CHANGES

Now images and other non html documents are copied under dest_dir. If one still wants to have symbolic links for those files, the new switch -k can be used.
Small changes to source to compile and work on Redhat 5.0
Fixing of a bad bug in scanner.lex
Changes to compile properly on Irix 5.3 (thanks to Andrei Mircea).
Some other small changes and enhancements.

The behavior of the -w switch has changed. Now, by default, pages are showed in another window, while with -w the same index window is used. -W is unchanged.
Html files are now created directly to the proper place under dest_dir. There is no longer need for temporary files and for write permission on /tmp. And there is no longer the problem of lacking html files when dest_dir and /tmp were on different partitions.
New option -l for ignoring lock file in $HOME/.netscape when one is sure that the cache isn't used by that Netscape.
Now, the Makefile, by default uses flex. This is to avoid problems on some Slackware systems where lex is a call to flex -l.
Better diagnostics.
Various small corrections to documentation.
Some changes to sources.

Default destination dir is now $HOME/cached.
Incorrect Html is supported better.
Bug regarding displaying of informations about main and related html files fixed.
Better protection against crashes in strange stituations.
Better management of user (Ctrl-C) interruptions.
Better support for using by a non-root user.
Some small bugs or malfunction fixing.
Source files now include short descriptions of what each function does.

Bug regarding http links contained in ftp documents fixed.
Now the program makes distinction between various command line errors.
Options, unless n_hours, may now be given grouped (i.e. -pt or pt rather than -p -t)
Fixing of a bug in the summary file occurring with names different from index.html.
More precise management of different documents with the same url.
If the program is launched without options, the user is asked if he really wants to proceed.
Url-type prefixes (http, ftp) are all converted to lower case.
Changes to documentation.

Changing of the name.
Summary file format has changed drastically: now it's based on the division into domains of retrieved pages.
Now the program can process cache directories created by Netscape for Windows present in non Linux-native partitions. Adding of the -p option.
Now n_hours condition is checked basing on informations contained in index.db, not on files' timestamps. This way is faster and better.
Temporary HTML files are now created in /tmp directory.
Adding of the -W option.
Now the program supports the situation in which there are in the cache multiple (different) files referring to the same url.
Adding of the script install.sh, for installing precompiled version (make install recompiles).
Various minor bugs fixing.
File organization in distribution changed.
Various changes to sources to remove MAX_PATH limitation on command line arguments.

More complete management of problems connected to the use of the n_hours parameter. Adding of the -f option.
Adding of the option -g, which permits to process only HTML files whose URL contains a specified string. Also adding -G, its opposite.
Bug related to -t option fixed.
Some links to local documents and images didn't function due to a bug now fixed.
The same problem as above was due to another bug also fixed.
Now netcache works even if Netscape Navigator isn't installed, that is even if the directory $HOME/.netscape doesn't exist.
Progress meter when reading index.db file.
Added support for disk-full situation.
Added support for multiple versions of the same document in the cache.
Heavy changes to documentation.

Adding of the option -t, which permits to see downloading date of each document in the summary_file .
When interrupted by the user, the program now deletes current temporary file in $HOME/.netscape/cache .
Some changes to documentation.

The behavior of the -m switch has changed. Now, by default, missing images are totally deleted, and -m is used to not delete them.
If no documents are retrieved, no summary file is written and the user is informed.
Messages during execution are cleaner and more compact. Informations now include total number of files to be processed.
Verifying of the n_hours condition is made before in the program, saving execution time.