--- gforth/BUGS 1994/09/05 17:36:15 1.8 +++ gforth/BUGS 2005/12/27 09:35:45 1.57 @@ -1,32 +1,99 @@ -name> does not take the same argument as e.g. .name. Remedy: add cell+ -before name>, but adapt all uses. anton 23apr94 +You can find more recent bug reports at +http://savannah.gnu.org/bugs/?group=gforth -revealing the same name several times (e.g., by using recursive) -results in redefined messages. anton 28jul94 +gforth.el: indentation does not work right on the first line of a +buffer. anton 27jan95 -[IF] is case-sensitive. anton 2aug94 - -if blocks.fb does not exist, 1 block creates the file, but cannot -read-file from it. Only if the file-id has been created with -open-file, not create-file, read-file works. - anton 6aug94 - -gforth.el: doing a forth-fill-paragraph on the following piece of code -uncomments the first comment line. This may be a bug in -fill-paragraph, but documentation says that a paragraph can start -without prefix (hanging indentation), so I guess it's all right. - -anton 19aug94 - -: bb-shortest-paths recursive { bb bb-from start-path -- } - \ compute the - \ paths from bb-from through bb, start-path is the path from - \ bb-from to bb - - -etags.fs: The TAGS file for the kernal is strange: there are no tags -for some aliases; many constants and variables appear twice. anton -2sep94 - -error-causing word indication no longer works well. Cause: relies on -name being HERE. Future trouble (once CATCH really restores the input -stream): it also relies on SOURCE delivering the line that caused the -error. anton 2sep94 +Conditional compilation continues after the file ends. This is allowed +by the standard (through an ambiguous condition), but the compiler +should at least produce a warning. anton 27jan95 + +QUERY and TIB may behave differently than some may expect. TIB always +points to the current SOURCE, and QUERY puts its result there. anton +28nov96 + +open-path-file expands "./" into the sourcefilename. It should either +not expand "./" or provide a mechanism that allows the application to +determine what "./" should expand to. anton 16jun98 + +gforth.el: Typing does not work when tabs separate words in a +line, and there is a \-comment at the end of the line. anton 19feb00 + +SEE does not work when the control structure is too complex (e.g., +load http://www.complang.tuwien.ac.at/forth/pentomino.fs and then do +SEE NEXT-PIECE). anton 5mar2000 + +Include cannot handle lines longer than 255 characters. anton 4sep00 + +Errors happening during a LOAD do not report the offending word and +its context (e.g., the 64-byte line). anton 8sep00 + +Our ecvt routine apparently does not work correctly for Infs and NaNs. +Try "ac_cv_func_ecvt=no ./configure; make" and then in Gforth: "1e 0e +f/ f. 0e 0e f/ f.". anton 25sep00 + +Our ecvt routine does not round correctly, e.g., 0.25->0.3. Marcel +Hendrix 3oct00 <8rdcmd$j96$1@news.IAEhv.nl> + +Newline has only LF (instead of CRLF) in DOS. Bruce Hoyt 25oct2000 +<39f7b14b$2@clear.net.nz> + +F. does not print trailing zeroes (e.g., "10 SET-PRECISION 125e f.") +anton 31may01 + +Block 0 does not work as it should: "0 block drop update save-buffers" +does not write to the blocks file. "0 block 1024 dump" seems to give +the previous contents of the buffer. Travis Bemann 10jul2001 +<3b4b4f57$0$42883$272ea4a1@news.execpc.com> + +When accessing a block beyond the end of the block file, the result is +filled with spaces (this is also documented). However, when accessing +a previously unwritten block before the end of the block file, we will +get a block full of zeroes on most (all?) OSs. This inconsistency +should be eliminated and the documentation fixed. anton 14jul2001 + +Path handling may have a bug: +strace $FORTH 'fpath= .|~+/gforth s" /home/anton/threading-variations/gforth/mach32b.fs" include gforth/kernel/main.fs bye' +tries to open gforth/kernel/gforth/kernel/aliases.fs, then +gforth/gforth/kernel/aliases.fs, but not gforth/kernel/aliases.fs. +anton 3 sep 2001 + +gforth-native does not deal correctly with inline arguments in static +superinstructions: when using lit_lit, the wrong literals are put into +the native code (probably due to wrong assumptions about argument +offsets). 6 nov 2003 + +finish-code changes the stack for gforth-fast with --enable-force-reg +(probably for gforth-native too). Apparently ecx (spTOS) is neither +saved by engine, nor by finish_code(). + +Wrong decompilation: +create foo +5 constant bar +: xxx foo bar ; +simple-see xxx +16 nov 2005 + +Debian's gforth's etags.fs crashes (probably due to path problems) +on b3: +gforth etags.fs +16 nov 2005 + +----- +Copyright (C) 1995,1996,1997,1998,2000,2003 Free Software Foundation, Inc. + +This file is part of Gforth. + +Gforth is free software; you can redistribute it and/or +modify it under the terms of the GNU General Public License +as published by the Free Software Foundation; either version 2 +of the License, or (at your option) any later version. + +This program is distributed in the hope that it will be useful, +but WITHOUT ANY WARRANTY; without even the implied warranty of +MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.#See the +GNU General Public License for more details. + +You should have received a copy of the GNU General Public License +along with this program; if not, write to the Free Software +Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111, USA.