A. Release Notes

The previous full release of MIT Scheme was version 7.5.17. This section describes changes that have occurred since that time. For more detailed information, see the `ChangeLog' files in the source code.

Note that MIT Scheme still conforms to the Revised^4 Report on the Algorithmic Language Scheme, but not to the Revised^5 Report on The Algorithmic Language Scheme.

A.1 Recent Changes
A.2 Older Changes

A.1 Recent Changes

These are the changes since release 7.6.0:

• Release 7.7.1 fixes several bugs in IMAIL; fixes a bug that prevented the use of server sockets on Windows systems; and fixes a bug that caused the debugger to generate errors in common circumstances.

• Release 7.7.0 provides "hygienic" macro support, as defined in R^4RS and R^5RS. This is a complete rewrite of the syntax engine, so any program that uses macros should be rewritten to use the new engine. A subset of the old macro-definition syntax is still supported, but this will eventually be removed. Note that the new syntax engine has no effect on the compiled-code format; most binaries compiled by release 7.6.x should continue to work.

  There are several user-visible consequences to this change:

  • These syntactic keywords have been eliminated: define-macro, in-package, macro, make-environment, scode-quote, unassigned?, and using-syntax.

  • The syntactic keyword the-environment has been restricted to use in top-level environments. It is no longer allowed in the body of any binding form (e.g. lambda, let).

  • Syntactic keywords are now stored in environments, rather than in a separate syntax-table structure. The environment abstraction has been enhanced to support this, as well as to make it more general. The changes are documented in the reference manual.

  • The syntax-table abstraction has been eliminated, and most procedures and arguments involving syntax tables have been removed. One exception is the load procedure, which still accepts a syntax-table argument, but ignores it.

  Although the 7.6.1 release had a workaround for problems with certain AMD Athlon processors, the workaround was ineffective on machines running Windows operating systems (and possibly OS/2 systems as well). This version fixes that problem.

  The hash-table abstraction is now always loaded. It's not necessary to call load-option prior to use of hash tables. For upwards compatibility, calling `(load-option 'hash-table)' is still permitted but does nothing.

• Release 7.6.1 provides a workaround for a bug on some AMD Athlon processors (specifically, models 1, 3, and 4). This bug causes those processors to incorrectly execute certain kinds of self-modifying code, including that used by MIT Scheme.

  Release 7.6.1 also fixes a problem that caused the *PARSER and XML subsystems not to load correctly on Windows and OS/2 systems.

A.2 Older Changes

These are the changes that occurred between releases 7.6.0 and 7.5.17:

• The variable-reference machinery has been completely redesigned. The new design provides two new features: (1) a procedure link-variables that can link two variable bindings with different names (previously environment-link-name could only link bindings with the same name); (2) a new procedure unbind-variable that can remove a variable binding. Both of these operations work only on "top-level" environments; they don't work on environments created by invoking compiled procedures.

  This is an incompatible change to the compiled-code representation. All programs must be recompiled.

• The "packaging" subsystem has been redesigned to make it more flexible. The new design takes advantage of the improved variable-linking mechanism to allow packages to export bindings using names different from their internal names. Also, each package now maintains an in-memory data structure that should (in principal) allow it to be uninstalled or reinstalled. This subsystem is still evolving; expect further changes in the future.

  This subsystem is now included in the released binaries; it can be loaded by `(load-option 'cref)'.

• Compiled-code debugging information is now keyed to prevent using the wrong debugging info when examining compiled code. This means that in-memory data structures have an identifying key that is matched against a corresponding key in the debugging-info file. The debugging info is used only when the keys match.

• MIT Scheme now has a non-validating XML parser that is mostly conformant (except that it doesn't support UTF-16).

• There is very limited support for UTF-8. At present this is only used by the XML parser.

• All string and character procedures now operate on ISO-8859-1 rather than US-ASCII as the base character set. This isn't real internationalization support, but at least it broadens the character support slightly. This involved changing a number of procedures that were formerly written in C and are now written in Scheme, which may have a performance impact on some uses.

• The procedure length now detects circularities in its argument and signals an error.

• A new rexp facility provides a more readable syntax for specifying regular expressions. This syntax can be compiled into the usual syntax.

• There is a new parser buffer abstraction that supports the building of parsers with backtracking. An associated high-level parser description language allows parsers to be described in a BNF-like language that is compiled into efficient Scheme code.

• A new procedure flo:finite? returns `#f' for infinite or `NaN' flonums and `#t' for all others.

• The signal-handling code for GNU/Linux systems has been changed to take advantage of additional information identifying the types of arithmetic exceptions. This allows more accurate error messages to be generated.

• The following procedure names were changed:

  Old name	New name
  list-transform-positive	keep-matching-items
  list-transform-negative	delete-matching-items
  list-search-positive	find-matching-item

  and the following new procedures were implemented: keep-matching-items! and delete-matching-items!.

• IMAIL now supports commands to sort folders according to various criteria. The design is similar to that of Rmail.

• IMAIL has been re-engineered to improve performance on very large folders. There is now a disk cache for each IMAP folder, to eliminate unnecessary network delays. Also, the summary buffer previously used a quadratic algorithm to associate lines in the summary with messages in the folder; it now uses a linear algorithm.

• A new IMAIL command imail-file-message appends the text of a message to a file more or less as it appears in the buffer. This is intended to allow collecting messages into a file in a human-readable format.

• IMAIL used to determine the type of file folders by their file names; now it examines the file contents instead. When creating a new file folder, it prompts for the type.