Faq-o-matic

Faq-o-matic

[Jonathan Larmour]

It appears that faq-o-matic is broken on the new system.

See: http://sources.redhat.com/fom/ecos or http://sources.redhat.com/fom/cgen/

I have absolutely no idea why from the error message, but I suspect it may 
be something as "simple" (ahem) as a newer version of perl on the new 
system breaking it - current perl is v5.6.1 but I don't know what it was 
on the old system... can someone knowledgeable have a look please?

Although it might be a simple cause, it probably isn't a simple fix :-|.

Thanks in advance,

Jifl
-- 
eCosCentric       http://www.eCosCentric.com/       <info@eCosCentric.com>
--[ "You can complain because roses have thorns, or you ]--
--[  can rejoice because thorns have roses." -Lincoln   ]-- Opinions==mine

Re: Faq-o-matic

[Jason Molenda]

On Wed, Jan 22, 2003 at 04:36:08AM +0000, Jonathan Larmour wrote:
> It appears that faq-o-matic is broken on the new system.
> 
> See: http://sources.redhat.com/fom/ecos or http://sources.redhat.com/fom/cgen/


I foolishly waded into this thinking I could somehow make it better.
I've never looked at FOM before.  On the up side, I reinstalled
fom and it's correctly working.

On the down side, I need to do some initial configuration before
the databases will work again.  I was given a secret of
3aff8b1ea7196576806c88113c3ee397 at install-time (it said it was
a temporary secret so hopefully it's not a big deal that I'm blasting
it across the known universe).

The total list of FOM databases on the system are

  automake
  binutils
  cgen
  ecos
  guile
  sid
  gcc

I believe only cgen, ecos, sid, and gcc have content in them.  I also
believe all of them are off-line right now, waiting for some kind of
initialization procedure to finish the installation.

I'll resort to drastic measures (reading the install documentation :-)
and see if I can't understand how this thing is supposed to be installed
before I go to sleep.


Jason
(I knew I'd some day pay for ignoring FOM all these years)

Status of FOM as I head off to bed

[Jason Molenda]

Still no joy.

I had to upgrade from 2.709 to 2.717 for it to work correctly with
perl 5.6.1.  Each of the existing databases --

  cgen
  sid
  ecos
  gcc

is currently inaccessible until the maintainer of each goes to some
"install" page.  I've asked jifl, the maintainer of the ecos
database, to see if he can get to that install page.  If that
doesn't work for him, I'll see about forcing new admin passwords
or whatever is necessary.

FOM stores hashes of each user's password in one of its config files.
In 2.709 these hashes were crypt() strings.  Now they're 128-bit
md5 hashes.  I _suspect_ that FOM will update the passwords as users
log in, but I don't know that for certain.  (I looked over all the
release notes foreach release and found no mention of this problem)

If anyone logs in as an admin, keep this caveat from one of the
release notes in mind for that first install-login:

  WHEN YOU UPGRADE TO 2.714, YOU MUST BE CERTAIN TO PASS THROUGH
  THE "Define Configuration Parameters" PAGE.

  You shouldn't have to actually define any parameters; the defaults
  are fine if you use GNU RCS. But you must pass through that page
  to get the args defined; otherwise RCS will get called with empty
  options and will end up making a mess of things. (A reversible
  mess, but a mess nonetheless.)

  If, when you upgrade, you consistently follow the directions (go
  to the installer page and check off the newly-unchecked boxes),
  you'll be fine. I discovered this problem because I cheated and
  just checked the "Mark the config file as upgraded" box, to make
  the "admin must run installer" message go away. :v)

Re: Status of FOM as I head off to bed

[Joseph S. Myers]

On Wed, 22 Jan 2003, Jason Molenda wrote:

>   gcc

The GCC FOM is dead; if someone fixes the remaining stray link to it in
the libstdc++ FAQ, we could put a tarball of the RCS files somewhere just
in case and remove that database.

-- 
Joseph S. Myers
jsm28@cam.ac.uk

Re: Status of FOM as I head off to bed

[Jonathan Larmour]

Jason Molenda wrote:
> I've asked jifl, the maintainer of the ecos
> database, to see if he can get to that install page.  If that
> doesn't work for him, I'll see about forcing new admin passwords
> or whatever is necessary.

Sorry, was asleep :-). Yes I can get to the install page - the old 
login/password worked fine, and after I went through all the checklist on 
the install page (not just the define configuration parameters appears) it 
is all happy. Thanks for working on this Jason!

> FOM stores hashes of each user's password in one of its config files.
> In 2.709 these hashes were crypt() strings.  Now they're 128-bit
> md5 hashes.  I _suspect_ that FOM will update the passwords as users
> log in, but I don't know that for certain.  (I looked over all the
> release notes foreach release and found no mention of this problem)
 >
> If anyone logs in as an admin, keep this caveat from one of the
> release notes in mind for that first install-login:
> 
>   WHEN YOU UPGRADE TO 2.714, YOU MUST BE CERTAIN TO PASS THROUGH
>   THE "Define Configuration Parameters" PAGE.
> 
>   You shouldn't have to actually define any parameters; the defaults
>   are fine if you use GNU RCS. But you must pass through that page
>   to get the args defined; otherwise RCS will get called with empty
>   options and will end up making a mess of things. (A reversible
>   mess, but a mess nonetheless.)

I don't know if it made a mess because the faq was online before I went 
through the configuration parameters. Probably not - at the very least it 
looks okay from the outside regardless.

>   If, when you upgrade, you consistently follow the directions (go
>   to the installer page and check off the newly-unchecked boxes),
>   you'll be fine. I discovered this problem because I cheated and
>   just checked the "Mark the config file as upgraded" box, to make
>   the "admin must run installer" message go away. :v)

I got 100 or so of those :-). At the very least it's stopped sending them now!

Thanks again for getting this going again Crash. You're a star.

Jifl
-- 
eCosCentric       http://www.eCosCentric.com/       <info@eCosCentric.com>
--[ "You can complain because roses have thorns, or you ]--
--[  can rejoice because thorns have roses." -Lincoln   ]-- Opinions==mine

Re: Status of FOM as I head off to bed

[Frank Ch. Eigler]

[-- Attachment #1: Type: text/plain, Size: 664 bytes --]

Hi -

On Wed, Jan 22, 2003 at 01:10:30AM -0800, Jason Molenda wrote:
> I had to upgrade from 2.709 to 2.717 for it to work correctly with
> perl 5.6.1.  Each of the existing databases --
> 
>   cgen
>   sid
>   ecos
>   gcc
> 
> is currently inaccessible until the maintainer of each goes to some
> "install" page.  [...]

I did the deed for cgen & sid; worked fine.
While I was in the neighbourhood, I installed an old GD.pm perl module,
along with one sucky hand-made change to force PNG images within
</usr/lib/perl5/site_perl/5.6.1/FAQ/OMatic/statgraph.pm>.
The effect is that those little fom access-statistics graphs now work.


- FChE

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]