Artifact 1c86f257f77db6bc540a8308bbe23af6586f29700249e2808ac1edf3c83e8c9c:
- File
psl-1983/x-psl/bug-fix.log
— part of check-in
[eb17ceb7f6]
at
2020-04-21 19:40:01
on branch master
— Add Reduce 3.0 to the historical section of the archive, and some more
files relating to version sof PSL from the early 1980s. Thanks are due to
Paul McJones and Nelson Beebe for these, as well as to all the original
authors.git-svn-id: https://svn.code.sf.net/p/reduce-algebra/code/historical@5328 2bfe0521-f11c-4a00-b80e-6202646ff360 (user: arthurcnorman@users.sourceforge.net, size: 10406) [annotate] [blame] [check-ins using] [more...]
Bug: Compress doesn't work on a list of ASCII values Fix: Make it call Lisp2Char on each element of the list. By: Eric Date: 4:51pm Tuesday, 12 October 1982 Source: PI:EXPLODE-COMPRESS.RED Module: IO in kernel Remarks: The numbers 0..9 no longer work the same as !0..!9 Bug: In open-coded arithmetic/vector access. Fix: Added (USESDEST USESDEST) clause to ASSOCPAT By: Eric Date: 11:10am Monday, 11 October 1982 Source: P20C:DEC20-COMP.RED and PVC:VAX-COMP.RED; P68C:M68K-COMP.RED should also be changed. Module: DEC20-COMP and VAX-COMP Remarks: Bug: Catch no longer needed in Read, due to change in EOF handling Fix: Removed CATCH($READ$, ChannelRead IN*) from READ, similarly for COMPRESS and IMPLODE By: Eric Date: 1:31pm Friday, 8 October 1982 Source: PI:READ.RED, PI:EXPLODE-COMPRESS.RED Module: IO in kernel Remarks: Bug: There is no EXPR for reading files Fix: Change DSKIN from a FEXPR to a one argument EXPR By: Eric Date: 12:14pm Tuesday, 5 October 1982 Source: PI:DSKIN.RED Module: IO in kernel Remarks: This change is incompatible for those using DSKIN with multiple arguments. These uses will have to change to multiple DSKINs. Feature: The printing functions use the variables PRINLEVEL and PRINLENGTH, as described in the Common Lisp Manual. By: Eric Date: 12:12pm Tuesday, 5 October 1982 Source: PI:PRINTERS.RED Module: IO in kernel Remarks: Bug: BIGNUM quotient, re-evaluation errors Fix: Improved BIGNUM and BIGFACE installed By: M. L. Griss, for A. C. Norman Date: 4 October 1982. Source: PU:BIGFACE.RED, PU:BIGBIG.RED Module: BIGNUM Remarks: Some errors still remain, in BLDIFF, etc. and minor typo's fixed. Bug: Scantable in POLY was inherited from CURRENTSCANTABLE!* not "nice" under PSL Fix: Added an ALGSCANTABLE!*, similar to RLISP table By: Martin Date: 3:41pm Tuesday, 28 September 1982 Source: PU:POLY.RED Module: POLY Remarks: Bug: (REMAINDER (RANDOM) n) wasnet good for 3,7 or 11 Fix: Defined RandomModulus variable and RANDOMMOD(N) function By: Martin Date: 3:38pm Tuesday, 28 September 1982 Source: PU:mathlib.red Module: MATHLIB Remarks: Maybe just a "quick" fix and needs further examination Bug: CopyStringToFrom wasn't safe Fix: Make it safe By: Cris Date: 10:37am Tuesday, 28 September 1982 Source: PI:COPIERS.RED Module: kernel Remarks: Bug: *THROW wasn't restoring the outer variable bindings Fix: Call on RestoreEnvironment. By: Eric Date: 8:55am Monday, 27 September 1982 Source: PI:CATCH-THROW.RED Module: EVAL in kernel Remarks: Bug: PRINTX in DEBUG didn't handle circular vectors. Fix: Now it does. By: Eric Date: 5:44pm Friday, 24 September 1982 Source: PU:DEBUG.RED Module: DEBUG Remarks: Also made DEBUG use CODE-NUMBER-OF-ARGUMENTS to find out the # of arguments to a compiled function. Feature: The printing function for code pointers prints the number of arguments expected, in the format #<Code 3 284313>, where 3 is the # of arguments and 284313 is the address. The address part is now printed in the "preferred" radix of the machine, defined by the WConst CompressedBinaryRadix, which is 8 on the Dec-20 and Cray, and 16 on the Vax, 68000, and 360. By: Eric Date: 5:38pm Friday, 24 September 1982 Source: PI:PRINTERS.RED and PXX:GLOBAL-DATA.RED (for constant definition) Module: IO in kernel Remarks: Bug: No way to find out how many arguments a compiled function gets. Fix: Put a header above the entry point with the # of arguments, accessed by the function CODE-NUMBER-OF-ARGUMENTS, which expects a code pointer as its argument and returns the number of arguments the code pointer expects, or NIL. By: Eric Date: 5:17pm Friday, 24 September 1982 Source: PC:PASS-1-LAP.SL (to add header word), PC:DATA-MACHINE.RED (to define access macro), PI:PUTD-GETD.RED (to define callable entry point). Module: PASS-1-LAP, SYSLISP, PROP in kernel Remarks: Only functions compiled since this change have the header word; old FASL files will have to be recompiled to make use of this feature. Bug: IDs (symbols) are not garbage collected. Fix: Allocate symbols as a free list linked through the name cell By: Eric Date: 5:02pm Friday, 24 September 1982 Source: PI:COPYING-GC.RED, PI:COMPACTING-GC.RED, PI:ALLOCATORS.RED, PC:LAP-TO-ASM.RED Module: LAP-TO-ASM, ALLOC in kernel Remarks: Bug: "FOO not compiled" messages in compiler are still unclear. Fix: Now says "Value of FOO not used, therefore not compiled", or "Top level FOO in (FOO BAR) not used, therefore not compiled" By: Eric Date: 11:43am Monday, 20 September 1982 Source: PC:COMPILER.RED Module: COMPILER Remarks: Bug: Printing {99} in ERROR is only noise. Fix: Only print message, don't print number By: Eric Date: 11:32am Monday, 20 September 1982 Source: PI:ERROR-ERRORSET.RED and PI:ERROR-HANDLERS.RED Module: ERROR in kernel Remarks: Bug: Unmatched right paren in a file is not an error. Fix: Only allow an unmatched right paren from the terminal By: Eric Date: 11:26am Monday, 20 September 1982 Source: PI:READ.RED Module: IO Remarks: Bug: CAR of a form is sometimes evaluated; compiler and Eval do not agree. Fix: CAR of a form is NEVER evaluated; only a LAMBDA form or globally defined function name is allowed. By: Eric Date: 10:41am Monday, 20 September 1982 Source: PC:COMPILER.RED and PI:EVAL-APPLY.RED Module: EVAL in kernel, and COMPILER Remarks: Bug: Backtrace is not very helpful Fix: Suppress printing of interpreter functions; better formatting By: Eric Date: 10:24am Monday, 20 September 1982 Source: PI:BACKTRACE.RED Module: EXTRA Remarks: It's still not too hot. Bug: The prettyprinter is weak, and conses a lot. Fix: Use the IMSSS prettyprinter, with a few modifications. By: Eric Date: 9:27am Monday, 20 September 1982 Source: Added PU:PRETTYPRINT.SL and PU:PRETTYPRINT.BUILD. Deleted PU:PRETTY.RED and PU:PRETTY.BUILD. Changed PI:AUTOLOAD.RED Module: Removed PRETTY, added PRETTYPRINT, changed FASL in kernel Remarks: Bug: Not all I/O functions have channel-specific counterparts Fix: Added ChannelTerPri, ChannelLineLength, ChannelPosn, ChannelEject ChannelReadCH, ChannelPrint, ChannelPrin2T, ChannelSpaces ChannelTab, ChannelSpaces2, ChannelPrinC By: Eric Date: 4:21pm Friday, 17 September 1982 Source: on PI: PRINTF.RED, OTHER-IO.RED, EASY-SL.RED, EASY-NON-SL.RED Module: IO and RANDM, in kernel Remarks: Bug: DO with no return forms returns T instead of NIL Fix: Typo in DO, DO*, DO-LOOP, DO-LOOP*, ((null (cdr result) nil)) ==> ((null (cdr result)) nil) By: Eric Date: 5:09pm Wednesday, 15 September 1982 Source: PU:ITER-MACROS.SL Module: USEFUL Remarks: Bug: Token scanner won't read 1+ and 1- as symbols Fix: Patch in ChannelReadToken By: Eric Date: 11:01am Wednesday, 15 September 1982 Source: PI:TOKEN-SCANNER.RED Module: IO in kernel Remarks: Still doesn't scan -1+ as a symbol Bug: InternP doesn't work for strings Fix: Checks to see if a symbol with that pname is interned By: Eric Date: 9:36am Wednesday, 15 September 1982 Source: PI:OBLIST.RED Module: SYMBL in kernel Remarks: Bug: (igetv (igetv x 5) y) generates bad code Fix: Add USESDEST clause to ASSOCPAT in xxx-COMP.RED By: Eric Date: 2:11pm Monday, 13 September 1982 Source: P20C:DEC20-COMP.RED and PVC:VAX-COMP.RED (Should also be done to P68C:M68K-COMP.RED). Module: DEC20-COMP and VAX-COMP Remarks: Bug: in EXP Fix: Changed 2**N to 2.0**N By: Eric Date: 8:50am Monday, 13 September 1982 Source: PU:MATHLIB.RED Module: MATHLIB Remarks: Bug: APPLY(x, list(1,2,3,4,5,6)) doesn't avoid consing Fix: Add a PA1FN for APPLY so that !&PaList isn't applied to the 2nd arg By: Eric Date: 4:26pm Friday, 10 September 1982 Source: PC:COMPILER.RED and PC:COMP-DECLS.RED Module: COMPILER, COMP-DECLS Remarks: Bug: Compiler error and warning messages are confusing Fix: Use more English, always print the function name By: Eric Date: 9:54am Friday, 10 September 1982 Source: PC:COMPILER.RED Module: COMPILER Remarks: Bug: FLUID and MACRO can't have the same name Fix: Use indicator VARTYPE for variables, instead of sharing TYPE with functions. By: Eric Date: 9:16am Friday, 10 September 1982 Source: PI:FLUID-GLOBAL.RED Module: PROP in kernel Remarks: Bug: DUMPLISP blows away the last page of the stack in rare cases on the 20 Fix: Add some slack in the call to UNMAP-SPACE from DUMPLISP By: Eric Date: 10:24am Friday, 3 September 1982 Source: P20:DUMPLISP.RED Module: EXTRA Remarks: Bug: WNOT was not caught by constant folding Fix: Added PA1REFORMFN = &DOOP for WNOT By: Eric Date: 9:47am Friday, 3 September 1982 Source: PC:COMP-DECLS.RED Module: COMP-DECLS Remarks: Bug: CHAR-UPCASE and CHAR-DOWNCASE returned NIL instead of their arguments if the function didn't modify them. Fix: Return the argument instead By: Eric Date: 2:25pm Thursday, 2 September 1982 Source: PU:CHARS.LSP Module: CHARS Remarks: Bug: Right parens cause an error at the top level Fix: Make ) a read macro to be ignored outside of list reading By: Eric Date: 2:08pm Thursday, 2 September 1982 Source: PI:READ.RED Module: IO in kernel Remarks: Bug: PSL-SAVE.CTL requires that you are connected to P20: Fix: add a logical name definition def DSK: DSK:,P20: By: Eric Date: 1:35pm Thursday, 2 September 1982 Source: P20:PSL-SAVE.CTL Module: None Remarks: Bug: XJsysError and JSYS constants are wrong Fix: Fixed. By: Eric Date: 1:28pm Thursday, 2 September 1982 Source: P20:20-INTERRUPT.RED Module: INTERRUPT Remarks: Bug: MACROEXPAND does not handle multiple argument SETQ Fix: Removed MACROEXPAND-SETQ, use MACROEXPAND-RANDOM instead By: Eric Date: 10:33am Thursday, 2 September 1982 Source: PU:MACROEXPAND.SL Module: USEFUL Remarks: Bug: Functions in Mathlib call REDERR which is only defined in Rlisp Fix: Have them call StdError instead By: Eric Date: 9:20am Thursday, 2 September 1982 Source: PU:MATHLIB.RED Module: MATHLIB Remarks: Bug: Prettyprint returns its argument, which is worse than useless Fix: Make it return NIL instead By: Eric Date: 9:15am Thursday, 2 September 1982 Source: PU:PRETTY.RED Module: PRETTY Remarks: Bug: ContError does not handle atoms as the ReEvalForm Fix: Now it does. By: Eric Date: 9:11am Thursday, 2 September 1982 Source: PI:CONT-ERROR.RED Module: MACRO in kernel Remarks: Bug: (QUOTE x y) is incorrectly printed Fix: Change ChannelPrintPair so that only (QUOTE x) prints as 'x By: Eric Date: 8:59am Thursday, 2 September 1982 Source: PI:PRINTERS.RED Module: IO in kernel Remarks: