[TYPO3-core] RFC: "Page is being generated" and search engines

Dmitry Dulepov dima at spamcop.net
Mon Jan 16 21:36:54 CET 2006


Type: fix

Problems:
1. Bug 1947 ("Page is being generated" is bad for search engines). There are
12900 results for this message in Google index. Proper content of those page was
not indexed and cannot be found
2. Refresh time is too small for "Page is being generated" message
3. Accompanying fix: Uninitialized variable $tableNameChar
4. Accompanying fix: Uninitialized variable $INTiS_key

Solution:
1. To let Google know about temporary problems with server, tslib_fe will set
HTTP status code to "503 Service unavailable" and supply a set of HTTP headers
that disable caching of this page (Cache-control, Pragma, Expire) and ask search
engine to return to the page later (Retry-after).
To make sure that these headers are returned, cache_pages table is extended with
new field named "temp_content" (int(1)). tslib_fe will send the above mentioned
headers if this value is set in the cache record.
2. "Page is being generated" often means server overload. Currently timeout is
set to 5 seconds which is too small to lower server overload. In this patch it
is changed to 20 seconds.
3. Replaced with proper variable.
4. Line swapped with the next line where this variable was set.

This is a serious change, so I ask everyone to take a look at it carefully.

(Sebastian, I made a simpler solution, not the one we talked about at bug party.
Question about http headers for USER plugins remains, I will make a patch for it
later).

Dmitry.
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: cms.txt
Url: http://lists.netfielders.de/pipermail/typo3-team-core/attachments/20060116/cced7f5c/attachment.txt 


More information about the TYPO3-team-core mailing list