Chris Hanson [Sat, 29 Sep 2001 03:00:13 +0000 (03:00 +0000)]
START-PROBE-FOLDER-THREAD is being called whenever a message is
selected, but unfortunately if there was already a thread, it was
being killed and a new one started, which generated network traffic on
an IMAP folder. Now it doesn't do anything if there is already a
thread.
Chris Hanson [Tue, 25 Sep 2001 05:30:24 +0000 (05:30 +0000)]
Extensive rewriting to make type-checking more uniform, and to
eliminate cases where it was being performed twice. Eliminate even
more primitives, and speed up the procedures as much as possible.
Chris Hanson [Tue, 25 Sep 2001 05:11:11 +0000 (05:11 +0000)]
Eliminate a large number of inlined primitives. Some of these are
necessary due to changes in the runtime system. Others are eliminated
because I don't want their names reserved, or because it's
inappropriate to inline them.
Chris Hanson [Mon, 24 Sep 2001 05:24:55 +0000 (05:24 +0000)]
Make ISO-8859-1 changes to string code, which involved rewriting all
of the case-aware procedures. Many of these were inherited from
primitives, so it was necessary to write Scheme versions of the
primitives.
Chris Hanson [Thu, 20 Sep 2001 18:13:01 +0000 (18:13 +0000)]
To each message shown in a summary buffer, attach a mark pointing at
the line on which the message is shown. Use this in the code that
maps from messages to marks, changing it from (slow) linear time in
the number of messages to constant time.
Chris Hanson [Mon, 10 Sep 2001 00:49:45 +0000 (00:49 +0000)]
Work around bug in "Microsoft Outlook IMO, Build 9.0.2416
(9.0.2910.0)" -- it sends documents of type application/octet-stream
with embedded hard line breaks, which is illegal according to RFC
2045, page 20:
Since the canonical representation of media types other than
text do not generally include the representation of line
breaks as CRLF sequences, no hard line breaks (i.e. line
breaks that are intended to be meaningful and to be displayed
to the user) can occur in the quoted-printable encoding of
such types.
Chris Hanson [Mon, 20 Aug 2001 21:02:43 +0000 (21:02 +0000)]
More changes to .pkd files: now the information about loading and
initialization is separated from the information about environments
and bindings. The load/initialization sequence as written in the .pkg
file is preserved in the .pkd file.
Chris Hanson [Mon, 20 Aug 2001 02:49:18 +0000 (02:49 +0000)]
Change .pkd file to have clearly defined exports and imports that
directly correspond to those written by the programmer in the .pkg
file. This eliminates the duplicate links that were present in the
previous design.
Chris Hanson [Sat, 18 Aug 2001 04:52:33 +0000 (04:52 +0000)]
Add additional information to .pkd file to support automatic evaluation
of package-initialization expressions. Major problems cropped up with
package extensions, which necessitated some redesign of CREF.
Chris Hanson [Thu, 16 Aug 2001 20:59:56 +0000 (20:59 +0000)]
Change "parent-name" field of package description to contain the names
of every ancestor of the package. This is necessary so that we can
replace the use of ".glo" files with ".pkd" files.
Chris Hanson [Thu, 16 Aug 2001 20:59:00 +0000 (20:59 +0000)]
Change "parent-name" field of package description to contain the names
of every ancestor of the package. This is necessary so that we can
replace the use of ".glo" files with ".pkd" files.
Chris Hanson [Thu, 16 Aug 2001 20:58:15 +0000 (20:58 +0000)]
Change "parent-name" field of package description to contain the names
of every ancestor of the package. This is necessary so that we can
replace the use of ".glo" files with ".pkd" files.
Chris Hanson [Thu, 16 Aug 2001 20:02:58 +0000 (20:02 +0000)]
Change "parent-name" field of package description to contain the names
of every ancestor of the package. This is necessary so that we can
replace the use of ".glo" files with ".pkd" files.
Chris Hanson [Wed, 15 Aug 2001 03:34:18 +0000 (03:34 +0000)]
Implement completely new format for compiled package descriptions.
Old format was code to build the packages and load files into them.
New format is a summary description of the packages, which is
interpreted at run time to provide the same functionality.
The purpose of this change is to support uninstallation and
replacement of packages. The new compiled package descriptions are
general enough for this purpose.
Chris Hanson [Wed, 15 Aug 2001 03:10:42 +0000 (03:10 +0000)]
Implement completely new format for compiled package descriptions.
Old format was code to build the packages and load files into them.
New format is a summary description of the packages, which is
interpreted at run time to provide the same functionality.
The purpose of this change is to support uninstallation and
replacement of packages. The new compiled package descriptions are
general enough for this purpose.
Chris Hanson [Wed, 15 Aug 2001 03:00:01 +0000 (03:00 +0000)]
Implement completely new format for compiled package descriptions.
Old format was code to build the packages and load files into them.
New format is a summary description of the packages, which is
interpreted at run time to provide the same functionality.
The purpose of this change is to support uninstallation and
replacement of packages. The new compiled package descriptions are
general enough for this purpose.
Chris Hanson [Fri, 10 Aug 2001 17:29:18 +0000 (17:29 +0000)]
Redesign compiled-code debugging information so that it is keyed. The
key in the .com file and the key in the .bci file must match, or the
.bci file will be ignored.
Chris Hanson [Fri, 10 Aug 2001 17:11:15 +0000 (17:11 +0000)]
Redesign compiled-code debugging information so that it is keyed. The
key in the .com file and the key in the .bci file must match, or the
.bci file will be ignored.
Chris Hanson [Tue, 7 Aug 2001 01:27:17 +0000 (01:27 +0000)]
Implement primitive to allow linking variables that have different
names. (Previously, linked variables were required to have the same
name.) This requires a major redesign of the cache data structures,
and since the garbage collector knows about these structures, all of
the garbage-collector files are affected too. The new data structures
have slightly different space requirements: a cache with no references
uses one word less than previously, while a cache with references uses
three words more than previously (independent of the number of
references).