On 2/27/10 Feb 27 -8:16 AM, David McClain wrote:
Hello Robert,
I took a look at your test shell script (run-tests.sh) and made the adjustment for running my saved image named lwm64.
After some tests I found that the Lisp image has its working directory set to the same directory from which you start it, and indeed there is a command line arg -eval to allow the evaluation of an initial form.
However, even after several attempts at quoting the initial command '(load "test/compile-asdf.lisp")', I cannot get it past Bash to send it to the Lisp image. Sorry for being so lame on shell scripting -- it has been many years since I was deep into them.
No need to apologize to me. I'm an unregenerate T shell user, so my understanding of these scripts is pretty much at the Cargo Cult level.
The error I keep getting is:
bash-3.2$ test/run-tests.sh lwm64
test/run-tests.sh: line 57: (load "test/compile-asdf.lisp"): No such file or directory
It looks to me as if the argument we are passing to the LW runtime is being interpreted as a file to load, not an s-expression to be evaluated. Does that sound right to you? E.g., in ACL, we start this up as something like alisp -e '(load "test/compile-asdf.lisp")' and for SBCL sbcl --eval '(load "test/compile-asdf.lisp")' Is there some equivalent to -e/--eval that works for lispworks? Their documentation seems relatively unfriendly to text-mode folks like me; seems very focused on the IDE.
And when I quote it variously, the same message appears with my quoting intact.
So, since you only want to execute that simple command in Lisp, here is the resulting output after I enter it manually:
....
;; Processing Cross Reference InformationASDF compiled cleanly Error: The variable CODE is unbound. 1 (continue) Try evaluating CODE again. 2 Return the value of :CODE instead. 3 Specify a value to use this time instead of evaluating CODE. 4 Specify a value to set CODE to. 5 Try loading test/compile-asdf.lisp again. 6 Give up loading test/compile-asdf.lisp. 7 Try loading another file instead of test/compile-asdf.lisp. 8 (abort) Return to level 0. 9 Return to top-level loop. 10 Return from multiprocessing.
Type :b for backtrace or :c <option number> to proceed. Type :bug-form "<subject>" for a bug report template or :? for other options.
It looks to me as if compile-asdf is running to completion, through the cond expression in that file (hence the "ASDF compiled cleanly"), and then something is failing. The Processing Cross Reference string there suggests that it might be something the LW IDE is doing, but I can't guess what that would be. Maybe someone else on the list can suggest something.... best, r
Hi Robert,
E.g., in ACL, we start this up as something like
alisp -e '(load "test/compile-asdf.lisp")'
and for SBCL
sbcl --eval '(load "test/compile-asdf.lisp")'
Is there some equivalent to -e/--eval that works for lispworks?
No, you have the correct form for LW also (-eval). The problem seems to be Bash, in not recognizing the quoted & parenthesized expression as a runtime argument to be passed to LWM. THe -eval appears to be correct, even though it is nowhere documented. It is Bash who is complaining, not LWM. Dr. David McClain dbm@refined-audiometrics.com
On 2/27/10 Feb 27 -5:13 PM, David McClain wrote:
Hi Robert,
E.g., in ACL, we start this up as something like
alisp -e '(load "test/compile-asdf.lisp")'
and for SBCL
sbcl --eval '(load "test/compile-asdf.lisp")'
Is there some equivalent to -e/--eval that works for lispworks?
No, you have the correct form for LW also (-eval). The problem seems to be Bash, in not recognizing the quoted & parenthesized expression as a runtime argument to be passed to LWM. THe -eval appears to be correct, even though it is nowhere documented. It is Bash who is complaining, not LWM.
Someone else will have to step up to the plate for this one. I just looked at run-tests.sh, and the command in question is wrapped in single-quotes for bash's benefit, which should keep it from really seeing the enclosed double-quoted strings and parentheses. I'm still suspicious of that error string you sent out --- it really looked like bash (or LW) thinking that your string should be interpreted as a filename... Question: what's the surrounding environment? Linux? MacOSX? Windows with Cygwin? Something else? Would you consider emailing a lispworks patch for run-tests.sh? If we could apply it to the repo (i.e., even if it doesn't work on LW, if it doesn't /break/ any of the other lisps), then people could rally round to debug it. HtH, r
Hi Guys, I don't have a problem dealing with LW. But I want to have them work on things that are within their purview. The script problem is being reported by Bash, not LWM. The environment is Mac OS X 10.6.2 (Snow Leopard). I can invoke LWM from a command line just fine, including passing an initial SExpr for eval. Here is a direct example, performed in Emacs in Shell mode: bash-3.2$ lwm64 -siteinit - -init - -eval "(print :hello)" LispWorks(R): The Common Lisp Programming Environment Copyright (C) 1987-2009 LispWorks Ltd. All rights reserved. Version 6.0.0 Saved by dbmcclain as lwm64, at 27 Feb 2010 6:37 User dbmcclain on RoadKill.local :HELLO CL-USER 1 > So the problem is one of the script in getting Bash to properly interpret the synthesized command. Dr. David McClain dbm@refined-audiometrics.com
Ahhh... yes, in doing my manual test, I discovered that the submitted -eval argument needs to be a string. Furthermore, when I modify your script with 'echo' so that I can examine what is being performed, here are the results: ;; Script as modified in the 4th line.. do_tests() { command=$1 eval=$2 fasl_ext=$3 rm -f *.$fasl_ext ~/.cache/common-lisp/"`pwd`"/*.$fasl_ext || true ( cd .. && echo $command $eval "(load \"test/compile-asdf.lisp\")" ) if [ $? -eq 0 ] ; then echo "Compiled OK" test_count=0 bash-3.2$ test/run-tests.sh lwm64 (load "test/compile-asdf.lisp") <<---- AHA! Wrong! Compiled OK Testing: *.script test/run-tests.sh: line 69: (load "*.script"): command not found Using , *.script failed -#--------------------------------------- Using Ran 1 tests: 0 passing and 1 failing failing test(s): *.script -#--------------------------------------- bash-3.2$ So $command, $eval, and $fasl_ext must not be getting set properly On Feb 27, 2010, at 16:54 PM, David McClain wrote:
Hi Guys,
I don't have a problem dealing with LW. But I want to have them work on things that are within their purview.
The script problem is being reported by Bash, not LWM. The environment is Mac OS X 10.6.2 (Snow Leopard).
I can invoke LWM from a command line just fine, including passing an initial SExpr for eval. Here is a direct example, performed in Emacs in Shell mode:
bash-3.2$ lwm64 -siteinit - -init - -eval "(print :hello)" LispWorks(R): The Common Lisp Programming Environment Copyright (C) 1987-2009 LispWorks Ltd. All rights reserved. Version 6.0.0 Saved by dbmcclain as lwm64, at 27 Feb 2010 6:37 User dbmcclain on RoadKill.local
:HELLO
CL-USER 1 >
So the problem is one of the script in getting Bash to properly interpret the synthesized command.
Dr. David McClain dbm@refined-audiometrics.com
_______________________________________________ asdf-devel mailing list asdf-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/asdf-devel
Dr. David McClain dbm@refined-audiometrics.com
What is your first argument to run-tests.sh? Did you modify run-tests.sh to support it? See the case "$lisp" statement. Please pull the latest git, it has some additional sanity checking. [ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] The problem with most conspiracy theories is that they seem to believe that for a group of people to behave in a way detrimental to the common good requires intent. The problem with most statist theories is that they seem to believe that for a group of people to behave in a way beneficial to the common good requires intent. On 27 February 2010 19:02, David McClain <dbm@refined-audiometrics.com> wrote:
Ahhh... yes, in doing my manual test, I discovered that the submitted -eval argument needs to be a string. Furthermore, when I modify your script with 'echo' so that I can examine what is being performed, here are the results:
;; Script as modified in the 4th line.. do_tests() { command=$1 eval=$2 fasl_ext=$3 rm -f *.$fasl_ext ~/.cache/common-lisp/"`pwd`"/*.$fasl_ext || true ( cd .. && echo $command $eval "(load \"test/compile-asdf.lisp\")" ) if [ $? -eq 0 ] ; then echo "Compiled OK" test_count=0
bash-3.2$ test/run-tests.sh lwm64 (load "test/compile-asdf.lisp") <<---- AHA! Wrong! Compiled OK Testing: *.script test/run-tests.sh: line 69: (load "*.script"): command not found Using , *.script failed -#--------------------------------------- Using Ran 1 tests: 0 passing and 1 failing failing test(s): *.script -#--------------------------------------- bash-3.2$
So $command, $eval, and $fasl_ext must not be getting set properly
On Feb 27, 2010, at 16:54 PM, David McClain wrote:
Hi Guys, I don't have a problem dealing with LW. But I want to have them work on things that are within their purview. The script problem is being reported by Bash, not LWM. The environment is Mac OS X 10.6.2 (Snow Leopard). I can invoke LWM from a command line just fine, including passing an initial SExpr for eval. Here is a direct example, performed in Emacs in Shell mode: bash-3.2$ lwm64 -siteinit - -init - -eval "(print :hello)" LispWorks(R): The Common Lisp Programming Environment Copyright (C) 1987-2009 LispWorks Ltd. All rights reserved. Version 6.0.0 Saved by dbmcclain as lwm64, at 27 Feb 2010 6:37 User dbmcclain on RoadKill.local :HELLO CL-USER 1 > So the problem is one of the script in getting Bash to properly interpret the synthesized command.
Dr. David McClain dbm@refined-audiometrics.com
_______________________________________________ asdf-devel mailing list asdf-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/asdf-devel
Dr. David McClain dbm@refined-audiometrics.com
_______________________________________________ asdf-devel mailing list asdf-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/asdf-devel
As shown in my last message: bash-3.2$ test/run-tests.sh lwm64 (load "test/compile-asdf.lisp") <<---- AHA! Wrong! The first argument is the name of my binary "lwm64" which was created to present a shell interface, no IDE. It is otherwise a bare LWM image with no extra features added. When LWM64 loads up, it assumes the working directory that was in effect in the shell at the time the command was invoked. So to run the test script properly, (according to your -eval argument) we have to be in the directory above the test subdirectory when the LWM64 is invoked. But again... the argument to LWM must be a *Lisp String* following the -eval command line argument. And that probably accounts for some of the problem. How you quote a string argument in a shell script is another problem... - DM On Feb 27, 2010, at 17:13 PM, Faré wrote:
What is your first argument to run-tests.sh? Did you modify run-tests.sh to support it? See the case "$lisp" statement.
Please pull the latest git, it has some additional sanity checking.
[ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] The problem with most conspiracy theories is that they seem to believe that for a group of people to behave in a way detrimental to the common good requires intent.
The problem with most statist theories is that they seem to believe that for a group of people to behave in a way beneficial to the common good requires intent.
On 27 February 2010 19:02, David McClain <dbm@refined-audiometrics.com> wrote:
Ahhh... yes, in doing my manual test, I discovered that the submitted -eval argument needs to be a string. Furthermore, when I modify your script with 'echo' so that I can examine what is being performed, here are the results:
;; Script as modified in the 4th line.. do_tests() { command=$1 eval=$2 fasl_ext=$3 rm -f *.$fasl_ext ~/.cache/common-lisp/"`pwd`"/*.$fasl_ext || true ( cd .. && echo $command $eval "(load \"test/compile-asdf.lisp\")" ) if [ $? -eq 0 ] ; then echo "Compiled OK" test_count=0
bash-3.2$ test/run-tests.sh lwm64 (load "test/compile-asdf.lisp") <<---- AHA! Wrong! Compiled OK Testing: *.script test/run-tests.sh: line 69: (load "*.script"): command not found Using , *.script failed -#--------------------------------------- Using Ran 1 tests: 0 passing and 1 failing failing test(s): *.script -#--------------------------------------- bash-3.2$
So $command, $eval, and $fasl_ext must not be getting set properly
On Feb 27, 2010, at 16:54 PM, David McClain wrote:
Hi Guys, I don't have a problem dealing with LW. But I want to have them work on things that are within their purview. The script problem is being reported by Bash, not LWM. The environment is Mac OS X 10.6.2 (Snow Leopard). I can invoke LWM from a command line just fine, including passing an initial SExpr for eval. Here is a direct example, performed in Emacs in Shell mode: bash-3.2$ lwm64 -siteinit - -init - -eval "(print :hello)" LispWorks(R): The Common Lisp Programming Environment Copyright (C) 1987-2009 LispWorks Ltd. All rights reserved. Version 6.0.0 Saved by dbmcclain as lwm64, at 27 Feb 2010 6:37 User dbmcclain on RoadKill.local :HELLO CL-USER 1 > So the problem is one of the script in getting Bash to properly interpret the synthesized command.
Dr. David McClain dbm@refined-audiometrics.com
_______________________________________________ asdf-devel mailing list asdf-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/asdf-devel
Dr. David McClain dbm@refined-audiometrics.com
_______________________________________________ asdf-devel mailing list asdf-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/asdf-devel
Dr. David McClain dbm@refined-audiometrics.com
And yes, I modified that portion of the script that you mentioned: lispworks) if type lispworks ; then fasl_ext="64xfasl" command=`which lwm64` command="$command -siteinit - -init -" eval="-eval" fi ;; On Feb 27, 2010, at 17:13 PM, Faré wrote:
What is your first argument to run-tests.sh? Did you modify run-tests.sh to support it? See the case "$lisp" statement.
Please pull the latest git, it has some additional sanity checking.
[ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] The problem with most conspiracy theories is that they seem to believe that for a group of people to behave in a way detrimental to the common good requires intent.
The problem with most statist theories is that they seem to believe that for a group of people to behave in a way beneficial to the common good requires intent.
On 27 February 2010 19:02, David McClain <dbm@refined-audiometrics.com> wrote:
Ahhh... yes, in doing my manual test, I discovered that the submitted -eval argument needs to be a string. Furthermore, when I modify your script with 'echo' so that I can examine what is being performed, here are the results:
;; Script as modified in the 4th line.. do_tests() { command=$1 eval=$2 fasl_ext=$3 rm -f *.$fasl_ext ~/.cache/common-lisp/"`pwd`"/*.$fasl_ext || true ( cd .. && echo $command $eval "(load \"test/compile-asdf.lisp\")" ) if [ $? -eq 0 ] ; then echo "Compiled OK" test_count=0
bash-3.2$ test/run-tests.sh lwm64 (load "test/compile-asdf.lisp") <<---- AHA! Wrong! Compiled OK Testing: *.script test/run-tests.sh: line 69: (load "*.script"): command not found Using , *.script failed -#--------------------------------------- Using Ran 1 tests: 0 passing and 1 failing failing test(s): *.script -#--------------------------------------- bash-3.2$
So $command, $eval, and $fasl_ext must not be getting set properly
On Feb 27, 2010, at 16:54 PM, David McClain wrote:
Hi Guys, I don't have a problem dealing with LW. But I want to have them work on things that are within their purview. The script problem is being reported by Bash, not LWM. The environment is Mac OS X 10.6.2 (Snow Leopard). I can invoke LWM from a command line just fine, including passing an initial SExpr for eval. Here is a direct example, performed in Emacs in Shell mode: bash-3.2$ lwm64 -siteinit - -init - -eval "(print :hello)" LispWorks(R): The Common Lisp Programming Environment Copyright (C) 1987-2009 LispWorks Ltd. All rights reserved. Version 6.0.0 Saved by dbmcclain as lwm64, at 27 Feb 2010 6:37 User dbmcclain on RoadKill.local :HELLO CL-USER 1 > So the problem is one of the script in getting Bash to properly interpret the synthesized command.
Dr. David McClain dbm@refined-audiometrics.com
_______________________________________________ asdf-devel mailing list asdf-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/asdf-devel
Dr. David McClain dbm@refined-audiometrics.com
_______________________________________________ asdf-devel mailing list asdf-devel@common-lisp.net http://common-lisp.net/cgi-bin/mailman/listinfo/asdf-devel
Dr. David McClain dbm@refined-audiometrics.com
Of course the if type lispworks looks for lispworks in your path, then fails silently. If you want the lispworks binary to be called lwm64 instead of lispworks, you have to change that, too... [ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] If a vegetarian eats vegetables, what does a humanitarian eat? — Mark Twain On 27 February 2010 19:21, David McClain <dbm@refined-audiometrics.com> wrote:
And yes, I modified that portion of the script that you mentioned:
lispworks) if type lispworks ; then fasl_ext="64xfasl" command=`which lwm64` command="$command -siteinit - -init -" eval="-eval" fi ;;
On Feb 27, 2010, at 17:13 PM, Faré wrote:
What is your first argument to run-tests.sh? Did you modify run-tests.sh to support it? See the case "$lisp" statement.
Please pull the latest git, it has some additional sanity checking.
[ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] The problem with most conspiracy theories is that they seem to believe that for a group of people to behave in a way detrimental to the common good requires intent.
The problem with most statist theories is that they seem to believe that for a group of people to behave in a way beneficial to the common good requires intent.
On 27 February 2010 19:02, David McClain <dbm@refined-audiometrics.com> wrote:
Ahhh... yes, in doing my manual test, I discovered that the submitted -eval
argument needs to be a string.
Furthermore, when I modify your script with 'echo' so that I can examine
what is being performed, here are the results:
;; Script as modified in the 4th line..
do_tests() {
command=$1 eval=$2 fasl_ext=$3
rm -f *.$fasl_ext ~/.cache/common-lisp/"`pwd`"/*.$fasl_ext || true
( cd .. && echo $command $eval "(load \"test/compile-asdf.lisp\")" )
if [ $? -eq 0 ] ; then
echo "Compiled OK"
test_count=0
bash-3.2$ test/run-tests.sh lwm64
(load "test/compile-asdf.lisp") <<---- AHA! Wrong!
Compiled OK
Testing: *.script
test/run-tests.sh: line 69: (load "*.script"): command not found
Using , *.script failed
-#---------------------------------------
Using
Ran 1 tests:
0 passing and 1 failing
failing test(s): *.script
-#---------------------------------------
bash-3.2$
So $command, $eval, and $fasl_ext must not be getting set properly
On Feb 27, 2010, at 16:54 PM, David McClain wrote:
Hi Guys,
I don't have a problem dealing with LW. But I want to have them work on
things that are within their purview.
The script problem is being reported by Bash, not LWM. The environment is
Mac OS X 10.6.2 (Snow Leopard).
I can invoke LWM from a command line just fine, including passing an initial
SExpr for eval. Here is a direct example, performed in Emacs in Shell mode:
bash-3.2$ lwm64 -siteinit - -init - -eval "(print :hello)"
LispWorks(R): The Common Lisp Programming Environment
Copyright (C) 1987-2009 LispWorks Ltd. All rights reserved.
Version 6.0.0
Saved by dbmcclain as lwm64, at 27 Feb 2010 6:37
User dbmcclain on RoadKill.local
:HELLO
CL-USER 1 >
So the problem is one of the script in getting Bash to properly interpret
the synthesized command.
Dr. David McClain
dbm@refined-audiometrics.com
_______________________________________________
asdf-devel mailing list
asdf-devel@common-lisp.net
http://common-lisp.net/cgi-bin/mailman/listinfo/asdf-devel
Dr. David McClain
dbm@refined-audiometrics.com
_______________________________________________
asdf-devel mailing list
asdf-devel@common-lisp.net
http://common-lisp.net/cgi-bin/mailman/listinfo/asdf-devel
Dr. David McClain dbm@refined-audiometrics.com
Ahh yes, of course.. However, when I try the command line: bash-3.2$ test/run-tests.sh lispworks test/run-tests.sh: line 162: type: lispworks: not found test/run-tests.sh: line 57: (load "test/compile-asdf.lisp"): No such file or directory bash-3.2$ So here is the script... what am I, or have I done incorrectly? #!/bin/sh # do_tests {lisp invocation} {scripts-regex} # - read lisp forms one at a time from standard input # - quit with exit status 0 on getting eof # - quit with exit status >0 if an unhandled error occurs export CL_SOURCE_REGISTRY="$PWD" unset DEBUG_ASDF_TEST while getopts "duh" OPTION do case $OPTION in d) export DEBUG_ASDF_TEST=t ;; u) usage exit 1 ;; h) usage exit 1 ;; esac done shift $(($OPTIND - 1)) if [ x"$1" = "xhelp" ]; then usage exit 1 fi if [ -z "$2" ]; then scripts="*.script" else scripts="$2" fi sok=1 usage () { echo "$0 [lisp invocation] [scripts-regex]" echo " - read lisp forms one at a time from matching scripts" echo " - quit with exit status 0 on getting eof" echo " - quit with exit status >0 if an unhandled error occurs" echo " you need to supply the .script in the second argument" echo " lisps include sbcl, clisp, allegro and allegromodern" echo "OPTIONS:" echo " -d -- debug mode" echo " -u -h -- show this message." } do_tests() { command=$1 eval=$2 fasl_ext=$3 rm -f *.$fasl_ext ~/.cache/common-lisp/"`pwd`"/*.$fasl_ext || true ( cd .. && $command $eval '(load "test/compile-asdf.lisp")' ) if [ $? -eq 0 ] ; then echo "Compiled OK" test_count=0 test_pass=0 test_fail=0 failed_list="" for i in $scripts ; do echo "Testing: $i" >&2 test_count=`expr "$test_count" + 1` rm -f *.$fasl_ext ~/.cache/common-lisp/"`pwd`"/*.$fasl_ext || true if $command $eval "(load \"$i\")" ; then echo "Using $command, $i passed" >&2 test_pass=`expr "$test_pass" + 1` else echo "Using $command, $i failed" >&2 test_fail=`expr "$test_fail" + 1` failed_list="$failed_list $i" sok=0 fi done echo >&2 echo "-#---------------------------------------" >&2 echo "Using $command" >&2 echo "Ran $test_count tests: " >&2 echo " $test_pass passing and $test_fail failing" >&2 if [ $test_fail -eq 0 ] ; then echo "all tests apparently successful" >&2 else echo "failing test(s): $failed_list" >&2 fi echo "-#---------------------------------------" >&2 echo >&2 fi } # terminate on error set -e lisp=$1 if [ -z $1 ] ; then lisp="sbcl" fi case "$lisp" in sbcl) if type sbcl ; then fasl_ext="fasl" command="sbcl --noinform --userinit /dev/null --sysinit /dev/null" nodebug="--disable-debugger" eval="--eval" fi ;; clisp) if type clisp ; then fasl_ext="fas" command=`which clisp` command="$command -norc -ansi -I " nodebug="-on-error exit" eval="-x" fi ;; allegro) if type alisp ; then fasl_ext="fasl" command="alisp -q " nodebug="-batch" eval="-e" fi ;; allegromodern) if type mlisp ; then fasl_ext="fasl" command="mlisp -q" nodebug="-batch" eval="-e" fi ;; ccl) if type ccl ; then case `uname -s` in Linux) fasl_os=lx ;; Darwin) fasl_os=dx ;; esac case `uname -m` in x86_64|ppc64) fasl_bits=64 ;; i?86|ppc) fasl_bits=32 ;; esac fasl_ext="${fasl_os}${fasl_bits}fsl" command="ccl --no-init --quiet" nodebug="--batch" eval="--eval" fi ;; cmucl) if type lisp ; then fasl_ext="x86f" command="lisp -noinit" nodebug="-batch" eval="-eval" fi ;; ecl) if type ecl ; then fasl_ext="fas" command=`which ecl` command="$command -norc" eval="-eval" fi ;; lispworks) if type lispworks ; then fasl_ext="64xfasl" command=`which lwm64` command="$command -siteinit - -init -" eval="-eval" fi ;; esac if [ -z "${DEBUG_ASDF_TEST}" ] ; then command="$command $nodebug" fi create_config () { mkdir -p ../tmp/test-source-registry-conf.d ../tmp/test-asdf-output-translations-conf.d } clean_up () { rm -rf ../tmp/test-source-registry-conf.d ../tmp/test-asdf-output-translations-conf.d } if [ -z "$command" ] ; then echo "Error: cannot find or do not know how to run Lisp named $lisp" else create_config mkdir -p results echo $command thedate=`date "+%Y-%m-%d"` do_tests "$command" "$eval" "$fasl_ext" 2>&1 | \ tee "results/${lisp}.text" "results/${lisp}-${thedate}.save" clean_up fi On Feb 27, 2010, at 18:53 PM, Faré wrote:
Of course the if type lispworks looks for lispworks in your path, then fails silently. If you want the lispworks binary to be called lwm64 instead of lispworks, you have to change that, too...
[ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] If a vegetarian eats vegetables, what does a humanitarian eat? — Mark Twain
On 27 February 2010 19:21, David McClain <dbm@refined-audiometrics.com> wrote:
And yes, I modified that portion of the script that you mentioned:
lispworks) if type lispworks ; then fasl_ext="64xfasl" command=`which lwm64` command="$command -siteinit - -init -" eval="-eval" fi ;;
On Feb 27, 2010, at 17:13 PM, Faré wrote:
What is your first argument to run-tests.sh? Did you modify run-tests.sh to support it? See the case "$lisp" statement.
Please pull the latest git, it has some additional sanity checking.
[ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] The problem with most conspiracy theories is that they seem to believe that for a group of people to behave in a way detrimental to the common good requires intent.
The problem with most statist theories is that they seem to believe that for a group of people to behave in a way beneficial to the common good requires intent.
On 27 February 2010 19:02, David McClain <dbm@refined-audiometrics.com> wrote:
Ahhh... yes, in doing my manual test, I discovered that the submitted -eval
argument needs to be a string.
Furthermore, when I modify your script with 'echo' so that I can examine
what is being performed, here are the results:
;; Script as modified in the 4th line..
do_tests() {
command=$1 eval=$2 fasl_ext=$3
rm -f *.$fasl_ext ~/.cache/common-lisp/"`pwd`"/*.$fasl_ext || true
( cd .. && echo $command $eval "(load \"test/compile-asdf.lisp\")" )
if [ $? -eq 0 ] ; then
echo "Compiled OK"
test_count=0
bash-3.2$ test/run-tests.sh lwm64
(load "test/compile-asdf.lisp") <<---- AHA! Wrong!
Compiled OK
Testing: *.script
test/run-tests.sh: line 69: (load "*.script"): command not found
Using , *.script failed
-#---------------------------------------
Using
Ran 1 tests:
0 passing and 1 failing
failing test(s): *.script
-#---------------------------------------
bash-3.2$
So $command, $eval, and $fasl_ext must not be getting set properly
On Feb 27, 2010, at 16:54 PM, David McClain wrote:
Hi Guys,
I don't have a problem dealing with LW. But I want to have them work on
things that are within their purview.
The script problem is being reported by Bash, not LWM. The environment is
Mac OS X 10.6.2 (Snow Leopard).
I can invoke LWM from a command line just fine, including passing an initial
SExpr for eval. Here is a direct example, performed in Emacs in Shell mode:
bash-3.2$ lwm64 -siteinit - -init - -eval "(print :hello)"
LispWorks(R): The Common Lisp Programming Environment
Copyright (C) 1987-2009 LispWorks Ltd. All rights reserved.
Version 6.0.0
Saved by dbmcclain as lwm64, at 27 Feb 2010 6:37
User dbmcclain on RoadKill.local
:HELLO
CL-USER 1 >
So the problem is one of the script in getting Bash to properly interpret
the synthesized command.
Dr. David McClain
dbm@refined-audiometrics.com
_______________________________________________
asdf-devel mailing list
asdf-devel@common-lisp.net
http://common-lisp.net/cgi-bin/mailman/listinfo/asdf-devel
Dr. David McClain
dbm@refined-audiometrics.com
_______________________________________________
asdf-devel mailing list
asdf-devel@common-lisp.net
http://common-lisp.net/cgi-bin/mailman/listinfo/asdf-devel
Dr. David McClain dbm@refined-audiometrics.com
Dr. David McClain dbm@refined-audiometrics.com
Please 1- use the latest ASDF from git 2- have a symlink to lwm64 in your PATH named lispworks [ François-René ÐVB Rideau | Reflection&Cybernethics | http://fare.tunes.org ] The Slogan of "Language Independence" is often but the pride that self-ignorant monolinguists put in not calling "language" the collection of barkings and grunts they invent to interact with their computers. — Faré On 27 February 2010 21:03, David McClain <dbm@refined-audiometrics.com> wrote:
Ahh yes, of course.. However, when I try the command line: bash-3.2$ test/run-tests.sh lispworks test/run-tests.sh: line 162: type: lispworks: not found test/run-tests.sh: line 57: (load "test/compile-asdf.lisp"): No such file or directory bash-3.2$
So here is the script... what am I, or have I done incorrectly?
participants (3)
-
David McClain -
Faré -
Robert Goldman