Hi!
For now, the following solution is available:

All URLs previously hardcoded in script check-prereq have been moved to
"master" config.txt, which resides on $HOME3P
(.../trunk/docs/scripts/3rd_party).

Master config.txt has the following syntax, which is the same as in all
other config.txt's on subdirectories:

  variable: value [another_value ...]

First the base URLs for most common mirrors are defined (like
CENTOS_MIRROR, EPEL_MIRROR, ...).

Then come the definitions for packages (always in pairs):

  package_url: base_package_url
  package_pkgs: package1.rpm [package2.rpm ...]

All packages listed in the <package_pkgs> are downloaded from the same
<package_url>.

I've also included new script "check-urls", which parses master
config.txt and tries to download all packages listed there (no data is
really downloaded, just the availability is checked). If any of the URLs
are not valid/found/..., their number is reported at the end and they
are listed in the file not_found.txt.

To reliably check the availability of packages, I had to use & enhance
already supplied wget replacement, written in java. Some of the sites
(like googlecode.com) return 404 Not Found if you check with the HEAD
request and 200 OK for the same URL requested with GET! Other sites
return 503 Service Unavailable if you are producing requests too quick.

The usage of this script is simple - just run it without parameters. To
see what's happening, I suggest logging of the execution like this:

  check-urls 2>&1 | tee check.log

Then see check.log (and possibly not_found.txt) where & what went wrong.
Keywords to search are WARNING: and ERROR:

Matjaz
_______________________________________________
Matterhorn mailing list
[email protected]
http://lists.opencastproject.org/mailman/listinfo/matterhorn


To unsubscribe please email
[email protected]
_______________________________________________

Reply via email to