$Id: Release-Notes-1.4.txt,v 1.6 1996/02/02 01:04:36 duane Exp $ TABLE OF CONTENTS 1. Gatherer C version of ftpget Rewriting the URL with post-processing Deleting objects with post-processing Generating RootNode/LeafNode URLs from a program NewsArticle.sum TTLs 2. Broker #Restart-Index-Server command Specifically setting GlimpseServer_Port 3. Miscellaneous Solaris and -lmalloc ACKNOWLEDGEMENTS ======================================================================== GATHERER C version of ftpget ------------------- The 'ftpget' program has been rewritten in C. The Perl version caused coredumps for a number of users. The C version should be more reliable and efficient. Rewriting the URL with post-processing -------------------------------------- Post-processing rules were introduced in version 1.3. At that time it was only possible to change attributes within a SOIF template. It may also be useful to change the URL associated with the SOIF template. A number of users have asked for the ability to gather objects from their HTTP server as file:// URLs but have them appear in the broker as http:// URLs. This can be done with a post- processing rule such as: url ~ 'file://localhost/web/htdocs/' url | fix-url.pl And the 'fix-url.pl' script might look like: #!/usr/local/bin/perl -p s'file://localhost/web/htdocs/'http://www.my.domain/'; Deleting objects with post-processing ------------------------------------- It is now possible to entirely delete an object with post- processing rules. This is done with the 'delete()' instruction. For example: keywords ~ /Limbaugh/ delete() Generating RootNode/LeafNode URLs from a program ------------------------------------------------ It is now possible to generate RootNode or LeafNode URLs automatically from program output. This might be useful when gathering a large number of Usenet newsgroups, for example. The program is specified inside the RootNode or LeafNode section, preceded by a pipe symbol. |generate-news-urls.sh The script must output valid URLs, such as news:comp.unix.voodoo news:rec.pets.birds ... In the case of RootNode URLs, enumeration parameters can be given after the program. |my-fave-sites.pl Depth=1 URL=5000,url-filter NewsArticle.sum TTL value ------------------------- In version 1.4 the TTL value for NewsArticles has been increased to seven days. If you wish to change the TTL for NewsGroup or NewsArticle objects you must edit the NewsGroup.sum or NewsArticle.sum scripts. ======================================================================== BROKER #Restart-Index-Server command ----------------------------- This addition to the broker admin interface makes it possible to force a restart of the index server (e.g. glimpseserver, waisserver). Under normal operations the index server is restarted after every 1000 queries. Specifically setting GlimpseServer_Port --------------------------------------- In version 1.2 the glimpseserver port number was specified in the admin/broker.conf file. On some systems (notably Solaris) this caused problems because the port would become "stuck" and restarting the broker would result in an "address already in use" error. In version 1.3 the glimpseserver port was chosen randomly every time it was restarted. This can make it difficult to debug glimpse and make use of the glimpseserver outside of the broker. In version 1.4 it is again possible to specifically set the glimpseserver port. If it is not defined in the broker.conf file, the port number is chosen at random. ======================================================================== MISCELLANEOUS Solaris and -lmalloc -------------------- We have found that libmalloc.a has memory leaks and/or other bugs in Solaris versions 2.4 and earlier. Therefore we now specifically avoid linking with -lmalloc on Solaris 2.0 through 2.4. ACKNOWLEDGEMENTS Thanks to the following list of people for contributing time, fixes and ideas: Peter J. Scott Grace Tamashiro Allen Forsyth Dirk Allyn Fratkin Ken Weiss Greg Sylvain