Stephen Adams [Fri, 19 May 1995 03:41:26 +0000 (03:41 +0000)]
Added call to DBG-INFO/REMEMBER.
Stephen Adams [Thu, 18 May 1995 20:34:21 +0000 (20:34 +0000)]
Changed open coding to be controlled by the size of the expression and
a limiting variable *SIMPLIFY/OPEN-CODE-EXPRESSION-LIMIT* which must
be a (small) integer to enable the open coding. The limiting varibale
roughtly counts the number of extra LOOKUPs or QUOTEs (including
operators) that will be tolerated.
Stephen Adams [Wed, 17 May 1995 03:37:46 +0000 (03:37 +0000)]
Initial revision
Stephen Adams [Wed, 17 May 1995 03:11:24 +0000 (03:11 +0000)]
Added wttree.scm.
Chris Hanson [Tue, 16 May 1995 09:21:28 +0000 (09:21 +0000)]
Add primitives to create and manage menus.
Stephen Adams [Tue, 16 May 1995 05:20:59 +0000 (05:20 +0000)]
Added wttree.
Stephen Adams [Tue, 16 May 1995 05:19:31 +0000 (05:19 +0000)]
*** empty log message ***
Stephen Adams [Tue, 16 May 1995 04:43:48 +0000 (04:43 +0000)]
Merged split-typecode and old versions.
Stephen Adams [Tue, 16 May 1995 02:43:14 +0000 (02:43 +0000)]
Tidying.
Stephen Adams [Sun, 14 May 1995 01:17:55 +0000 (01:17 +0000)]
Added expansions for multiplication by a constant fixnum and some hair
for (EXPT -1.0 <fixnum>) and (EXPT 2 <fixnum>)
Stephen Adams [Sun, 14 May 1995 00:53:04 +0000 (00:53 +0000)]
Added %car %cdr %set-car! %set-cdr! %vector-ref %vector-set!
%vector-length unchecked pair and vector operators.
Stephen Adams [Sun, 14 May 1995 00:52:41 +0000 (00:52 +0000)]
Added %car %cdr %set-car! %set-cdr! %vector-ref %vector-set!
%vector-length unchecked pair and vector operators.
Added *RTLGEN/PRE-LOAD-STACK-FRAME?*.
When #T behaves as before.
When #F it keeps stack frame elements in stack instead of loading them
into registers (which sometimes causes register spills!). Not
completed so dont set this flage to #F yet.
Stephen Adams [Sat, 13 May 1995 19:01:38 +0000 (19:01 +0000)]
Fixed prefix nodes to compose with highlighted nodes.
Chris Hanson [Sat, 13 May 1995 11:10:10 +0000 (11:10 +0000)]
Don't use WinQueryFocus to decide whether the cursor has been created.
There is a short period of time in which it answers "yes" when in fact
the answer is "no", which can result in errors from WinShowCursor.
Chris Hanson [Fri, 12 May 1995 09:30:55 +0000 (09:30 +0000)]
OS/2 directory-listing procedures must canonicalize filename case,
because filename completion code depends on case-sensitive string
matching operations.
Stephen Adams [Thu, 11 May 1995 16:13:54 +0000 (16:13 +0000)]
Tidying.
Stephen Adams [Thu, 11 May 1995 14:06:13 +0000 (14:06 +0000)]
Removed useless definitions.
Stephen Adams [Thu, 11 May 1995 05:39:59 +0000 (05:39 +0000)]
Added ATAN.
Chris Hanson [Wed, 10 May 1995 21:27:03 +0000 (21:27 +0000)]
Fix random editing error.
Chris Hanson [Wed, 10 May 1995 21:19:49 +0000 (21:19 +0000)]
Patch to work around sporadic error code in DosRequestMutexSem.
Chris Hanson [Wed, 10 May 1995 21:15:57 +0000 (21:15 +0000)]
Patch to work around bugs in OS/2 PM API cursor procedures.
Chris Hanson [Wed, 10 May 1995 20:48:39 +0000 (20:48 +0000)]
Fix think-o in recent change.
Chris Hanson [Wed, 10 May 1995 20:26:21 +0000 (20:26 +0000)]
Fix think-o in recent change.
Stephen Adams [Wed, 10 May 1995 03:33:08 +0000 (03:33 +0000)]
Changed some LAMBDAs in the arity-dispatched operators to
NAMED-LAMBDAs for PP clarity.
Chris Hanson [Tue, 9 May 1995 10:32:36 +0000 (10:32 +0000)]
Fix two small bugs in the paragraph-parsing code.
Chris Hanson [Sun, 7 May 1995 05:54:12 +0000 (05:54 +0000)]
In the console readahead thread, when getch() returns EOF, send an
empty readahead message to inform Scheme that the console channel has
been closed.
Chris Hanson [Sun, 7 May 1995 05:53:20 +0000 (05:53 +0000)]
Pay attention to the console_closedp flag, and take appropriate action
when it is set.
Chris Hanson [Sun, 7 May 1995 05:52:34 +0000 (05:52 +0000)]
When WM_QUIT is posted to the PM message queue, kill the Scheme
program. This normally happens at one of two times: (1) the user
clicks on "close" in the pop-up menu in the window list, and (2) the
system is shutting down.
Chris Hanson [Sat, 6 May 1995 23:01:05 +0000 (23:01 +0000)]
Implement support for fixed-width outline fonts.
Stephen Adams [Sat, 6 May 1995 20:57:08 +0000 (20:57 +0000)]
Tidying and adding integration declarations.
Stephen Adams [Sat, 6 May 1995 18:28:45 +0000 (18:28 +0000)]
Fixed typo.
Stephen Adams [Sat, 6 May 1995 18:25:35 +0000 (18:25 +0000)]
Code for BEGIN now generates simpler code for sequences fo `simple'
expressions.
Stephen Adams [Sat, 6 May 1995 18:04:53 +0000 (18:04 +0000)]
Changed BEGINNIFY to work incrementally, i.e. assume that BEGIN
subforms have been BEGINNIFY-ied.
Chris Hanson [Sat, 6 May 1995 02:22:02 +0000 (02:22 +0000)]
Add code to do news-group posting.
Chris Hanson [Fri, 5 May 1995 23:15:47 +0000 (23:15 +0000)]
Change definition of MAIL-INSERT-FIELD to make it more general.
Chris Hanson [Fri, 5 May 1995 22:53:18 +0000 (22:53 +0000)]
If "To:" header field not given to MAIL-SETUP, don't insert one.
Chris Hanson [Fri, 5 May 1995 22:35:09 +0000 (22:35 +0000)]
Reorganize mail sending code into more general pieces, then export the
pieces for use by other programs.
Chris Hanson [Fri, 5 May 1995 22:32:44 +0000 (22:32 +0000)]
Add OS/HOSTNAME procedure to get a host name suitable for use in an
email address. (Implemented for unix and OS/2.)
Stephen Adams [Fri, 5 May 1995 12:58:36 +0000 (12:58 +0000)]
Added DBG-USE structure.
Stephen Adams [Fri, 5 May 1995 12:57:56 +0000 (12:57 +0000)]
Tidying.
Stephen Adams [Fri, 5 May 1995 12:57:26 +0000 (12:57 +0000)]
Made debugging info handling code be more graceful with badly formed
dbg info.
Stephen Adams [Fri, 5 May 1995 12:55:51 +0000 (12:55 +0000)]
Fiddled with kmp printing.
Chris Hanson [Fri, 5 May 1995 07:20:41 +0000 (07:20 +0000)]
Change MAKE-MAIL-BUFFER to return the buffer it makes. If an old
buffer exists and the user wants to keep it, return #F instead. Also
make the SELECTOR argument optional, and allow it to be #F meaning
that the buffer will not be selected but just returned (the default).
Change the MAIL-SETUP-HOOK variable so that it is invoked with the
mail buffer as an argument.
Chris Hanson [Fri, 5 May 1995 06:53:05 +0000 (06:53 +0000)]
Change MAKE-MAIL-BUFFER to return the buffer it makes. If an old
buffer exists and the user wants to keep it, return #F instead. Also
make the SELECTOR argument optional, and allow it to be #F meaning
that the buffer will not be selected but just returned (the default).
Change the MAIL-SETUP-HOOK variable so that it is invoked with the
mail buffer as an argument.
Chris Hanson [Fri, 5 May 1995 06:25:50 +0000 (06:25 +0000)]
Fix bug: select-buffer-hook not being invoked-with correct arguments.
Chris Hanson [Thu, 4 May 1995 07:06:12 +0000 (07:06 +0000)]
Do CRLF translation when cutting to and pasting from the clipboard.
Stephen Adams [Wed, 3 May 1995 21:37:11 +0000 (21:37 +0000)]
Changed SHOW-TIME to break process time into non-gc and gc times.
Chris Hanson [Wed, 3 May 1995 07:51:01 +0000 (07:51 +0000)]
Initial revision
Chris Hanson [Wed, 3 May 1995 07:50:39 +0000 (07:50 +0000)]
Add auto-loaded news reader.
Chris Hanson [Wed, 3 May 1995 07:35:56 +0000 (07:35 +0000)]
Initial revision
Chris Hanson [Wed, 3 May 1995 07:34:40 +0000 (07:34 +0000)]
Add new file implementing operations on ordered vectors.
Chris Hanson [Tue, 2 May 1995 22:30:32 +0000 (22:30 +0000)]
Integrate support for OS-dependent cut&paste buffer into the kill and
yank commands, as is done in Emacs 19. This support is available in
the OS/2 implementation.
Chris Hanson [Tue, 2 May 1995 21:47:43 +0000 (21:47 +0000)]
Integrate support for OS-dependent cut&paste buffer into the kill and
yank commands, as is done in Emacs 19. This support is available in
the OS/2 implementation.
Chris Hanson [Tue, 2 May 1995 21:19:22 +0000 (21:19 +0000)]
Integrate support for OS-dependent cut&paste buffer into the kill and
yank commands, as is done in Emacs 19. This support is available in
the OS/2 implementation.
Chris Hanson [Tue, 2 May 1995 20:58:41 +0000 (20:58 +0000)]
Add primitives to read and write the OS/2 clipboard.
Chris Hanson [Tue, 2 May 1995 20:54:09 +0000 (20:54 +0000)]
Add primitives to read and write the OS/2 clipboard. Also implement
Shift-Insert command on the console window to allow pasting from the
clipboard.
Chris Hanson [Tue, 2 May 1995 20:49:06 +0000 (20:49 +0000)]
If a file insertion doesn't insert any data, don't do any of the
post-processing that sets the modification bit, etc.
Chris Hanson [Tue, 2 May 1995 00:40:13 +0000 (00:40 +0000)]
Previously a POP password was discarded any time an error happened
while fetching mail from the POP server. Now, once mail has been
fetched successfully, the password is kept regardless of any errors.
Chris Hanson [Tue, 2 May 1995 00:29:23 +0000 (00:29 +0000)]
If popclient program is successful, kill buffer showing its output.
Stephen Adams [Mon, 1 May 1995 03:36:18 +0000 (03:36 +0000)]
Moved call to GRAPH/INTERESTING to before call to GRAPH/CLEANUP!
Stephen Adams [Mon, 1 May 1995 02:43:24 +0000 (02:43 +0000)]
Removed 2 old fg related bindings.
Stephen Adams [Sun, 30 Apr 1995 17:22:37 +0000 (17:22 +0000)]
Added a hand-crafted vector-8b based bit-string operation for a fast
connectivity predicate when the in-degree exceeds a certain value.
Speeds up dataflow of large graphs.
Stephen Adams [Sun, 30 Apr 1995 15:09:15 +0000 (15:09 +0000)]
Changed COMPUTE-KEY-HASH to compute hash value optimistically and only
go into the SET-INTERRUPT-ENABLES! guarded loop if, as a post-check,
the table needs rehashing.
On a test case (the new compiler on list.scm) this reduces the
profiled time in SET-INTERRUPT-ENABLES! from 9.9% to 3.4%
Chris Hanson [Sun, 30 Apr 1995 07:12:12 +0000 (07:12 +0000)]
Repaginate.
Chris Hanson [Sun, 30 Apr 1995 07:02:18 +0000 (07:02 +0000)]
Change select-buffer hooks to take a second argument, the window that
the buffer is being selected in. This is often useful, and if it's
not passed in there's no way to get it.
Chris Hanson [Sun, 30 Apr 1995 06:56:13 +0000 (06:56 +0000)]
Add optional argument to WINDOW-DELETE! procedure. This argument is a
window that should get the space released by the deleted window. If
it is not adjacent to the deleted window, the argument is ignored.
Chris Hanson [Sun, 30 Apr 1995 06:54:43 +0000 (06:54 +0000)]
* Generalize and simplify sendmail interface, to give more control
over the configuration of a mail buffer by a program.
* Restructure parts of the RMAIL code that deal with replies and
output to mail files. Export some of this code for use by other
programs.
Stephen Adams [Sat, 29 Apr 1995 22:26:36 +0000 (22:26 +0000)]
Fixed bug when a coerced call site has been transformed into a binder
for another coerced procedure.
Stephen Adams [Sat, 29 Apr 1995 14:09:09 +0000 (14:09 +0000)]
(LOAD-OPTION name #!optional no-error?) now returns #F when NO-ERROR?
is true and the option NAME is not available. It used to return NAME
in all cases.
Stephen Adams [Sat, 29 Apr 1995 13:57:56 +0000 (13:57 +0000)]
Fixed RTLGEN/INTEGER-CONSTANT? to reject non-integers.
Stephen Adams [Sat, 29 Apr 1995 13:18:05 +0000 (13:18 +0000)]
Added expansion for SYMBOL?
Stephen Adams [Sat, 29 Apr 1995 13:08:29 +0000 (13:08 +0000)]
Added expansion for SYMBOL? (similar to those for numeric predicates).
Stephen Adams [Sat, 29 Apr 1995 01:11:40 +0000 (01:11 +0000)]
Tidying.
Stephen Adams [Sat, 29 Apr 1995 01:05:08 +0000 (01:05 +0000)]
Tidying.
Stephen Adams [Sat, 29 Apr 1995 00:57:30 +0000 (00:57 +0000)]
Tidying.
Stephen Adams [Fri, 28 Apr 1995 13:47:22 +0000 (13:47 +0000)]
Added some (incomplete) documantation about garbage collection.
Stephen Adams [Fri, 28 Apr 1995 12:35:59 +0000 (12:35 +0000)]
Yet another case where a BODY ended in a DEFINITION instead of an
EXPRESSION.
Chris Hanson [Fri, 28 Apr 1995 07:20:08 +0000 (07:20 +0000)]
Eliminate compiler warning.
Chris Hanson [Fri, 28 Apr 1995 07:05:10 +0000 (07:05 +0000)]
* The procedure OS_set_working_dir_pathname in "os2env.c" had a
serious bug: passing it a string like "d:\\" caused it to signal a
Scheme error. What was serious was that this could occur while
spawning a subprocess, which for some as-yet unexplained reason
would crash Scheme with no error message. To compound matters, the
bug appeared to be unreproduceable under a debugger (not true). The
remaining changes were side effects of the blind debugging process:
* Added OS2_initialize_early to "boot.c" and "os2top.c". This allows
the microcode to do the initialization necessary to put up dialog
boxes. Which in turn is needed to report boot-time errors, such as
non-existent band file names.
* "option.c" calls outf but did not include "outf.h". Under ANSI C,
since outf takes a variable number of arguments, this results in
unspecified behavior. Under OS/2 with IBM's compiler, this causes a
mismatch in calling conventions, which can result in bizarre errors.
* Modified "os2.c" to have an alternate implementation of OS_malloc
and related procedures. This was generated while attempting to
track down a nasty bug that appeared to be a memory-allocation
error. The alternate implementation is #ifdef'ed out, but left in
place should it be useful in the future.
* Improved the definition of OS2_logic_error_1 so that it treats
errors in non-essential subthreads as non-fatal. Also improved the
wording of the error messages produced, and included more
information for future reference.
* All subthreads now have exception handlers and PM message queues, so
they can catch and report any exceptions that occur in them.
Previously, an exception occurring in a subthread would crash Scheme
with no explanation.
* Enhanced the acknowledgement messages sent to a channel readahead
thread to carry a flag indicating whether the thread is being asked
to read some more or to kill itself. This eliminates a race
condition associated with closing of the channel.
* Changed the input-pipe readahead thread to accept its arguments in a
freshly malloc'ed structure that the thread is responsible for
freeing. Previously, the arguments were accepted in a channel,
which contained a context structure which in turn contained the
arguments. The previous implementation had a race condition in that
the argument data structures were owned by the Scheme thread, and
could be deallocated before the readahead thread had a chance to
dereference them.
* In "os2proc.c", moved the allocation of the structure representing
the child process into the locked region of the process
initialization. This is necessary because the child-wait thread
accesses the same structures as the allocator.
* In "os2proc.c", the working directory was being saved and later
restored. But this was a static string, and there weren't good
guarantees as to it retaining its value -- so now it is copied
before being stored.
* When putting up a message box for a fatal error, the buffer used to
compose the message was static. This caused garbling when the two
threads attempted to produce fatal errors at the same time (this was
occurring regularly during the debugging process). Now the message
box uses a per-thread static buffer.
Chris Hanson [Fri, 28 Apr 1995 07:01:32 +0000 (07:01 +0000)]
Primitive OS2WIN-GET-EVENT was sometimes calling OS_free with a null
pointer. This is legal with IBM's C library, but is not something I
like to rely on.
Chris Hanson [Fri, 28 Apr 1995 06:59:29 +0000 (06:59 +0000)]
In primitive OS2-SELECT-REGISTRY-TEST, there was a window where input
could be waiting to be read, but the primitive would be blocked
waiting for some input. This was because the primitive is composed of
two phases, and input detected in the first phase was not used to
prevent blocking in the second.
Chris Hanson [Fri, 28 Apr 1995 06:45:37 +0000 (06:45 +0000)]
Readahead message types belong to the channel-thread implementation
and are already initialized there. The initialization here was an
unnecessary duplicate.
Stephen Adams [Fri, 28 Apr 1995 00:29:13 +0000 (00:29 +0000)]
Added more SCode<->binding relations.
Stephen Adams [Fri, 28 Apr 1995 00:01:21 +0000 (00:01 +0000)]
Added the SCode expressions of subproblems to the debugging info. We
might want to use this in the debugger to get some kind of subproblem
history for compiled code.
Stephen Adams [Thu, 27 Apr 1995 23:23:18 +0000 (23:23 +0000)]
Converted to new dbg-info scheme.
Stephen Adams [Thu, 27 Apr 1995 23:22:00 +0000 (23:22 +0000)]
Converted to new dbg-info scheme. Rather than model the environmnet
at every stage, we keep a collection of all the micro-transformations
that occur. At the very end we will have to reconstruct the
envrionment from the available information.
Stephen Adams [Thu, 27 Apr 1995 23:16:34 +0000 (23:16 +0000)]
Added DBG-INFO/REMEMBER.
Chris Hanson [Thu, 27 Apr 1995 20:33:28 +0000 (20:33 +0000)]
Implement some renames that have been nagging for some time:
current-buffer => selected-buffer
current-buffer? => selected-buffer?
current-window => selected-window
current-window? => selected-window?
With these renames, the accessors now have names that match the
selectors. The old names remain as aliases.
This was stimulated by a recent change that accidentally inserted a
reference to one of the new names -- when I misremember the names I
chose, it's time to change the names.
Stephen Adams [Thu, 27 Apr 1995 02:48:47 +0000 (02:48 +0000)]
Added frame size statistic.
Stephen Adams [Wed, 26 Apr 1995 03:23:03 +0000 (03:23 +0000)]
Fiddled with a sequence of definitions and expressions to ensure the
last form is an expression to comply with the new analchecking software.
Stephen Adams [Wed, 26 Apr 1995 01:53:25 +0000 (01:53 +0000)]
Removed old flowgraph code.
Stephen Adams [Wed, 26 Apr 1995 01:51:52 +0000 (01:51 +0000)]
Initial revision
Stephen Adams [Mon, 24 Apr 1995 23:22:30 +0000 (23:22 +0000)]
Added SET-GC-NOTIFICATION!
Stephen Adams [Mon, 24 Apr 1995 23:19:23 +0000 (23:19 +0000)]
Tweaked error checking code to make the integrated portion smaller.
Stephen Adams [Mon, 24 Apr 1995 22:30:44 +0000 (22:30 +0000)]
Fixed TRANSFORM/PACKAGE to always return UNSPECIFIC, which ensures
that the last thing in the generated LET is not a definition.
Stephen Adams [Mon, 24 Apr 1995 22:15:33 +0000 (22:15 +0000)]
Fixed definition ordering bug in macros.scm
Stephen Adams [Mon, 24 Apr 1995 16:07:34 +0000 (16:07 +0000)]
Fixed bugs in multicell rewrites.
Stephen Adams [Mon, 24 Apr 1995 16:06:45 +0000 (16:06 +0000)]
Changed to use multicells. Uses singleton multicells unless there are
many mutated variables, in which case it uses one large multicell.
Ideally it should partition varibales according to their dynamic
extent and create one multcell per extent.
Chris Hanson [Mon, 24 Apr 1995 01:10:07 +0000 (01:10 +0000)]
When doing completing prompt with REQUIRE-MATCH? set true, delay
running completion procedure on final value. First try VERIFY-MATCH?
procedure, and if it succeeds, don't do any completion. This
shouldn't change the behavior since VERIFY-MATCH? must be run anyway,
and if the result is already a match, completion does nothing.
This is helpful when the completion procedure is expensive but
VERIFY-MATCH? is relatively cheap.
Chris Hanson [Sun, 23 Apr 1995 06:08:53 +0000 (06:08 +0000)]
Fix think-o in POP server support.