Revisions for trean/scripts/check_links.php

Happy New Year

Happy New Year

Link checker works now, though favicons still need to be re-done.

Assign copyright to The Horde Project for those copyright holders that already agreed to it.

Trean_Bookmarks does not use a singleton now.

[cjh] Move bookmarks from DataTree storage to a SQL table.

Probably still a few kinks to work out, but it works.

style nit

It took me until today, when I was tracking down the usage and meaning
of all attributes used by bookmarks, that I found the feature to save
a bookmark's redirected URL. Great idea but given that I never knew it
was there... I've instead included (commented out for now) code in
check_links.php to go ahead and update the actual url of the bookmark.

I can't find anything that http-status-count was used for.

http-Content-Type is a nice idea, but trim it in the interest of
reducing the number of fields bookmarks use.

Take advantage of share caching in Trean.
Maybe we want a separate configuration setting to turn this off or on in
Trean, in that case we need to check inside Trean_Bookmarks::singleton() if we
have a sufficiently new Horde_Share version.

Consistently use /usr/bin/php in script shebangs.

Happy new year\!

Fix string generation.

Use file_get_contents().

Fix a few errors/warnings and always use _getHeaders, since we can't
control the timeout on the builtin function. Also lower the timeout to
10 seconds; this script takes _forever_ to run for me, and I figure 10
seconds is fine as a quick check for whether a URL is there or not and
for trying to get a favicon.

give each url 30 seconds to respond.  Continue with the next url if it
doesn't

Don't send errors from get_headers() calls.

array_shift() needs a variable.

space

We probably want to check for favicons for more status codes than 200 and
302, but 302 was common for my bookmarks.

Fix favicon checking for bookmarks without paths.

whitespace

for some reason the 'favicon' attribute wasn't always being set.  This
fixes it

* Try to get favicon.ico if a favicon could not be parsed
* Store Content-Type header and try to fall back on a mime icon

Get favicons with cron job

fix "Undefined Variable" error

silence parse_url call, problems are caught later anyway.

check_links cron script.  Stores http status code, and if necessary, the
'Location:' header for redirects.  I've done a little with the UI to show
this, but it still needs polishing.