
Changes to curl since v1.0
==========================

Listed below are the major changes to the curl
distribution since v1.0. More implementation details
can be found at the start of the C files.


v1.1 (July 1995)
================
Links can now be represented by user-specified icons.
The details on how to do this are in the user guide.


v1.2 (September 1995)
=====================
For those people with curl v1.1, the only file that needs
to be updated is curl.c. The changes have been mostly at 
the implementation level.

* Each filename string used by the main curl data structures
  are stored only once. Previously, each reference to a 
  filename was stored as a separate string. The change means 
  that curl can deal with larger sets of files.
  This idea came from Jan Vitek (jvitek@cui.unige.ch).
  Thanks Jan.

* Lines in the keywords file keys.html can now be much longer.
  If this length is too short then an error message is 
  output. Previously, curl bombed.

* curl now allows two files to have the same name, but
  the directories where they are located must be uniquely named.


v2.0 (May 1996)
===============
no_link, the word occurrence tools, and the link tools are
additional to the main curl application, and do not require any
changes to documents created by earlier versions of curl.

However, the changes to the curl application mean that documents
created in earlier versions will need to be reprocessed to
take advantage of the new features.

The most visible changes to the curl application are the
introduction of a new additions page, a search page (and engine),
and the use of icons to represent the absence of links.

The easiest way of adding these features to a document
is to decurl it, delete prev_contents, and start afresh.

The following tasks need to be carried out in order to upgrade:

* Update curl and decurl (modtime has not been altered).

* Add the new GIF images to the 'arrows' directory: new.gif,
  search.gif, and any `no_' GIF files.

* Modify the style files to include references to %new and %search.

* Compile sk.c (the search engine) and install it in the local
  machine's CGI scripts directory.

* Places copies of new.html and search.html in the top-level
  directory of the document. Modify search.html to refer to
  the directory and sk. See section 5.3.6 of the user guide
  for details, or 'srch-setup' in the 'docs' directory.

The document can now be reprocessed by curl.

