[BACK]Return to BUGS CVS log [TXT] [DIR] Up to [gforth] / gforth

Diff for /gforth/BUGS between versions 1.19 and 1.56

version 1.19, 1995/09/06 21:00:07 version 1.56, 2003/11/06 10:23:31
Line 1 Line 1
 etags.fs crashes one of my applications (gs.fs). anton 12jan95  You can find more recent bug reports at
   http://savannah.gnu.org/bugs/?group=gforth
 not all aliases are in the etags file. Bug in etags.fs? anton 24jan95  
 -> wrote special tag generation for cross compilation. bernd 6sep95  
   
 emacs often finds the wrong tag. anton 24jan95  
 -> emacs does not search for a complete word, but for parts.  
 -> Solution: add blanks in front and end of tag name (etags.fs)  
 -> and use own forth-find-tag. bernd 6sep95  
   
 gforth.el: indentation does not work right on the first line of a  gforth.el: indentation does not work right on the first line of a
 buffer. anton 27jan95  buffer. anton 27jan95
 -> Solved?  
   
 Conditional compilation continues after the file ends. This is allowed  Conditional compilation continues after the file ends. This is allowed
 by the standard (through an ambiguous condition), but the compiler  by the standard (through an ambiguous condition), but the compiler
 should at least produce a warning.  anton 27jan95  should at least produce a warning.  anton 27jan95
   
 .1 is interpreted as floating-point number, not as double number or  
 error. anton 4may95  
 -> All numbers ecvt converts are (by implementation) floating-point.  
 -> Should we add a check againgst . at the start of a number?  
 -> .1 isn't a double number; it's either an error or a FP number.  
 -> bernd 5sep95  
   
 DOS and OS/2 don't like . at begin of a file (.gforth-history). bernd 6sep95py  
   
   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 <ret> 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().
   
   -----
   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.
Removed from v.1.19  
changed lines
  Added in v.1.56

FreeBSD-CVSweb <freebsd-cvsweb@FreeBSD.org>