changelog

网络爬虫程序

* it is now posible to disable processing of some URLs by using of Tree
  preview

version 0.9pl2 (Sep 6 1998)
--------------
* minor bug fixes reported by some users
* repared bug ,when -cdir ends with '/' and using -base_level switch results 
  to broken filenames
* implemented interactive downloading using URL tree preview dialog
* solved problem in GTK URL tree preview with more starting URLs
* URL tree preview dialog in Xt interface is now not modal
* basic support for sending and receiving HTTP cookies (writing to cookie file
  not supported yet, GUI can't hand cookie parameters - only via cmd-line)

version 0.9pl3 (Sep 20 1998)
--------------
* inteligent updating of cookie file implemented (the some file may be updated
  with more proceses concurently without cookie looses)
* GUI interface for cookies setup
* HTML file on FTP server is processed right
* repared rewriting of redirected url with fragment name specification
* you can now download from URL tree preview manualy files which were broken or 
  rejected

version 0.9pl4 (Jan 6 1999)
--------------
* cookie file may contain any coments started by '#' 
  (not saved back after update)
* host name translation errors are reported now right
* buffered IO implemented
* some minor bug fixes 
* repared any segfaults
* new & more icons for URL tree preview
* HTML tag & attribute restrictions for slection of URL's from HTML docs
* checking cookies if source domain is equal with domain attribute of 
  Set-Cookie MIME entry
* cookie file is now right ordered (not reversed each time :-)
* new czech message catalog in ISO8859-2 encoding by Petr Vyhnalek
* added new switch -gui_font , which allows you to set font used in 
  GUI interface
* added new switch -language for used to set language of messages while
  compiled with GNU gettext support
* added very simple SOCKS(4/5) support (not tested yet)
* -pattern accepts comma-separated list of documentname matching patterns
* new option -url_pattern to enter comma-separated list of url matching
  patterns
* -user_condition options added to provide option for user to specify by
  external script or program if URL should be processed or not
* repared bug when extra space characters in scenario file are not removed 
* repared seg-fault while doing HTTP reget (thank to Orestes Sanchez Benavente)
* added -disabled_cookie_domains option

version 0.9pl5 (Jan 28 1999)
--------------
* you can now immediately change communication language from GTK GUI
* added gtk-config script to configure script for GTK configuration checkings
* added client certification stuff for HTTPS (SSL) (not tested yet)
* some segfaults repared in GUI code
* repared time handling bugs
* added realm info to authinfo file
* HTTP authorization schemes are now handled properly
* HTTP digest access authorization implemented (it work with my apache server)

version 0.9pl6 (Feb 28 1999)
--------------
* when compiling with SSLeay lib using md5 computing rutines from libcrypto.a
  instead of apaches md5c.c
* reuse of HTTP digest access nonce in more following requests is now 
  implemented 
* digest authorization with proxy server
* added QueryGeometry to all Nws widgets for windows autosizing 
  (finaly - I am so lazy :-))
* filename conversion rutines for changing local filename 
  (delete set of characters , change string to string , tr like char to char)
* language change now work too if some files were processed 
  (Tree preview not destroyed)
* while changing language all visible windows stay visible
* menu entry labels are GNOME compliant
* beautify of xinterface.c
* rewriten Xt interface to support language change from GUI
* each file selection entry now have browse button
* send QUIT signal while running in text mode and pavuk will exit safe
* added sample of Xt resources file for Pavuk
* thank to H鍁ard Skinnemoen added some features from gtk+-1.1.* 
        - new style of adding childs to scrolled windows
        - parsing of ~/.pavuk-gtkrc
* solved win32/cygwin32/unix file path madnes

version 0.9pl7 (Mar 30 1999)
--------------
* changes for support GTK+-1.2.0
* removed sk and cs ascii message catalogs from distribution
* repared comandline time parameter scaning routine
* all labels in GTK interface are now left justified
* scheduling now work well
* solved problems when compiling without GNU gettext support and with GUI 
  support
* a lot of GTK improvements
* better processing of some stupid HTML constructions
* HTML comments and inline scripts are not parsed && processed
* default location of system pavukrc changed from $(prefix)/lib/pavukrc
  to $(prefix)/etc/pavukrc
* added a lot of new HTML tags for processing

version 0.9pl8 (Apr 12 1999)
--------------
* now compile with gettext support on systems without LC_MESSAGES defined
* checking of robots.txt now work again (thank to Stefan Stidl)
  - checking disabled in many previous versions because of oddly writen 
    condition :-(
* better detection of cyclic HTTP redirections
* repared SEG fault while in GUI and HTTP redirection to already processed
  document occurs
* new icons for buttons added from Andreas Kraska . If you want old buttons, 
  execute configure script with --disable-new_buttons option.
* accelerated menubar with GTK+-1.2<
* using putenv on system where setenv & unsetenv not found
* a lot of minor bug fixes

version 0.9pl9 (Apr 18 1999)
--------------
* repared bug, when all documents downloaded over HTTP/HTTPS were processed as
  HTML documents (a lot of rewriting operations on binary files :-()
* repared implementation of setenv/unsetenv on systems where not implemented
  (thank to Orestes Sanchez Benavente)
* timeout on connect() call
* now pavuk work on filesystems, where doesn't work link() call (FAT)
* better detection of already downloaded directories
* not buffered read while reading document data from net
* new Action menu
* enhanced use of GTK+-1.2 < features (GTK 1.0.x compatibility preserved)

version 0.9pl10 (Apr 25 1999)
---------------
* repared bugs in net_connect() function
* repared bug while using active ftp connection
* you can now miniaturize main pavuk window (GTK+ only)
* !!!!! -progres option repared to -progress
* new option -runX (you can immediately start downloading files after GUI 
  interface is started)
* simple support for CSS
* a lot of bugs fixed

version 0.9pl11 (May 2 1999)
---------------
* new -index_name option used to change default name of directory index
* new -store_name option used to set filename for document downloaded with
  -mode singlepage
* changed version of used autoconf (1.3) and automake (1.4)
* support for processing standalone CSS files
* doesn't get SIGPIPE when decoding encoded file (not fork-ing in GUI)
* using CTree widget instead of Tree with GTK+-1.2

version 0.9pl12 (May 5 1999)
---------------
* new option -ftplist to use wide listing of FTP directories (using LIST 
  ftp cmd instead of NLST) (only unix style of list supported)
* new option -preserve_perm to preserve options of ftp files 
  (assume -ftplist option)
* now pavuk saves ftp symbolic links as symbolick links not normal files
* new option -preserve_slinks to leave point symbolic link to same location
  as on remote server.
* Go Bg button now work properly with GTK+ (thank to Jan Kratochvil)
* new option -FTPhtml/-noFTPhtml to enable/disable processing of files
  downloaded over FTP protocol
* anchor names for FTP urls now parsed right

version 0.9pl13 (May 16 1999)
---------------
* pavuk now removes empty directories in local document tree
* directories are now processed right
* new option -min_size to eliminate transfer of small documents 
* new options -skip_url_pattern and -skip_pattern
* repared bug in document time preservation (thank to Tomas Dobrovolny)
* while updating parent document links, and it is locked, pavuk will wait
  until lock will be released
* locked document is allways rescheduled

version 0.9pl14 (May 23 1999)
---------------
* thank to Steffen Kern added droping of URL's to url list and pavuk main window
  (for example from netscape)
* thank to Tomas Dobrovolny fixed some minor bugs in configure.in script
* new HTML tags for table backgrounds added (thank to Szabolcs Szakacsits)
* new -htDig option for cooperation with htDig web indexing program
* new option -check_size/-nocheck_size for enabling/disabling checking of
  document size (some HTTP servers report bad Content-length: header)
* minor bug fixes

version 0.9pl15 (Jun 21 1999)
---------------
* many fixes and changes in HTML parser code
* better support for Cascading Style Sheets
* lot of patches from Szabolcs Szakacsits and Stefen Kern added
* fettching of URLs from clipboard implemented for GTK and Xt GUI
* repared encoding of URLs (thank to Marc Haber and Szabolcs Szakacsits)
* new option -urls_file (for reading URLs from file or stdin)
* get SSL stuff working again (was broken because of non-blocking IO)
* updated czech message catalog (by Petr Cech)
* new icons in icons/ directory
* a lot of changes / bug fixes

version 0.9pl16 (Jun 29 1999)
---------------
* checking for zero size of file
* fixed bug with using -store_name option (thank to Marc Haber)
* new type of log file added (option -slogfile)
* -mode resumeregets now recurse through links
* removed many memory leaks inside new HTML and CSS parser code
* removed some random crashes with Xt GUI

version 0.9pl17 (Jul 06 1999)
---------------
* bigger read buffer -> better read performance on fast connections
* new option -identity for specifying User-Agent: HTTP request field
* new option -nosend_from for deny sending From: field with HTTP request
* new option -nostore_index used to tell pavuk not to store documents 
  referenced with directory URLs
* new option -acharset used to specify set of preffered document encodings
  for HTTP protocol
* changed selection retrieving with GTK+ GUI
* better native language switching in internetionalized environment
* bug fixes

version 0.9pl18 (Jul 26 1999)
---------------
* support for EPLF format listing of FTP directories
* support for Novel format listing of FTP directories
* repared one typo which breaks compilation without GUI
* automatical prefferences saving/loading to file ~/.pavuk_prefs
* loading & saving of menu accelerator keys to prefs file
* fixed type casting bug in html/css parser code (thank to Robert Gasch)
* support for newer openssl versions (0.9.3<)
* better & nicer progress meter
* limatation of transfer speed (max/min)
* my CERN HTTP/proxy server is somehow odd - synchronization of WWW pages
  wont work if you specify port number in URL (currious), so port number
  was removed from URL if portnumber is default.
* sync mode work now well when spaning to another server
* sync mode work again with servers which not respond right 304 code (mea culpa)
* added Apply button to configuration dialogs
* fixed lot of bugs in net_connect function
* instaltion of pavuk icons to $(prefix)/share/icons/
* new quota options (quota for file size, transfer amount and free space on
  filesystem)
* solved bug, when Gtk+ URL list not show its contents
* solved bug, when pavuk crashes on redirection to unsupported URL
* corrected fetching of URI: header content for redirected URLs
* several bug fixes and improvements

version 0.9pl19 (Sep 06 1999)
---------------
* changed URL equivalence checking from filename based to URL based
* internal URL representation now contains its local filename  , this means
  lower memory footprint, but bigger memory consumption
* several minor memoryleaks removed
* implemented universal & flexible mapping mechanism URL -> local filename
  based on RE or wildcard patterns and simple rules (see manual ,