ein wenig unterstützung schadet garnix.
also rann ans netz und die leitung glühen lassen


Code: Alles auswählen
v1.6.4 05/08/08
+ New approach to versions: even versions will be stable release builds and odd ones will be testing betas.
! Fixed rare condition when anchor text won't be taken
! Fixed rare entity decoding problem in HTML parsing
! Improved support for XML content types
+ Added support for parsing of XML data streams (like RSS/Atom feeds)
+ Added support for text analysis for domain/url mention without being explicitly put between proper HTML tags
! Blank content-type no longer prevents crawling of the page
! Fixed bug resulting in inability to saved cached urls in some fairly rare cases
! Switched off web server request logging
! Longer urls are now taken for analysis + too long urls are handled in a smarter way then just dropping them
! Fixed bug with buckets failing to archive 3 times and then never getting a chance again until restart
! Dup url debug message removed
! Fixed incorrect message about buckets that failed archiving 3 times - ETA for next attempt should be correct now
! Less useless entries in log
! Fixed issue with some rare buckets hanging to be archived but never making it due to archiving being done before
! Fixed rare url decoding case
! Better support for character encoding detection
! If robots.txt block is detected during redirection then Abort is issued for compatibiity with
some web servers that would otherwise cause timeout, same result but cleaner error.
! We will avoid crawling site if we get hard errors like 501 Server unavailable when requesting robots.txt
! Domains that disallow all requests will get finished faster (main benefit for deep crawl ones)
! Better session ID detection in urls
! Fixed parser bug not taking ?xml element from feeds thus not setting correct enconding
! Smarter handling of too long urls if they were url encoded (mainly helps non-Latin based content types)
! Improved removal of sessionIDs used in URLs
! Fixed possibility of some buckets never get selected for crawling - now oldest received bucket for its priority
will be chosen for crawling
! MJ12bot version increased to 1.2.2 due to enhanced behavior when dealing with robots.txt errors 
Code: Alles auswählen
v1.6.5 09/08/08
! Back to old version scheme (stable/release odd/even versions is scraped)
! More session IDs filtered
! Fixed bug with second pass parsing in XML parsing when encoding was non-Latin :(
! Some plain text urls were not taken from analysed data
! External links that look like images (when not part of <img src> tag are saved now (won't be crawled though)
(excludes XML like RSS feeds) 
Code: Alles auswählen
v1.6.6 14/08/08
! Fixed failure to clean up some urls with session ids that were already removed (also make session clean up process more reliable)
! Removed more unnecessary warnings
! Change in how Accept header is set for Mono builds 




Code: Alles auswählen
v1.6.7 19/08/08
! Changed referer set when robots.txt is requested to make it shorter and avoid potential
false positive by faulty security software that should have never checked it in the first place 
Code: Alles auswählen
[29.12.2008 14:31:48] Failed to get server stats: ConnectivityProblem -- TIMEOUT!

