Use isset() to check for $base value and not NO_BASE...
Now the search script should work without the base if on the same site.
This makes URLs shorter...
this way now (as $MYSITE also ends in /).
Apply urlencode to URL parts needed encoding in prev and next links
Adding a security check for show, so there is no need to escape it anywhere
Also add security check for lang and move language detection to up (will be
replaced with langchooser.inc functionality).
Apply $base modification to new search link too...
updated automatically from local information. This means:
- There is no need to contact mirrors@php.net to activate local search
anymore.
- Unofficial mirror sites can also support searching without any tweaks.
Made base check more intelligent, so there is no need to list all special
.php.net sites. Check for mirror pattern in the name. There is only one
exception here "qa.php.net" which looks like a mirror ;)
More secure page number handling and exclude handling
Urlencode != htmlspecialchars, so use urlencode for URL parameters on
"no documents found" page
Use preg instead of ereg everywhere for consistency
Add language information to manual search dropdown
index db of php.net. See bug #20870 for more info... The index
should be updated, but it does not seem to happen, or bogus
entrys keep staying there...
(is in place, and PHP can excute it).
Still commented, so my mails sent out recently will still
be valid ;). After those problems are sorted out, this
can be uncommented, and will provide another check for
local search mirror sites before they go into an error.
- Reorganize code to have two parts instead of four
(form processing, no form processing)
- Clean up code, adding comments (use switch instead
of a series of ifs, use list() instead of several
value copies, etc.)
- Try to at least preserve the $lang setting (which
is still not used, as before)
- Always exclude "printed" and "printed with notes"
pages from search results
- Provide better defaults, if the search page is
invoked directly (for form target and language)
- Removed long unused search targets (phplib mls, phpkb)
- Resolved a few bugs in parameter passing in
"remote search mode"
- Introduced XHTML formatting to direct output from
this script (like in other files)
- Made the result page's navigation more accessible
for non-image browsers (already discussed for manual
pages)
- Probably other improvements I cannot remember now ;)
This still needs some work, particularly on the language
side, and if any bugs arise from this new code (regarding
the printed page exclusions for example). I hope this is
a good start at least...
