity to it is somewhat slower than BerliOS.
> I also set up a repository for Linux Israeli activities (linuxisrael) and
> we can put the code somewhere there.
>
> At the moment I'm leaning towards BerliOS.
>
> 4. Organizing /iglu
> -------------------
>
> Once upon a time, we had the /iglu volume for all the web-hosted files and
> everyone were happy. Then, came several virtual hosts which were directed
> to eskimo: Hackers-IL, Python-IL, Welcome-to-Linux, wiki.perl.org.il. I
> placed the first three straight under /iglu as /iglu/Hackers-IL/ etc.
>
> Then I decided putting them there together with the rest of the /iglu
> sub-directories was not such a good idea, and created a
> sub-directory /iglu/Hosts for them. However, only wiki.perl.org.il can be
> found there at the moment.
>
> I'd like to invest some spare cycles into moving the rest of the virtual
> hosts into the /Hosts. It shouldn't be too hard - just re-configuring
> apache, doing the move, and reloading apache.
>
> 5. Managing the MediaWikis with ease:
> --------------------------------------
>
> At the moment Eskimo hosts 3 MediaWikis: Python-IL's, Hackers-IL's and
> Perl-IL's. They all contain the entire unpacked distribution where it was
> unpacked, and all require separate upgrades or patch applications when
> these things are needed.
>
> I'd like to ask for any ideas from people here, regarding how one can
> easily centrally manage three such MediaWikis. Is there a script or
> somethings.
>
> At times like this I really hate the fact that PHP does not have modules.
> ;-)
>
> Let me know what you think of all of the above.
>
> Regards,
>
> Shlomi Fish
>
> [1] - The XML::RSS abstraction of the various RSS versions is very
> minimalistic, if not non-existent.
>
>
> ---------------------------------------------------------------------
> Shlomi Fish [EMAIL PROTECTED]
> Homepage: http://www.shlomifish.org/
>
> Tcl is LISP on drugs. Using strings instead of S-expressions for closures
> is Evil with one of those gigantic E's you can find at the beginning of
> paragraphs.
--
---------------------------------------------------------------------
Shlomi Fish [EMAIL PROTECTED]
Homepage: http://www.shlomifish.org/
Tcl is LISP on drugs. Using strings instead of S-expressions for closures
is Evil with one of those gigantic E's you can find at the beginning of
paragraphs.
