* Change numeric OPCODES SUBRCALL NN to use the LLSUBRS name
* more opcodes subr# in maikoloadupfns
* even more OPCODES SUBRCALL
* Recover BIGBMAPS definitions dup (but more recent) from LLCOLOR
* IOCHAR: Fix daylight savings time
* TMAX: Y2K fix
Also a little code cleanup, changing default font to TERMINAL from GACHA and making text more legible
* DATABASEFNS, ATBL: DUMPDB with DEFINE-FILE-INFO
New database files will have standard headers, then a little special stuff for LOADDB to synchronize, old database files default to a new interlisp environment.
MAKE-READER-ENVIRONMENT in ATBL extended for easier specification, plus better type-testing.
* Remove duplicate comment
* Eliminate implicit calls to \FILEOUTCHARFN
Also, update DATE to modern readtable (don't know what it does), add LLETHER to EXPORTFILES in FILESETS (may also need the file that exports pup records).
* Further fixups for EXPORTFILES
also fixing/compiling PLAINTEXTSTREAM
* Remove garbage files DATE and PLAINTEXTSTREAM from checkin
Co-authored-by: Larry Masinter <LMM@acm.org>
* FILESETS, TEDITWINDOW, TEDIT-PF-SEE
Add DTDECLARE to EXPORTFILES, fix TEDIT so that READONLY windows and processes are collected
* PRINTFN: Eliminate PMORE
* Use COPYCHARS instead of COPYBYTES when HPRINT is copying from NODIRCORE buffer
* Change HPRINT of non-random-access files to use FORMAT of ultimate destination
Introduced an external format (:TEDIT) for Tedit, initialized TEXTOFD to use it. Parmeterized the window split-window region to stop confusions with modernwindows. TEDIT-SEE starts out the defaultfont for non-Tedit-format files. Restored git-lost edits to COPY.TEXT.TO.IMAGE
For some reason, in going back and forth, the hard link between the versioned and the unversioned got lost, and the unversioned was effectively deleted. I did a copyfile to get things back in order
Keep the \INITSUBRS in sorted order: move (YIELD 210) to the end of
the list.
Change the output of WRITECALLSUBRS so that the guard wraps the whole
subrs.h include file.
Avoid using literal tabs in the output #defines, pad with spaces and
arrange the minimum field widths so that everything lines up better.
Local BOUTCCODE macro replaced by call to \OUTCHAR etc. Previously it called PRINTCCODE probably because that was the encapsulation of NSIN. Now avoid the user-level function call, so it can be broken, go directly to the generic.
* Start loadup using lisp.venuesysout as the base on which to build new
* Change 'starter.sysout' (taken from recent lisp.sysout)
* Fix problem with GREET ordering
Keyaction lists have also been updated so that they contain only character specifications, not a mixture of character codes and specifications.
KEYACTION now tests its KEY argument for validity
New function \KEYNUMBERTONAME, potentially convenient for future keyboard work
* TTYIN: Meta chars defined by Meta charset, not specific value 1
Also, changed \TTYIN.RPEOF to use \FORMATBYTESTREAM generic function so that it does not make assumptions about the stream's external format
* LLKEY: Respecify 535 as "Function,^W" in \ORIGKEYACTIONS
First step in anticipating moving meta and signaling characters to unoccupied parts of the Unicode code space
* Have to commit the build artifacts in order to unstash changes
* CMLSTRING: No change, just upgrade filemap for CL Functions
* LLREAD: A little tighter on CHARCODE.DECODE
Allow hex and unicode in comma-separated specs, but must have 0X or U+ indicator. Don't allow junk characters
* LLKEY: Revert attempt at Meta,x and Function,x
For unknown reason, CHARTABLE isn't declared when LLREAD appears in loadup sequence.
* Unicode: Unicode character-constants weren't properly prefixed
Needed U+ or x0
* THINFILES (again): Add version number for mis-ordered HELP
* ATBL: Change 1,0...1,377 to Meta,0...Meta,377
In anticipation of a future migration of the meta charset.
* TEDITFNKEYS: Convert 1,x 2,x to Meta,x Function,x
In anticipation of eventual deconfliction with Unicode
* FILEBROWSER: remove reference to reader-environment field, add fix-directory-dates
Added a new submenu item under Recompute to make directory dates match the filecreated dates for Medley source/compiled files.
* SEDIT-COMMANDS: 1/2 to Meta/Function in Sedit commands
Codes in the 512+ range are now Function, anticipating eventual code-space translation
* FILEPKG: Consistency of exports.all external format
* Adds fields to reader-environment to remember evaluation forms for reprinting
For package and readtable, not for base or external format. This restores previous package/readtable behavior.
ATBL also is now not radix 8
* THINFILES: now can add extensions/file names to operate on
Not related to other things in this branch, just a useful extension to a simple lispusers package. You can now add dribble as an extension, and things like I-NEW as a name, to make it easy to clean up the tmp/ loadup directory in particular.
* Format implementation functions set a known variable *BYTECOUNTER*
The generic functions deal with updating the application variable.
\INCHAR eliminated in favor of \INCCODE.EOLC to make clear what it does.
OPENSTRINGSTREAM streams have their own format, and the string is always fattened.
READBITMAP doesn't mix character and byte reading
* AOFD: Don't execute \STRINGSTREAM.INIT
This creates a file device that is not used anywhere. The function OPENSTRINGSTREAM provides the functionality that this file device suggests that it would provide, but that functionality seems suspect at best. The function is left in the system for now, probably should be deleted at some point in the future so we don't try to maintain it if we trip over it.
* TTYIN: Fix an ancient coding error
but still doesn't solve the (DIRECTORY ?= problem #402
* LLREAD, FILEIO, XCCS: Improve charcode backing, copychars
Added \BACKCCODE.EOLC that backs up over EOL encoding bytes, simplifies \RSTRING2.
\XCCSBACKCCODE returns T/NIL according to whether it succeeded.
\XCCSOUTCHAR uses IPLUS16 for CHARPOSITION
COPYCHARS makes no assumptions about EOL encoding
But still no solution for #402
* LLREAD, TTYIN.LCOM fix#402
The bug showed up in TTYIN, but it was actually a bad edit in the generic backccode.
TTYIN.LCOM is just a recompile--that had never been done with various new declarations.
it took a long time to figure out how to restore old versions, using the './scripts/restore-versions file'. Now that it's there and tested it should be ok to remove them from new 'git clone' of medley
