From: SourceForge.net <no...@so...> - 2007-12-29 21:53:24
|
Bugs item #1860677, was opened at 2007-12-29 16:53 Message generated for change (Tracker Item Submitted) made by Item Submitter You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=101355&aid=1860677&group_id=1355 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: None Group: None Status: Open Resolution: None Priority: 5 Private: No Submitted By: Dan Muller (dmuller) Assigned to: Nobody/Anonymous (nobody) Summary: DIRECTORY of dirs ignores :FULL Initial Comment: A call to DIRECTORY with a pathname that selects for subdirectories will not honor a non-nil value for the :FULL keyword argument. AFAICT, this leaves no way for a CLISP program to discover the name of a symbolic link file which references a valid directory. Example: cd /tmp mkdir test-dir cd test-dir mkdir a-dir ln -s /tmp/test-dir/a-dir a-dir-link (directory #P"/tmp/*/" :full t) => (#P"/tmp/test-dir/a-dir" #P"/tmp/test-dir/a-dir") I reproduced this effect on both Cygwin and Debian. --------------------------------------------------- On Windows: Output of uname -a: CYGWIN_NT-5.1 pegasus 1.5.25(0.156/4/2) 2007-12-14 19:21 i686 Cygwin Actually running on Windows XP Home Edition SP2. Binaries are latest from mirrors.kernel.org, as of 2007-12-29, installed via Cygwin Setup. I did not build this CLISP myself. Output of clisp --version: GNU CLISP 2.41 (2006-10-13) (built on ATGRZWN502840.avl01.avlcorp.lan [127.0.0.1]) Software: GNU C 3.4.4 (cygming special, gdc 0.12, using dmd 0.125) gcc -O2 -pipe -W -Wswitch -Wcomment -Wpointer-arith -Wimplicit -Wreturn-type -Wmissing-declarations -Wno-sign-compare - O2 -fexpensive-optimizations -falign-functions=4 -DUNICODE -DDYNAMIC_FFI -I. -Wl,--enable-auto-image-base -x none -lint l -liconv libcharset.a libavcall.a libcallback.a -lreadline -lncurses -lsigsegv SAFETY=0 HEAPCODES STANDARD_HEAPCODES SPVW_PAGES SPVW_MIXED libsigsegv 2.4 libiconv 1.11 libreadline 5.1 Features: (READLINE REGEXP SYSCALLS I18N LOOP COMPILER CLOS MOP CLISP ANSI-CL COMMON-LISP LISP=CL INTERPRETER SOCKETS GENERIC-STREAMS LOGICAL-PATHNAMES SCREEN FFI GETTEXT UNICODE BASE-CHAR=CHARACTER PC386 UNIX CYGWIN) C Modules: (clisp i18n syscalls regexp readline) Installation directory: /usr/lib/clisp/ User language: ENGLISH Machine: I686 (I686) pegasus [192.168.0.3] --------------------------------------------------- On Debian Linux: Output of uname -a: Linux scorpion 2.6.18-4-k7 #1 SMP Mon Mar 26 17:57:15 UTC 2007 i686 GNU/Linux Binaries are latest available for Debian testing branch, as of 2007-12-29, installed via aptitude. I did not build this CLISP myself. Output of clisp --version: GNU CLISP 2.41 (2006-10-13) (built 3369758291) (memory 3407938871) Software: GNU C 4.1.2 20061007 (prerelease) (Debian 4.1.1-16) /usr/bin/gcc-4.1 -g -O2 -W -Wswitch -Wcomment -Wpointer-arith -Wimplicit -Wreturn-type -Wmissing-declarations -Wno-sign-compare -O2 -fexpensive-optimizations -falign-functions=4 -DUNICODE -DDYNAMIC_FFI -DDYNAMIC_MODULES -I. -x none libcharset.a libavcall.a libcallback.a /usr/lib/libreadline.so -lncurses -ldl -L/usr/lib -lsigsegv SAFETY=0 HEAPCODES LINUX_NOEXEC_HEAPCODES GENERATIONAL_GC SPVW_BLOCKS SPVW_MIXED TRIVIALMAP_MEMORY libsigsegv 2.4 libreadline 5.1 Features: (ASDF CLC-OS-DEBIAN COMMON-LISP-CONTROLLER CLX-ANSI-COMMON-LISP CLX READLINE REGEXP SYSCALLS I18N LOOP COMPILER CLOS MOP CLISP ANSI-CL COMMON-LISP LISP=CL INTERPRETER SOCKETS GENERIC-STREAMS LOGICAL-PATHNAMES SCREEN FFI GETTEXT UNICODE BASE-CHAR=CHARACTER PC386 UNIX) C Modules: (clisp i18n syscalls regexp readline linux clx) Installation directory: /usr/lib/clisp/ User language: ENGLISH Machine: I686 (I686) scorpion.home [192.168.0.2] ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=101355&aid=1860677&group_id=1355 |
From: SourceForge.net <no...@so...> - 2007-12-29 23:11:29
|
Bugs item #1860677, was opened at 2007-12-29 16:53 Message generated for change (Comment added) made by dmuller You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=101355&aid=1860677&group_id=1355 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: None Group: None Status: Open Resolution: None >Priority: 6 Private: No Submitted By: Dan Muller (dmuller) Assigned to: Nobody/Anonymous (nobody) Summary: DIRECTORY of dirs ignores :FULL Initial Comment: A call to DIRECTORY with a pathname that selects for subdirectories will not honor a non-nil value for the :FULL keyword argument. AFAICT, this leaves no way for a CLISP program to discover the name of a symbolic link file which references a valid directory. Example: cd /tmp mkdir test-dir cd test-dir mkdir a-dir ln -s /tmp/test-dir/a-dir a-dir-link (directory #P"/tmp/*/" :full t) => (#P"/tmp/test-dir/a-dir" #P"/tmp/test-dir/a-dir") I reproduced this effect on both Cygwin and Debian. --------------------------------------------------- On Windows: Output of uname -a: CYGWIN_NT-5.1 pegasus 1.5.25(0.156/4/2) 2007-12-14 19:21 i686 Cygwin Actually running on Windows XP Home Edition SP2. Binaries are latest from mirrors.kernel.org, as of 2007-12-29, installed via Cygwin Setup. I did not build this CLISP myself. Output of clisp --version: GNU CLISP 2.41 (2006-10-13) (built on ATGRZWN502840.avl01.avlcorp.lan [127.0.0.1]) Software: GNU C 3.4.4 (cygming special, gdc 0.12, using dmd 0.125) gcc -O2 -pipe -W -Wswitch -Wcomment -Wpointer-arith -Wimplicit -Wreturn-type -Wmissing-declarations -Wno-sign-compare - O2 -fexpensive-optimizations -falign-functions=4 -DUNICODE -DDYNAMIC_FFI -I. -Wl,--enable-auto-image-base -x none -lint l -liconv libcharset.a libavcall.a libcallback.a -lreadline -lncurses -lsigsegv SAFETY=0 HEAPCODES STANDARD_HEAPCODES SPVW_PAGES SPVW_MIXED libsigsegv 2.4 libiconv 1.11 libreadline 5.1 Features: (READLINE REGEXP SYSCALLS I18N LOOP COMPILER CLOS MOP CLISP ANSI-CL COMMON-LISP LISP=CL INTERPRETER SOCKETS GENERIC-STREAMS LOGICAL-PATHNAMES SCREEN FFI GETTEXT UNICODE BASE-CHAR=CHARACTER PC386 UNIX CYGWIN) C Modules: (clisp i18n syscalls regexp readline) Installation directory: /usr/lib/clisp/ User language: ENGLISH Machine: I686 (I686) pegasus [192.168.0.3] --------------------------------------------------- On Debian Linux: Output of uname -a: Linux scorpion 2.6.18-4-k7 #1 SMP Mon Mar 26 17:57:15 UTC 2007 i686 GNU/Linux Binaries are latest available for Debian testing branch, as of 2007-12-29, installed via aptitude. I did not build this CLISP myself. Output of clisp --version: GNU CLISP 2.41 (2006-10-13) (built 3369758291) (memory 3407938871) Software: GNU C 4.1.2 20061007 (prerelease) (Debian 4.1.1-16) /usr/bin/gcc-4.1 -g -O2 -W -Wswitch -Wcomment -Wpointer-arith -Wimplicit -Wreturn-type -Wmissing-declarations -Wno-sign-compare -O2 -fexpensive-optimizations -falign-functions=4 -DUNICODE -DDYNAMIC_FFI -DDYNAMIC_MODULES -I. -x none libcharset.a libavcall.a libcallback.a /usr/lib/libreadline.so -lncurses -ldl -L/usr/lib -lsigsegv SAFETY=0 HEAPCODES LINUX_NOEXEC_HEAPCODES GENERATIONAL_GC SPVW_BLOCKS SPVW_MIXED TRIVIALMAP_MEMORY libsigsegv 2.4 libreadline 5.1 Features: (ASDF CLC-OS-DEBIAN COMMON-LISP-CONTROLLER CLX-ANSI-COMMON-LISP CLX READLINE REGEXP SYSCALLS I18N LOOP COMPILER CLOS MOP CLISP ANSI-CL COMMON-LISP LISP=CL INTERPRETER SOCKETS GENERIC-STREAMS LOGICAL-PATHNAMES SCREEN FFI GETTEXT UNICODE BASE-CHAR=CHARACTER PC386 UNIX) C Modules: (clisp i18n syscalls regexp readline linux clx) Installation directory: /usr/lib/clisp/ User language: ENGLISH Machine: I686 (I686) scorpion.home [192.168.0.2] ---------------------------------------------------------------------- >Comment By: Dan Muller (dmuller) Date: 2007-12-29 18:11 Message: Logged In: YES user_id=358502 Originator: YES BTW, bugs 1860489 and 1860677 are impediments to getting unit tests for ASDF-INSTALL working smoothly, as they prevent the unit tests from cleaning up test directory trees reliably. These trees contain internal symlinks which, due to these bugs, prevent the trees from being reliably deleted. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=101355&aid=1860677&group_id=1355 |
From: SourceForge.net <no...@so...> - 2007-12-31 21:03:17
|
Bugs item #1860677, was opened at 2007-12-29 16:53 Message generated for change (Comment added) made by sds You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=101355&aid=1860677&group_id=1355 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. >Category: clisp >Group: lisp error >Status: Closed >Resolution: Fixed Priority: 6 Private: No Submitted By: Dan Muller (dmuller) >Assigned to: Sam Steingold (sds) Summary: DIRECTORY of dirs ignores :FULL Initial Comment: A call to DIRECTORY with a pathname that selects for subdirectories will not honor a non-nil value for the :FULL keyword argument. AFAICT, this leaves no way for a CLISP program to discover the name of a symbolic link file which references a valid directory. Example: cd /tmp mkdir test-dir cd test-dir mkdir a-dir ln -s /tmp/test-dir/a-dir a-dir-link (directory #P"/tmp/*/" :full t) => (#P"/tmp/test-dir/a-dir" #P"/tmp/test-dir/a-dir") I reproduced this effect on both Cygwin and Debian. --------------------------------------------------- On Windows: Output of uname -a: CYGWIN_NT-5.1 pegasus 1.5.25(0.156/4/2) 2007-12-14 19:21 i686 Cygwin Actually running on Windows XP Home Edition SP2. Binaries are latest from mirrors.kernel.org, as of 2007-12-29, installed via Cygwin Setup. I did not build this CLISP myself. Output of clisp --version: GNU CLISP 2.41 (2006-10-13) (built on ATGRZWN502840.avl01.avlcorp.lan [127.0.0.1]) Software: GNU C 3.4.4 (cygming special, gdc 0.12, using dmd 0.125) gcc -O2 -pipe -W -Wswitch -Wcomment -Wpointer-arith -Wimplicit -Wreturn-type -Wmissing-declarations -Wno-sign-compare - O2 -fexpensive-optimizations -falign-functions=4 -DUNICODE -DDYNAMIC_FFI -I. -Wl,--enable-auto-image-base -x none -lint l -liconv libcharset.a libavcall.a libcallback.a -lreadline -lncurses -lsigsegv SAFETY=0 HEAPCODES STANDARD_HEAPCODES SPVW_PAGES SPVW_MIXED libsigsegv 2.4 libiconv 1.11 libreadline 5.1 Features: (READLINE REGEXP SYSCALLS I18N LOOP COMPILER CLOS MOP CLISP ANSI-CL COMMON-LISP LISP=CL INTERPRETER SOCKETS GENERIC-STREAMS LOGICAL-PATHNAMES SCREEN FFI GETTEXT UNICODE BASE-CHAR=CHARACTER PC386 UNIX CYGWIN) C Modules: (clisp i18n syscalls regexp readline) Installation directory: /usr/lib/clisp/ User language: ENGLISH Machine: I686 (I686) pegasus [192.168.0.3] --------------------------------------------------- On Debian Linux: Output of uname -a: Linux scorpion 2.6.18-4-k7 #1 SMP Mon Mar 26 17:57:15 UTC 2007 i686 GNU/Linux Binaries are latest available for Debian testing branch, as of 2007-12-29, installed via aptitude. I did not build this CLISP myself. Output of clisp --version: GNU CLISP 2.41 (2006-10-13) (built 3369758291) (memory 3407938871) Software: GNU C 4.1.2 20061007 (prerelease) (Debian 4.1.1-16) /usr/bin/gcc-4.1 -g -O2 -W -Wswitch -Wcomment -Wpointer-arith -Wimplicit -Wreturn-type -Wmissing-declarations -Wno-sign-compare -O2 -fexpensive-optimizations -falign-functions=4 -DUNICODE -DDYNAMIC_FFI -DDYNAMIC_MODULES -I. -x none libcharset.a libavcall.a libcallback.a /usr/lib/libreadline.so -lncurses -ldl -L/usr/lib -lsigsegv SAFETY=0 HEAPCODES LINUX_NOEXEC_HEAPCODES GENERATIONAL_GC SPVW_BLOCKS SPVW_MIXED TRIVIALMAP_MEMORY libsigsegv 2.4 libreadline 5.1 Features: (ASDF CLC-OS-DEBIAN COMMON-LISP-CONTROLLER CLX-ANSI-COMMON-LISP CLX READLINE REGEXP SYSCALLS I18N LOOP COMPILER CLOS MOP CLISP ANSI-CL COMMON-LISP LISP=CL INTERPRETER SOCKETS GENERIC-STREAMS LOGICAL-PATHNAMES SCREEN FFI GETTEXT UNICODE BASE-CHAR=CHARACTER PC386 UNIX) C Modules: (clisp i18n syscalls regexp readline linux clx) Installation directory: /usr/lib/clisp/ User language: ENGLISH Machine: I686 (I686) scorpion.home [192.168.0.2] ---------------------------------------------------------------------- Comment By: Sam Steingold (sds) Date: 2007-12-31 16:03 Message: Logged In: YES user_id=5735 Originator: NO thank you for your bug report. the bug has been fixed in the CVS tree. you can either wait for the next release (recommended) or check out the current CVS tree (see http://clisp.cons.org) and build CLISP from the sources (be advised that between releases the CVS tree is very unstable and may not even build on your platform). ---------------------------------------------------------------------- Comment By: Dan Muller (dmuller) Date: 2007-12-29 18:11 Message: Logged In: YES user_id=358502 Originator: YES BTW, bugs 1860489 and 1860677 are impediments to getting unit tests for ASDF-INSTALL working smoothly, as they prevent the unit tests from cleaning up test directory trees reliably. These trees contain internal symlinks which, due to these bugs, prevent the trees from being reliably deleted. ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=101355&aid=1860677&group_id=1355 |
From: Sam S. <sd...@gn...> - 2007-12-31 21:26:44
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 > https://sourceforge.net/tracker/?func=detail&atid=101355&aid=1860677&group_id=1355 >> Status: Closed >> Resolution: Fixed to fix this bug, I had to modify some woe32-specific code. I was careful (yeah, right), but I might have broken things for you. please test cvs head - if it builds, try $ ./clisp -q -norc -K boot [1]> (directory "*/" :full t) ((#P"/homedata/ssteingold/src/clisp/current/build-64-g/benchmarks/" #P"/homedata/ssteingold/src/clisp/current/build-64-g/benchmarks/" (9 43 10 20 12 2007) 4096) (#P"/homedata/ssteingold/src/clisp/current/build-64-g/ansi-tests/" #P"/homedata/ssteingold/src/clisp/current/build-64-g/ansi-tests/" (31 41 10 20 12 2007) 52224) (#P"/homedata/ssteingold/src/clisp/current/build-64-g/sacla-tests/" #P"/homedata/ssteingold/src/clisp/current/build-64-g/sacla-tests/" (26 17 10 20 12 2007) 2048) (#P"/homedata/ssteingold/src/clisp/current/build-64-g/tests/" #P"/homedata/ssteingold/src/clisp/current/build-64-g/tests/" (56 47 11 31 12 2007) 4096) (#P"/homedata/ssteingold/src/clisp/current/build-64-g/stage/" #P"/homedata/ssteingold/src/clisp/current/build-64-g/stage/" (3 8 16 12 12 2007) 10240) (#P"/homedata/ssteingold/src/clisp/current/build-64-g/boot/" #P"/homedata/ssteingold/src/clisp/current/build-64-g/boot/" (10 7 16 12 12 2007) 1024) (#P"/homedata/ssteingold/src/clisp/current/build-64-g/full/" #P"/homedata/ssteingold/src/clisp/current/build-64-g/full/" (32 14 16 31 12 2007) 1024) (#P"/homedata/ssteingold/src/clisp/current/build-64-g/data/" #P"/homedata/ssteingold/src/clisp/current/build-64-g/data/" (4 1 16 12 12 2007) 1024) (#P"/homedata/ssteingold/src/clisp/current/build-64-g/locale/" #P"/homedata/ssteingold/src/clisp/current/build-64-g/locale/" (3 1 16 12 12 2007) 1024) ..... and play with symbolic links to directories (if such things make sense under woe32). thanks! Sam. PS. happy new year! -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFHeV6HPp1Qsf2qnMcRAh5BAJ41ySo6UgY7Ohabad6sczvLZOadFwCfVwax 0gK2MAI5rUqxysZADX/p8So= =Uzmp -----END PGP SIGNATURE----- |
From: Michael K. <mic...@gm...> - 2008-01-03 00:17:29
|
Sam Steingold wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > >> https://sourceforge.net/tracker/?func=detail&atid=101355&aid=1860677&group_id=1355 >>> Status: Closed >>> Resolution: Fixed > > to fix this bug, I had to modify some woe32-specific code. > I was careful (yeah, right), but I might have broken things for you. > please test cvs head - if it builds, try It builds "almost": lisp.exe -B . -E 1:1 -Efile UTF-8 -Eterminal UTF-8 -norc -m 1800KW -x "(and (setq custom::*load-paths* (quote (\"../src/\"))) (load \"../src/init.lisp\") (sys::%saveinitmem) (ext::exit)) (ext::exit t)" i i i i i i i ooooo o ooooooo ooooo ooooo I I I I I I I 8 8 8 8 8 o 8 8 I \ `+' / I 8 8 8 8 8 8 \ `-+-' / 8 8 8 ooooo 8oooo `-__|__-' 8 8 8 8 8 | 8 o 8 8 o 8 8 ------+------ ooooo 8oooooo ooo8ooo ooooo 8 Welcome to GNU CLISP 2.43 (2007-11-18) <http://clisp.cons.org/> Copyright (c) Bruno Haible, Michael Stoll 1992, 1993 Copyright (c) Bruno Haible, Marcus Daniels 1994-1997 Copyright (c) Bruno Haible, Pierpaolo Bernardi, Sam Steingold 1998 Copyright (c) Bruno Haible, Sam Steingold 1999-2000 Copyright (c) Sam Steingold, Bruno Haible 2001-2007 Type :h and hit Enter for context help. *** - READ: input stream #<INPUT STRING-INPUT-STREAM> ends within an object Bye. mv lispimag.mem interpreted.mem mv: cannot stat `lispimag.mem': No such file or directory make: *** [interpreted.mem] Error 1 > $ ./clisp -q -norc -K boot $ ./lisp -q -norc -K boot WARNING: No initialization file specified. Please try: d:\Projects\clisp-cvs\clisp\build.mingw\lisp.exe -M lispinit.mem WARNING: No installation directory specified. Please try: d:\Projects\clisp-cvs\clisp\build.mingw\lisp.exe -B /usr/local/lib/clisp > (directory "*/" :full t) *** - Win32 error 2 (ERROR_FILE_NOT_FOUND): The system cannot find the file specified. 1. Break> (directory ".") NIL 1. Break> (directory "./") *** - READ from #<INPUT CONCATENATED-STREAM #<INPUT STRING-INPUT-STREAM> #<IO TERMINAL-STREAM>>: illegal character #\Escape 2. Break> (directory "./") (#P"D:\\Projects\\clisp-cvs\\clisp\\build.mingw\\") 2. Break> *** - READ from #<INPUT CONCATENATED-STREAM #<INPUT STRING-INPUT-STREAM> #<IO TERMINAL-STREAM>>: illegal character #\Escape 3. Break> > PS. happy new year! New Year's Pledge: Let woe32 cause less trouble... Regards Michael |
From: Sam S. <sd...@gn...> - 2008-01-03 15:31:38
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Michael Kappert wrote: > Sam Steingold wrote: >>> https://sourceforge.net/tracker/?func=detail&atid=101355&aid=1860677&group_id=1355 >>>> Status: Closed >>>> Resolution: Fixed >> to fix this bug, I had to modify some woe32-specific code. >> I was careful (yeah, right), but I might have broken things for you. >> please test cvs head - if it builds, try > > It builds "almost": > > lisp.exe -B . -E 1:1 -Efile UTF-8 -Eterminal UTF-8 -norc -m 1800KW -x "(and (setq custom::*load-paths* (quote (\"../src/\"))) (load \"../src/init.lisp\") (sys::%saveinitmem) (ext::exit)) (ext::exit t)" > *** - READ: input stream #<INPUT STRING-INPUT-STREAM> ends within an object this is frighteningly similar to the bug reported by Yaroslav. could you please debug this? you need to run ./configure --with-debug and play with gdb. IIRC, "::" is somehow transformed into ";". >> $ ./clisp -q -norc -K boot > $ ./lisp -q -norc -K boot > > (directory "*/" :full t) > > *** - Win32 error 2 (ERROR_FILE_NOT_FOUND): The system cannot find the file specified. please build --with-debug and report the stack trace. > 1. Break> (directory ".") > > NIL should be *** - DIRECTORY: ".../." names a directory, not a file > 1. Break> (directory "./") > > *** - READ from #<INPUT CONCATENATED-STREAM #<INPUT STRING-INPUT-STREAM> #<IO TERMINAL-STREAM>>: illegal character #\Escape > 2. Break> (directory "./") > > (#P"D:\\Projects\\clisp-cvs\\clisp\\build.mingw\\") > 2. Break> > *** - READ from #<INPUT CONCATENATED-STREAM #<INPUT STRING-INPUT-STREAM> #<IO TERMINAL-STREAM>>: illegal character #\Escape > 3. Break> looks like spurious control characters in your input stream. this is probably related to the above "::"-->";" issue. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFHfP/QPp1Qsf2qnMcRAgYTAJ4p2+9O/a51bm8vwLPsCUrVybbABwCgl+gZ GySllmOMzaYZh0e7MPAaRzs= =vAIY -----END PGP SIGNATURE----- |
From: Michael K. <mic...@gm...> - 2008-01-05 16:50:59
|
Sam Steingold wrote: > this is frighteningly similar to the bug reported by Yaroslav. > could you please debug this? > you need to run ./configure --with-debug and play with gdb. > IIRC, "::" is somehow transformed into ";". Ok, I played with gdb... gdb seems to jump back and forth in the source code without any visible reason?! I got to a breakpoint in eval, which I think is good, but I don't know how to inspect the one_o field (see below). I think I need to spend some more time reading gdb documentation... ------- Current directory is d:/Projects/clisp-cvs/clisp/build.mingw-debug/ GNU gdb 6.5.50.20060706-cvs (cygwin-special) [..] (gdb) run Starting program: /cygdrive/d/Projects/clisp-cvs/clisp/build.mingw-debug/lisp.exe -x "(and (setq custom::*load-paths* (quote (\"../src/\"))) (load \"../src/init.lisp\") (sys::%saveinitmem) (ext::exit)) (ext::exit t)" [..] ;; Loading file ..\src\defseq.lisp ... ;; Loaded file ..\src\defseq.lisp [..] ;; Loading file D:\Projects\clisp-cvs\clisp\src\room.lisp ... ;; Loaded file D:\Projects\clisp-cvs\clisp\src\room.lisp [..] interpret_bytecode_ (closure={one_o = 434341629}, codeptr=0x19e38640, byteptr_in=0x19e38652 ";\001\003\016") at ../src/eval.d:7694 eval1 (form={one_o = 2221808}) at ../src/eval.d:3848 eval (form={one_o = 1526398627}) at ../src/eval.d:2919 (gdb) ------- Regards Michael |
From: Sam S. <sd...@gn...> - 2008-01-06 02:15:27
|
> * Michael Kappert <zvp...@tz...g> [2008-01-05 17:50:54 +0100]: > > Sam Steingold wrote: >> this is frighteningly similar to the bug reported by Yaroslav. >> could you please debug this? >> you need to run ./configure --with-debug and play with gdb. >> IIRC, "::" is somehow transformed into ";". > > Ok, I played with gdb... > gdb seems to jump back and forth in the source code without any > visible reason?! that's because of "-O". edit CFLAGS in Makefile and do $ rm -fv *.o && make lisp.run in the future, do "./configure CFLAGS="" --with-debug" > I got to a breakpoint in eval, which I think is good, > but I don't know how to inspect the one_o field (see below). "xout form" will print using nobject_out(), "zout form" with print it using lisp PRIN1. -- Sam Steingold (http://sds.podval.org/) on Fedora release 8 (Werewolf) http://ffii.org http://iris.org.il http://mideasttruth.com http://dhimmi.com http://honestreporting.com http://israelunderattack.slide.com War has never solved anything - except for ending slavery, fascism, communism. |
From: Michael K. <mic...@gm...> - 2008-01-06 19:09:18
|
Sam Steingold wrote: >> * Michael Kappert [2008-01-05 17:50:54 +0100]: >> >> Sam Steingold wrote: >>> this is frighteningly similar to the bug reported by Yaroslav. >>> could you please debug this? > edit CFLAGS in Makefile and do > $ rm -fv *.o && make lisp.run > > in the future, do "./configure CFLAGS="" --with-debug" Hmm, this is getting strange. The error ("input stream #<INPUT STRING-INPUT-STREAM> ends with an object") does not occur when the command is run in the debugger. The double colons are still there after the -x arg was read. I was not able to trace processing of the expression in more detail yet. ./configure CFLAGS="" ... did not change anything. Could this be the same bug as the one reported by Vadim? However, compiling with -O0 also breaks (gcc version "3.4.5(mingw special)") Regards Michael |
From: Sam S. <sd...@gn...> - 2008-01-06 19:57:29
|
> * Michael Kappert <zvp...@tz...g> [2008-01-06 20:09:12 +0100]: > > Sam Steingold wrote: >>> * Michael Kappert [2008-01-05 17:50:54 +0100]: >>> >>> Sam Steingold wrote: >>>> this is frighteningly similar to the bug reported by Yaroslav. >>>> could you please debug this? >> edit CFLAGS in Makefile and do >> $ rm -fv *.o && make lisp.run >> >> in the future, do "./configure CFLAGS="" --with-debug" > > Hmm, this is getting strange. > The error ("input stream #<INPUT STRING-INPUT-STREAM> ends with an object") > does not occur when the command is run in the debugger. > The double colons are still there after the -x arg was read. > I was not able to trace processing of the expression in more detail yet. not having access to a woe32 system, all I can suggest is either -- spray io.d, spvw.d and stream.d with printfs and nobject_outs -- or do a binary search on time to figure out which specific cvs patch broke clisp. > Could this be the same bug as the one reported by Vadim? unlikely. gcc3 has been working fine. -- Sam Steingold (http://sds.podval.org/) on Fedora release 8 (Werewolf) http://jihadwatch.org http://iris.org.il http://honestreporting.com http://truepeace.org http://pmw.org.il http://mideasttruth.com Parachute for sale, used once, never opened, small stain. |
From: Michael K. <mic...@gm...> - 2008-01-08 00:16:39
|
Sam Steingold wrote: >> * Michael Kappert <zvp...@tz...g> [2008-01-06 20:09:12 +0100]: >> >> Sam Steingold wrote: >>>> * Michael Kappert [2008-01-05 17:50:54 +0100]: >>>> >>>> Sam Steingold wrote: >>>>> this is frighteningly similar to the bug reported by Yaroslav. >>>>> could you please debug this? It's a shell problem... $ cat test.c #include <stdio.h> int main (int argc, char* argv[]) { fprintf(stdout, "test: %s\n", argv[1]); exit(0); } $ ./test "(../src)(foo::bar)" test: (..\src)(foo;bar) $ ... and it looks like a bug: $ ./test "(foo::bar)" test: (foo::bar) $ Happens both with bash 2.0.4 and 3.1.0. I'll ask the bash/mingw folks about this. I did not find a workaround yet. Regards Michael |
From: Michael K. <mic...@gm...> - 2008-01-11 13:55:54
|
Michael Kappert wrote: > It's a shell problem... Ok, some more info on this. It's not a shell problem but a problem with MSYS <http://www.mingw.org/msys.shtml> itself. MSYS does automagic unix to windows pathname conversion. If an argument contains both a forward slash and double colons, it is converted. Using double slashes is supposed to suppress conversion, but this doesn't seem to work (in msys 1.0.10 and 1.0.11). The arguments to lisp.exe for target interpretd.mem and halfcompild.mem did not contain slashes in rel. 2.43, this is why the build worked there. Would it be possible to revert to that state? regards Michael |
From: Sam S. <sd...@gn...> - 2008-01-11 14:08:35
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Michael Kappert wrote: | Michael Kappert wrote: | |> It's a shell problem... | | Ok, some more info on this. | It's not a shell problem but a problem with MSYS <http://www.mingw.org/msys.shtml> | itself. MSYS does automagic unix to windows pathname conversion. If an argument | contains both a forward slash and double colons, it is converted. Using double slashes | is supposed to suppress conversion, but this doesn't seem to work (in msys 1.0.10 and 1.0.11). this sucks. | The arguments to lisp.exe for target interpretd.mem and halfcompild.mem did not contain | slashes in rel. 2.43, this is why the build worked there. | | Would it be possible to revert to that state? not really. this was introduced to avoid linking files from src/ to build/ - making the build directory more manageable. even if we make this behavior optional on mingw, I cannot commit to this restriction in the future. msys people should make this behavior optional (or, better yet, eliminate it altogether). what do they say about this "feature"? -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFHh3hYPp1Qsf2qnMcRAvFFAKCaJ7y3v23e5F6b27/eK/q1AsMdugCglv6z 6sitn8oj3bVfkqgPhy3fyTY= =HHtz -----END PGP SIGNATURE----- |
From: Michael K. <mic...@gm...> - 2008-01-11 21:47:36
|
Sam Steingold wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > Michael Kappert wrote: > > Using double slashes > > is supposed to suppress conversion, but this doesn't seem to work (in > > msys 1.0.10 and 1.0.11). > > this sucks. > > what do they say about this "feature"? See <news://news.gmane.org:119/120...@ub...m> In summary - works for 99% of the cases; we are in the unfortunate 1% - problem was known and under discussion, but rather low prio - suggestions include - introducing an environment variable to switch off conversion - introduce argument prefixes, interpreted by msys, to control conversion - we are probably better off with cygwin for the time being I had no luck with cygwin on a quick try; I'd like to hear form other win32 users. Michael |
From: Sam S. <sd...@gn...> - 2008-01-14 16:06:17
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Michael Kappert wrote: | See | <news://news.gmane.org:119/120...@ub...m> | In summary | - works for 99% of the cases; we are in the unfortunate 1% | - problem was known and under discussion, but rather low prio | - suggestions include | - introducing an environment variable to switch off conversion | - introduce argument prefixes, interpreted by msys, to control conversion | - we are probably better off with cygwin for the time being | | I had no luck with cygwin on a quick try; I'd like to hear form other win32 users. Please tell them about our perspective: the portability one. we write code which has to work on many different platforms, and each "feature" like that adds to our workload. ideally, there should be a way to make msys shell behave just like a vanilla bash on a vanilla unix, an environment variable being the most obvious choice. until they do that, I am afraid we will have to stick with recommending that people get cygwin and build clisp with $ ./configure --with-mingw .... Sorry. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFHi4hwPp1Qsf2qnMcRAlYUAJ46vK+W3VCbRGTmqdKCxiRB47cP0ACgj3yK KF8O1CmxGnhbg60JnK4svbE= =8aS8 -----END PGP SIGNATURE----- |
From: Sam S. <sd...@gn...> - 2008-01-11 15:02:19
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Michael Kappert wrote: | Michael Kappert wrote: | |> It's a shell problem... | | Ok, some more info on this. | It's not a shell problem but a problem with MSYS <http://www.mingw.org/msys.shtml> | itself. MSYS does automagic unix to windows pathname conversion. If an argument | contains both a forward slash and double colons, it is converted. OK, so building in src (without the build-dir configure argument) should work, right? please build clisp in src and see if (directory "*/" :full t) works thanks (as requested in <http://permalink.gmane.org/gmane.lisp.clisp.devel/17337>) -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFHh4TvPp1Qsf2qnMcRAk3yAKCqaI0EKdz29omzegzwkg7Eid/fHwCePiH3 fornjRF1p++TY/CeiQpb2v8= =N0V1 -----END PGP SIGNATURE----- |
From: Michael K. <mic...@gm...> - 2008-01-11 17:42:50
|
Sam Steingold wrote: > OK, so building in src (without the build-dir configure argument) should > work, right? > please build clisp in src and see if (directory "*/" :full t) works > thanks (as requested in <http://permalink.gmane.org/gmane.lisp.clisp.devel/17337>) Doesn't work: [2]> (directory "*/" :full t) *** - Win32 error 2 (ERROR_FILE_NOT_FOUND): The system cannot find the file specified. The following restarts are available: ABORT :R1 ABORT (directory "*/") works. I can try to debug this but I need to restore my mingw environment first. Currently libsigsegv and libiconv are missing (as well as gdb). Btw, I noticed that target lispimag.mem does not have the (setq custom::*load-paths* ...) form. Is this ok? Michael |
From: Sam S. <sd...@gn...> - 2008-01-11 18:18:58
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Michael Kappert wrote: | Sam Steingold wrote: |> OK, so building in src (without the build-dir configure argument) should |> work, right? |> please build clisp in src and see if (directory "*/" :full t) works |> thanks (as requested in <http://permalink.gmane.org/gmane.lisp.clisp.devel/17337>) | | Doesn't work: | | [2]> (directory "*/" :full t) | | *** - Win32 error 2 (ERROR_FILE_NOT_FOUND): The system cannot find the file | specified. | The following restarts are available: | ABORT :R1 ABORT | | (directory "*/") works. please compile --with-debug, then the above message should carry file:lineno information. | I can try to debug this but I need to restore my mingw environment | first. Currently libsigsegv and libiconv are missing (as well as gdb). please do. | Btw, I noticed that target lispimag.mem does not have the | (setq custom::*load-paths* ...) form. Is this ok? there is no lispimag.mem target in my makefiles. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFHh7MIPp1Qsf2qnMcRAtItAJ4rfgeFXgTyBaaXHSsujrSpZwtESgCcDf4X KY1seXGSL0BqC1IoInC29Zg= =4DNU -----END PGP SIGNATURE----- |
From: Michael K. <mic...@gm...> - 2008-01-11 21:34:22
|
Sam Steingold wrote: > please compile --with-debug, then the above message should carry > file:lineno information. I did ./configure --with-mingw --with-debug --ignore-absence-of-libsigsegv but clisp printed no file or lineno information -- same message as before. I got an additional OS error message (Please send problem report to M$) which I think wasn't there before. Note that I currently don't have sigsegv. I did a fresh cvs co because ./configure --with-debug didn't work with the already-configured clisp. Before that I tried (cd src; make clean) but that seems to delete config.lisp when it shouldn't ? > | Btw, I noticed that target lispimag.mem does not have the > | (setq custom::*load-paths* ...) form. Is this ok? > > there is no lispimag.mem target in my makefiles. Oops. The target name is lispinit.mem. Michael |
From: Sam S. <sd...@gn...> - 2008-01-14 16:08:58
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Michael Kappert wrote: | Sam Steingold wrote: |> please compile --with-debug, then the above message should carry |> file:lineno information. | I did ./configure --with-mingw --with-debug --ignore-absence-of-libsigsegv | but clisp printed no file or lineno information -- same message as before. this is controlled by DEBUG_OS_ERROR. just add -DDEBUG_OS_ERROR to CFLAGS in your Makefile. | I got an additional OS error message (Please send problem report to M$) | which I think wasn't there before. Note that I currently don't have sigsegv. | I did a fresh cvs co because ./configure --with-debug didn't work with the | already-configured clisp. Before that I tried (cd src; make clean) but that seems | to delete config.lisp when it shouldn't ? yes, it should. use "make config.lisp" to re-create it. |> | Btw, I noticed that target lispimag.mem does not have the |> | (setq custom::*load-paths* ...) form. Is this ok? |> there is no lispimag.mem target in my makefiles. | Oops. The target name is lispinit.mem. this is correct. lispinit.mem should load only compiled files in the build directory, not sources from the src directory. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFHi4kRPp1Qsf2qnMcRAkczAJ4zInBj7rPLxnSYQJ/kaSRc6c/YIQCfVL2s q2JQ4VgjNhwMm2EdiZAw1RE= =RqaP -----END PGP SIGNATURE----- |
From: Sam S. <sd...@gn...> - 2008-01-23 18:26:36
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Michael Kappert wrote: | Sam Steingold wrote: |> OK, so building in src (without the build-dir configure argument) should |> work, right? |> please build clisp in src and see if (directory "*/" :full t) works |> thanks (as requested in <http://permalink.gmane.org/gmane.lisp.clisp.devel/17337>) | | Doesn't work: | | [2]> (directory "*/" :full t) | | *** - Win32 error 2 (ERROR_FILE_NOT_FOUND): The system cannot find the file | specified. | The following restarts are available: | ABORT :R1 ABORT please edit Makefile and add -DDEBUG_OS_ERROR to CFLAGS, then do rm -f pathname.o make lisp.run ./clisp -q -norc -K boot -x 'directory "*/" :full t)' thanks. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFHl4bRPp1Qsf2qnMcRAp/rAJ9+qy9eKVca2huG14wsRwwYlSM5tQCfVRGn dgw6HB3UgEhoMzxXKyZu/Us= =Gupz -----END PGP SIGNATURE----- |
From: Michael K. <mic...@gm...> - 2008-01-24 21:25:28
|
Sam Steingold wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > Michael Kappert wrote: > | Sam Steingold wrote: > |> OK, so building in src (without the build-dir configure argument) should > |> work, right? > |> please build clisp in src and see if (directory "*/" :full t) works > |> thanks (as requested in > <http://permalink.gmane.org/gmane.lisp.clisp.devel/17337>) > | > | Doesn't work: > | > | [2]> (directory "*/" :full t) > | > | *** - Win32 error 2 (ERROR_FILE_NOT_FOUND): The system cannot find the > file > | specified. > | The following restarts are available: > | ABORT :R1 ABORT > > please edit Makefile and add -DDEBUG_OS_ERROR to CFLAGS, then do > > rm -f pathname.o > make lisp.run > ./clisp -q -norc -K boot -x 'directory "*/" :full t)' > > thanks. I tried the above commands (with the obvious corrections). I also did cvs up; ./configure --with-debug from scratch and tried removing the optimize flags from CFLAGS - clisp still refuses to print line number in the error message. Sorry. Suggestions are welcome. Debugging wasn't really successful either - the error seems to happen in a call issued from the with_sstring_0 macro, in pathname.d:6800, immediately after the macro arguments are processed ?! I'd expect the error to be raised during READDIR_findfirst (on line pathname.d:6803) but pathname:6803 is never reached. Are my line numbers messed up? Below is a transscript from a debug session. Maybe someone can suggest. Sorry for the long post. Regards Michael [3]> (directory "*/" :full t) Breakpoint 15, get_time_size (timepoint=0x22b4c0, entry_size=0x76632d70) at ../src/pathname.d:6800 (gdb) next (gdb) next (gdb) where #0 get_time_size (timepoint=0x22b4c0, entry_size=0x22b4b8) at ../src/pathname.d:6800 #1 0x0043405e in directory_search (pathname={one_o = 2274496}, dsp=0x22b518) at ../src/pathname.d:6815 #2 0x00434358 in C_directory () at ../src/pathname.d:7612 #3 0x00419bf3 in eval_subr (fun={one_o = 5034194}) at ../src/eval.d:3542 #4 0x004184f4 in eval1 (form={one_o = 1527135739}) at ../src/eval.d:3034 #5 0x00419697 in eval (form={one_o = 1527135739}) at ../src/eval.d:2908 #6 0x0048e007 in C_read_eval_print () at ../src/debug.d:407 #7 0x004122f7 in funcall_subr (fun={one_o = 5023026}, args_on_stack=2) at ../src/eval.d:5184 #8 0x004130f1 in interpret_bytecode_ (closure={one_o = 434603841}, codeptr=0x19e785e8, byteptr_in=0x43 <Address 0x43 out of bounds>) at ../src/eval.d:6743 #9 0x0041aef8 in funcall_closure (closure={one_o = 434603841}, args_on_stack=0) at ../src/eval.d:5587 #10 0x004225e6 in C_driver () at ../src/control.d:1971 #11 0x00413141 in interpret_bytecode_ (closure={one_o = 434603925}, codeptr=0x19e7859c, byteptr_in=0x18 <Address 0x18 out of bounds>) at ../src/eval.d:6749 #12 0x0041aef8 in funcall_closure (closure={one_o = 434603925}, args_on_stack=0) at ../src/eval.d:5587 #13 0x004133a1 in interpret_bytecode_ (closure={one_o = 434754545}, codeptr=0x19dd558c, byteptr_in=0x3e <Address 0x3e out of bounds>) at ../src/eval.d:6798 #14 0x0041aef8 in funcall_closure (closure={one_o = 434754545}, args_on_stack=0) at ../src/eval.d:5587 #15 0x0048e222 in driver () at ../src/debug.d:476 #16 0x0040b432 in main_actions (p=0x593570) at ../src/spvw.d:3279 #17 0x0040f1fe in main (argc=11, argv=0x3d2970) at ../src/spvw.d:3407 (gdb) step whole_namestring (pathname={one_o = 435033553}) at ../src/pathname.d:2536 (gdb) step directory_namestring_parts (pathname={one_o = 435033796}) at ../src/pathname.d:2395 (gdb) step (gdb) step (gdb) step (gdb) step (gdb) finish Run till exit from #0 directory_namestring_parts (pathname= {one_o = 435033553}) at ../src/pathname.d:2400 0x0042a5d7 in whole_namestring (pathname={one_o = 435033553}) at ../src/pathname.d:2536 Value returned is $1 = 13 (gdb) finish Run till exit from #0 0x0042a5d7 in whole_namestring (pathname= {one_o = 435033553}) at ../src/pathname.d:2536 0x00432bd9 in get_time_size (timepoint=0x22b4c0, entry_size=0x22b4b8) at ../src/pathname.d:6800 Value returned is $2 = {one_o = 435033797} (gdb) xout $2 "D:\Projects\clisp-cvs\clisp\build.mingw.debug\base\"{one_o = 435033797} (gdb) step nls_asciiext_wcslen (encoding={one_o = 434759421}, src=0x19ee16cc, srcend=0x19e9e6fd) at ../src/encoding.d:1787 (gdb) finish Run till exit from #0 nls_asciiext_wcslen (encoding={one_o = 434759421}, src=0x19ee16cc, srcend=0x19e9e6fd) at ../src/encoding.d:1787 0x00432c7f in get_time_size (timepoint=0x22b4c0, entry_size=0x22b4b8) at ../src/pathname.d:6800 Value returned is $3 = 51 (gdb) step nls_asciiext_wcstombs (encoding={one_o = 434759421}, stream={one_o = 0}, srcp=0x22b338, srcend=0x19e9e6fd, destp=0x22b33c, destend=0x19e9e6fd "æé\0310") at ../src/encoding.d:1822 (gdb) finish Run till exit from #0 nls_asciiext_wcstombs (encoding={one_o = 434759421}, stream={one_o = 0}, srcp=0x22b338, srcend=0x19e9e6fd, destp=0x22b33c, destend=0x19e9e6fd "æé\0310") at ../src/encoding.d:1822 0x00432cd7 in get_time_size (timepoint=0x22b4c0, entry_size=0x22b4b8) at ../src/pathname.d:6800 (gdb) step Breakpoint 12, OS_file_error (pathname={one_o = 2009275280}) at ../src/errwin32.d:3898 (gdb) where #0 OS_file_error (pathname={one_o = 2009275280}) at ../src/errwin32.d:3898 #1 0x00432d73 in get_time_size (timepoint=0x22b4c0, entry_size=0x22b4b8) at ../src/pathname.d:6800 #2 0x0043405e in directory_search (pathname={one_o = 2274496}, dsp=0x22b518) at ../src/pathname.d:6815 #3 0x00434358 in C_directory () at ../src/pathname.d:7612 #4 0x00419bf3 in eval_subr (fun={one_o = 5034194}) at ../src/eval.d:3542 #5 0x004184f4 in eval1 (form={one_o = 1527135739}) at ../src/eval.d:3034 #6 0x00419697 in eval (form={one_o = 1527135739}) at ../src/eval.d:2908 #7 0x0048e007 in C_read_eval_print () at ../src/debug.d:407 #8 0x004122f7 in funcall_subr (fun={one_o = 5023026}, args_on_stack=2) at ../src/eval.d:5184 #9 0x004130f1 in interpret_bytecode_ (closure={one_o = 434603841}, codeptr=0x19e785e8, byteptr_in=0x43 <Address 0x43 out of bounds>) at ../src/eval.d:6743 #10 0x0041aef8 in funcall_closure (closure={one_o = 434603841}, args_on_stack=0) at ../src/eval.d:5587 #11 0x004225e6 in C_driver () at ../src/control.d:1971 #12 0x00413141 in interpret_bytecode_ (closure={one_o = 434603925}, codeptr=0x19e7859c, byteptr_in=0x18 <Address 0x18 out of bounds>) at ../src/eval.d:6749 #13 0x0041aef8 in funcall_closure (closure={one_o = 434603925}, args_on_stack=0) at ../src/eval.d:5587 #14 0x004133a1 in interpret_bytecode_ (closure={one_o = 434754545}, codeptr=0x19dd558c, byteptr_in=0x3e <Address 0x3e out of bounds>) at ../src/eval.d:6798 #15 0x0041aef8 in funcall_closure (closure={one_o = 434754545}, args_on_stack=0) at ../src/eval.d:5587 #16 0x0048e222 in driver () at ../src/debug.d:476 #17 0x0040b432 in main_actions (p=0x593570) at ../src/spvw.d:3279 #18 0x0040f1fe in main (argc=11, argv=0x3d2970) at ../src/spvw.d:3407 (gdb) |
From: Sam S. <sd...@gn...> - 2008-01-24 21:53:12
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Michael Kappert wrote: | clisp still refuses to print line number in the error message. Sorry. | Suggestions are welcome. check CFLAGS: does it contain -DDEBUG_OS_ERROR? try adding "#define OS_error foo" to pathname.d - do you get a cpp error? | I'd expect the error to be raised during READDIR_findfirst (on line pathname.d:6803) | but pathname:6803 is never reached. Are my line numbers messed up? alas, all macros, such as with_sstring_0, are treated by C and thus gdb as being all on one line. | [3]> (directory "*/" :full t) | | Breakpoint 15, get_time_size (timepoint=0x22b4c0, entry_size=0x76632d70) | at ../src/pathname.d:6800 | (gdb) next | (gdb) next | (gdb) where | #0 get_time_size (timepoint=0x22b4c0, entry_size=0x22b4b8) | at ../src/pathname.d:6800 print STACK_0 with "xout STACK[-1]" when inside OS_file_error(), do "up" and "p resolved_asciz" and "p notfound" please try the appended patch: - --- pathname.d.~1.428.~ 2008-01-22 17:33:45.000000000 -0500 +++ pathname.d 2008-01-24 16:52:18.005334000 -0500 @@ -6802,7 +6802,7 @@ local maygc void get_time_size (decoded_ ~ begin_system_call(); ~ READDIR_findfirst(resolved_asciz, notfound = true; , notfound = true; ); ~ end_system_call(); - - if (notfound || READDIR_entry_ISDIR()) /* just to be paranoid */ + if (notfound) /* just to be paranoid */ ~ OS_file_error(STACK_1); ~ READDIR_entry_timedate(timepoint); ~ *entry_size = READDIR_entry_size(); -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFHmQi6Pp1Qsf2qnMcRAj4SAJ4oFeL55BOugV9iWB+OWuly85UdUACfU/bL 4wUsA7LuaP5kz8Q+7D9A0Gc= =xSsg -----END PGP SIGNATURE----- |
From: Sam S. <sd...@gn...> - 2008-01-25 16:13:37
|
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Sam Steingold wrote: | please try the appended patch: did you try the patch? if (directory "*/" :full t) now works, does "make check" pass? thanks. -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFHmgqkPp1Qsf2qnMcRAm5zAJ9LDSPM96azSDAMGldVG54gxNnjKwCcC+uW M0j9ciCiKDP7eCV/K9Xx5S8= =M4OH -----END PGP SIGNATURE----- |
From: Michael K. <mic...@gm...> - 2008-01-25 20:59:35
|
Sam Steingold wrote: > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > Sam Steingold wrote: > | please try the appended patch: > > did you try the patch? Sorry I wasn't able to look at this sooner. The patch failed (error in line 4). I changed line pathame.d:6805 to read if (notfound) /* just to be paranoid */ manually but still get the error: -------------------------------------------------- [1]> (directory "*/" :full t) Breakpoint 12, OS_file_error (pathname={one_o = 434766261}) at ../src/errwin32.d:3898 (gdb) where #0 OS_file_error (pathname={one_o = 434766261}) at ../src/errwin32.d:3898 #1 0x0044f2d3 in get_time_size (timepoint=0x22b3a0, entry_size=0x22b398) at ../src/pathname.d:6800 #2 0x0044f374 in with_stat_info_computed () at ../src/pathname.d:6815 #3 0x0045089a in directory_search (pathname={one_o = 434762977}, dsp=0x22b468) at ../src/pathname.d:7456 [20 more] (gdb) xout STACK[-1] #(CL::PATHNAME :HOST CL::NIL :DEVICE "D" :DIRECTORY (:ABSOLUTE "Projects" "clisp-cvs" "clisp" "build.mingw.debug" "base") :NAME CL::NIL :TYPE CL::NIL :VERSION :NEWEST){one_o = 434836325} (gdb) up #1 0x0044f2d3 in get_time_size (timepoint=0x22b3a0, entry_size=0x22b398) at ../src/pathname.d:6800 warning: Source file is more recent than executable. (gdb) p resolved_asciz $1 = 0x22b180 "D:\\Projects\\clisp-cvs\\clisp\\build.mingw.debug\\base\\" (gdb) p notfound $2 = true (gdb) quit ------------------------------------------------------ I disabled the OS_file_error statement completely and now I get a proper directory listing: [1]> (directory "*/" :full t) ((#P"D:\\Projects\\clisp-cvs\\clisp\\build.mingw.debug\\zlib\\" #P"D:\\Projects\\clisp-cvs\\clisp\\build.mingw.debug\\zlib\\" (31 37 15 5 9 7519) 8286724913799310704) [..] (#P"D:\\Projects\\clisp-cvs\\clisp\\build.mingw.debug\\base\\" #P"D:\\Projects\\clisp-cvs\\clisp\\build.mingw.debug\\base\\" (32 24 18 18 3 7519) 8286724913799310704)) [2]> Regards Michael |