release 0.55

git-svn-id: https://svn.berlios.de/svnroot/repos/yacy/trunk@4145 6c8d7289-2bf4-0310-a012-ef5d649a1542
pull/1/head
orbiter 18 years ago
parent ed20531e68
commit 1a0f89d7e8

@ -3,7 +3,7 @@ javacSource=1.4
javacTarget=1.4
# Release Configuration
releaseVersion=0.548
releaseVersion=0.55
releaseFile=yacy_v${releaseVersion}_${DSTAMP}_${releaseNr}.tar.gz
proReleaseFile=yacy_pro_v${releaseVersion}_${DSTAMP}_${releaseNr}.tar.gz
releaseFileParentDir=yacy

@ -85,6 +85,60 @@ globalheader();
</ul>
-->
<br><p><a name="4145">v0.55_20071004_4145</a>
<ul>
<li>Enhanced Search Interface and Search Behaviour</li>
<ul>
<li>Search result pages are displayed directly after search request without any waiting time. Search results are displayed one-by-one using SSI includes when they had been verified by the snippet-fetcher.</li>
<li>Redesigned search process: search results are checked internally for snippet matching and results are presented using SSIs.</li>
<li>All search interface pages now use the progress bar that had been introduced first for image search.</li>
<li>Added search history cache that keeps search results for 10 minutes. Repeated searches for the same topic use the cache for results.</li>
<li>Introduced multiple search result pages.</li>
<li>Enhanced ranking normalization time during search result computation.</li>
<li>Removed Ajax methods from search interface.</li>
<li>Enhanced display of favicons beside search results (less verticall space, better floating of text around, correct position beside result text)</li>
<li>Added profiling information to search result computation: URL fetch time and snippet computation time. These values can be seen in the access tracker.</li>
<li>Added security layer for ViewImage, images are only retrieved if another process has created a view-license before.</li>
<li>The opensearch rss search result (/yacysearch.rss) has always result snippets</li>
</ul>
<li>Enhanced Web Crawls</li>
<ul>
<li>Redesigned appearance of crawl start and monitoring pages in menu. Introduced some intelligent forwarding rules that automatically shows monitoring pages of interest accoriding to crawler/indexing situation.</li>
<li>Extended crawl profile menu; crawls that are finished are not deleted any more but stored in a new table which can also be seen on the profile servlet page.</li>
<li>The size of the remote crawl stack is added to seed information, so other peers can see how much links are provided for remote crawl.</li>
<li>Added option to simple crawl start: complete domain. This is also added to complex crawl start menu and replaces the 'only one domain' button.</li>
<li>Added cache to favicon display.</li>
<li>Fixed bad synchronization in Crawl Balancer that caused inconsistencies.</li>
<li>Added a yacy interface servlet that can be used to download links from the remote crawl stack. This interface is not yet used on client-side.</li>
<li>Enhanced crawl start with link file: URLs are placed in the preNURL table instead of NURL table. This enables crawl starts with a huge number of links in crawl start files.</li>
<li>Fixed problem during crawling with redirection: these urls had not been double-checked.</li>
</ul>
<li>Speed-up Enhancements</li>
<ul>
<li>Enhanced memory allocation methods in kelondro, less overall memory copying (about 50%).</li>
<li>Adopted memory calculation functions for thread control to reduced memory consumption of collection objects.</li>
<li>Removed object pool from httpc and introduced client connection supervision. Long-running idle connections are closed automaticaly.</li>
<li>Redesigned URL hash computation, this used a is-local method which itself nedded a DNS lookup, and that needed time. Now the URL hash computation is prevented in many cases where the URL can be read from a YaCy database which had been accessed with the URL hash. The result is a overall speed up and less network operation in all parts of YaCy.</li>
</ul>
<li>Other new Features</li>
<ul>
<li>Enhanced SSIs, refences may now also be servlets, not only files. SSIs work now also for non-chunked-encoding clients.</li>
<li>In pro-releases all parsers are activated by default.</li>
<li>Added client connection table to Connections servlet.</li>
<li>New peers compute their peer hash according to present DHT situation, a optimal position is calculated and used as new peer hash.</li>
</ul>
<li>Bugfixes</li>
<ul>
<li>Auto-Update functionality now works on Solaris/OpenSolaris.</li>
<li>Various fixes and cleanups for blacklist handling.</li>
<li>Fix problems with some web servers which couldn't handle the way yacy was sending requests.</li>
<li>Fixed round-robin method of peer-ping order (the successfully pinged peer was not updated to current last-seed date).</li>
<li>Fixed connection problem on http client side implementation: disconnections are not treated as error any more since sudden disconnections are allowed as EOF signal if no file length is given in http header. Because the YaCy protocol uses exactly this method, many peers had been not reacheable after the delay at the end of a server session was removed. After this fix there was much better availability between peers.</li>
<li>Many more minor bug fixes.</li>
</ul>
</ul>
<br><p><a name="4021">v0.54_20070802_4021</a>
<ul>
<li>Option to define separated YaCy networks</li>
@ -126,7 +180,7 @@ globalheader();
<br><p><a name="3914">v0.53_20070618_3914</a>
<ul>
<li>Enhance Search Interface</li>
<li>Enhanced Search Interface</li>
<ul>
<li>Speed-up of web interface (more responsive)</li>
<li>Search results without snippets are blinded out</li>

@ -1,197 +0,0 @@
<?xml version="1.0"?>
<rss version="2.0">
<channel>
<title>YaCy: News</title>
<link>http://www.yacy.net/yacy/News.html</link>
<description>This is essentially the YaCy release change-log.</description>
<language>en-us</language>
<lastBuildDate>Tue, 30 Jun 2007 09:00:00 GMT</lastBuildDate>
<image>
<url>http://www.yacy.net/yacy/grafics/yacy.gif</url>
<title>YaCy</title>
<link>http://www.yacy.net/</link>
</image>
<item>
<title>New Release V0.52 (20070512_3715)</title>
<link>http://www.yacy.net/yacy/News.html#3715</link>
<pubDate>Sat, 12 May 2007 00:12:29 GMT</pubDate>
<guid>3715</guid>
<description><![CDATA[
<ul>
<li>New Functions</li>
<ul>
<li>Added exclusion-search (a search with '-' to exclude specific words from the search results)</li>
<li>Added extraction of sitemap-url from robots.txt, which can be used for crawl starts</li>
<li>Added a network configuration menu for new cluster configuration functions: a set of peers may now operate as an isle within the YaCy network, without exchange of index data over the border of the isle. Peers within the cluster can trigger internal remote crawls and search only within the own cluster.</li>
<li>Added a postscript parser</li>
</ul>
<li>Interface Enhancements</li>
<ul>
<li>Redesigned the status page, shows now also hints and warnings</li>
<li>Better layout for image search results</li>
<li>The peer profile can now be displayed as vcard, e.g. http://localhost:8080/ViewProfile.vcf?hash=localhash</li>
</ul>
<li>Performance Enhancements</li>
<ul>
<li>Added an option to configure a path to a secondary index location.
This shall be used to store a fragment of the index on another physical device,
to split IO load and enhance access speed. The index is splitted in such a way
that the LURLs are stored to the secondary location, and the RWIs to the primary
location.</li>
<li>Optimized memory allocation when accessing the web-index (now half of memory throughput as before)</li>
<li>Fixed bugs in database engine that corrupted the data when entries had been removed</li>
</ul>
</ul>
]]></description>
</item>
<item>
<title>New Release V0.51 (20070321_3501)</title>
<link>http://www.yacy.net/yacy/News.html#3501</link>
<pubDate>Wed, 21 Mar 2007 15:00:00 GMT</pubDate>
<guid>3501</guid>
<description><![CDATA[
<ul>
<li>Better Crawling</li>
<ul>
<li>Higher crawling speed possible caused by better ram cache flush methods</li>
<li>The crawl balancer now has a security function which prevents that remote web servers are accessed more than two times in one second. In case a crawling from a single domain is made, this means the crawling speed is restricted to not more than 120 pages per minute</li>
<li>The crawl balancer chooses better urls. Newly added urls are now prevented from beeing hidden by masses of links generated by the crawler. The effect is that in most cases the security function described above is not needed.</li>
<li>Added a crawling speed button on the crawling monitor page.</li>
<li>Crawl targets get informed about the yacy bot; a link to http://yacy.net/yacy/bot.html is attached to each crawl request; the page explains YaCy and that YaCy respects robots.txt</li>
</ul>
<li>Better Monitoring</li>
<ul>
<li>New search result page SearchStatistics_p.html shows local and remote search requests; remote requests are anonymized</li>
<li>Added network-wide QPM (queries per minute) computation to show how much the network is used for web search. The statistics are not reported from searching peers, but from searched peers; therefore the accumulation preserves privacy of the searcher</li>
<li>New page LogStatistics_p.html which shows an evaluation of entries from the log.</li>
<li>New page BlacklistCleaner_p.html to clean up wrong blacklist entries. The page allows categorization of blacklist error case, correction of error and the optional deletion of the blacklist entry.</li>
<li>Added RSS feed for YaCyNews</li>
</ul>
<li>Enhanced User Interface</li>
<ul>
<li>Added a robots.txt configuration menu to enable/disable external crawlers to access the yacy user interface</li>
<li>New wiki-parser</li>
<li>Blog entries may now have user-comments</li>
<li>The network list page now provides links to the users blog pages</li>
<li>The menu points had been rearranged</li>
</ul>
<li>Less Memory Usage and Better Memory Management</li>
<ul>
<li>All caches (node cache, object cache) now have enhanced self-organization and dont need fixed size assigments</li>
<li>Memory protection by disallowing collection arrays beyond kca-7. Collections larger than those are written to 'common' files.</li>
<li>The network picture uses less memory</li>
</ul>
<li>Bugfixes: a very large number of bugfixes were made.</li>
</ul>
]]></description>
</item>
<item>
<title>New Release V0.50 (20061222_3124)</title>
<link>http://www.yacy.net/yacy/News.html#3124</link>
<pubDate>Tue, 30 Jan 2007 09:00:00 GMT</pubDate>
<guid>3124</guid>
<description><![CDATA[
<ul>
<li>Added Media Search</li>
<ul>
<li>Added search pages for Images, Audio, Video and Application search.</li>
<li>Added media link presentation during snippet fetch; the Image Search presents search results as image thumbnails.</li>
<li>Better recognition of search hits for text snippet generation.</li>
<li>Media search results are indexed again after remote search results are collected; only media links are used to update the index.</li>
</ul>
<li>Better Result Ranking</li>
<ul>
<li>New ranking parameters and appearance attributes are now considered.</li>
<li>Faster ranking; more references can be ranked and sorted within given search time.</li>
<li>Ranking Parameters can be handed over to remote peers and are applied there.</li>
<li>Adopted Detailed Search to new ranking parameters.</li>
<li>Coefficients from detailed search can be set as default ranking for search page; this replaces the old ranking alternatives.</li>
</ul>
<li>Better Crawl Monitoring</li>
<ul>
<li>After a crawl start was initialized, the Crawler Monitor is shown.</li>
<li>The Crawl Monitor now shows all queue elements in one table.</li>
<li>Monitoring of index size.</li>
<li>The Crawl Profiles are shown; crawls can be interrupted within the profile table.</li>
<li>A crawl may now distinguish between text indexing and media link indexing.</li>
</ul>
<li>Migration to new Database Structure</li>
<ul>
<li>The new Collection Database is now the only database structure that can be used; Assortments are switched off.</li>
<li>Added functions to migrate Assortment databases and WORDS databases to Collection database.</li>
<li>Removed all methods to write Assortment data structures.</li>
<li>Migrated DHT position computation to base64-decoded values; this changes the DHT structure slightly and closes the gaps in the old DHT structure.</li>
</ul>
</ul>
]]></description>
</item>
<item>
<title>New Release V0.49 (20061202_3040)</title>
<link>http://www.yacy.net/yacy/News.html#3040</link>
<pubDate>Sat, 02 Dec 2006 09:00:00 GMT</pubDate>
<guid>3040</guid>
<description><![CDATA[
<ul>
<li>Enhanced search service</li>
<ul>
<li>Web searches are faster because of the new data structures implemented in this version (see below)
and because bugs had been found and fixed.</li>
<li>Searches can be re-done with changed search properties. Please use the 'more options' link at the search page.</li>
<li>Added search constraints. These are search restrictions to web searches which are applied to information that is scraped
from the web pages during page parsing. The first application of search constraints is a search restricted to
index pages ('index of'). Please use the flag at the extended search functions.</li>
<li>Enabled index-abstracts search; this should solve the distributed-combined search challenge (still beeing tested).</li>
</ul>
<li>New Database Structures for Index and URL storage</li>
<ul>
<li>The new 'Collections' Data Structure is now the default data structure.</li>
<li>Index entries and URL entries carry more ranking and selection attributes, e.g. for image, video, audio and application search.</li>
<li>Enhanced Storage of URLs: they are now divided into different creation times. This enables easy deletion
of outdated URLs, enables a index-limitation function and solves the problem that the URL database was too
big to fit into a 2 GB file.</li>
<li>Search requests can now be answered in less time.</li>
<li>The index organization needs less IO.</li>
<li>Index transfers will now only be done to latest peers supporting the collection data structure.</li>
<li>Index transfers from old peers to new peers are translated automatically to new data format.</li>
<li>Assortments are no longer supported.</li>
</ul>
<li>Enhanced SOAP support</li>
<ul>
<li>Added protocol for peer administration, custom services, status queries, blacklist management,
file share management, support for outgoing transfer- and content-encoding, better error handling,
function to get and set message forwarding, handling of YaCy bookmarks, log display,
manage peer messages, get and set peer profile, query peer status, query the pause/resume state of the crawling queues,
and a check if a specific URL is blacklisted.</li>
<li>Added new ANT target to allow generation of client stub classes for YaCy SOAP api.</li>
</ul>
<li>Other new Features</li>
<ul>
<li>Added DNS-cache-miss caching.</li>
<li>Added Flash (experimental), MS Excel and Powerpoint parser.</li>
<li>New mint-green and dark skin.</li>
<li>Better non-7bit ascii character support.</li>
<li>Added ant support for rpms.</li>
<li>Added ant target for windows installer.</li>
<li>Added template to display file share in xml format.</li>
<li>Better object caching for kelondro database (combined read/write object cache with synergy effects).</li>
<li>More anonymization in logging.</li>
<li>New HTCACHE layout using files hashes; tree- and hash-layout can be used simultanously; hash-layout is now default.</li>
<li>Access to wiki is now limited to administrator, if wanted. This can be configured at the wiki page.</li>
<li>..and many bugfixes.</>
</ul>
<li>New 'satellite' Projects: these applications work as service applications for the YaCy application (start-up/experimental status)</li>
<ul>
<li>YaCy admin: a swing-based client, that is able to administrate yacy using the SOAP interface.</li>
<li>YaCy Screen Saver: presentation of the peer status in a screen saver</li>
<li>YaCy Updater: automated donwloads/updates</li>
<li>YaCy logalizer: analyzer for the YaCy log</li>
</ul>
</ul>
]]></description>
</item>
</channel>
</rss>
Loading…
Cancel
Save