Steel Bank Common Lisp

Kevin Smith <k2msmith@...> writes:

> Please bear with me as I am a beginner with sbcl and lisp environments in general.  I am using SBCL 1.0.39 compiled on Mac OSX from macports. The primary packages I am interested in using with SBCL right now are asdf, slime, cl-opengl (and cl-glu, cl-glut which come with it and cffi (for building wrappers on "c" code functions).
>
>
> Question 1
>
> All seems to be working.  I have written some test code as well as loaded a few example files from cl-opengl file and things seem to work - although I am wondering if I have asdf and my package collection properly installed.  
>
> When I execute function from the sbcl prompt:  (compile-file "intro.lisp"), I get
>
> ; compiling file "/Users/kevinsmith/intro.lisp" (written 29 JUN 2010 04:58:26 PM):
> ; compiling (REQUIRE :ASDF)
> ; compiling (REQUIRE (QUOTE CL-OPENGL))
> ; compiling (REQUIRE (QUOTE CL-GLU))
> ; compiling (REQUIRE (QUOTE CL-GLUT)); 
> ; compilation unit aborted
> ;   caught 1 fatal ERROR condition
> ;
> ; compilation aborted because of fatal error:
> ;   SB-INT:SIMPLE-READER-PACKAGE-ERROR at 231 (line 6, column 32) on #<SB-SYS:FD-STREAM
> ;                                                                      for "file /Users/kevinsmith/intro.lisp"
> ;                                                                      {100381E971}>:
> ;     package "GLUT" not found
> ;   
> ; compilation aborted after 0:00:00.005
> NIL
> T
>
> an excerpt of the source file looks like this:
>
> (require :asdf)                 ; need ASDF to load other things
> (require 'cl-opengl)   ; load OpenGL bindings
> (require 'cl-glu)      ; load GLU bindings
> (require 'cl-glut)     ; load GLUT bindings
>
> (defclass my-window (glut:window)
>   ()
>   (:default-initargs :width 400 :height 300
>                      :title "My Window Title"
>                      :x 100 :y 100
>                      :mode '(:double :rgb :depth)))
>
Put all your requires inside (eval-when (:compile-toplevel :load-toplevel :execute) ...)

> Question 2
>
> What are my options for the sbcl shell/evaluation loop ?   With no arrow keys for history and the annoying way it handles errors, requiring you to type in a different number to go back to the previous state, it makes it a bit rough for me. Running everything through emacs seems to be the only alternative, but a nicer shell (similar to haskell shell or others) would be nice if available.  (or maybe something like the inspector window in other lisps I've seen).
Try rlwrap (http://utopia.knoware.nl/~hlub/rlwrap/ )

-- 
With Best Regards, Stas.

Stas, Thanks much for your tips here. The "eval-when" function and rlwrap  
worked well for me. I found the cliki has some example configurations:

http://www.cliki.net/rlwrap



On Jun 30, 2010 1:16am, Stas Boukarev <stassats@...> wrote:
> Kevin Smith k2msmith@...> writes:



> > Please bear with me as I am a beginner with sbcl and lisp environments  
> in general. I am using SBCL 1.0.39 compiled on Mac OSX from macports. The  
> primary packages I am interested in using with SBCL right now are asdf,  
> slime, cl-opengl (and cl-glu, cl-glut which come with it and cffi (for  
> building wrappers on "c" code functions).

> >

> >

> > Question 1

> >

> > All seems to be working. I have written some test code as well as  
> loaded a few example files from cl-opengl file and things seem to work -  
> although I am wondering if I have asdf and my package collection properly  
> installed.

> >

> > When I execute function from the sbcl prompt:  
> (compile-file "intro.lisp"), I get

> >

> > ; compiling file "/Users/kevinsmith/intro.lisp" (written 29 JUN 2010  
> 04:58:26 PM):

> > ; compiling (REQUIRE :ASDF)

> > ; compiling (REQUIRE (QUOTE CL-OPENGL))

> > ; compiling (REQUIRE (QUOTE CL-GLU))

> > ; compiling (REQUIRE (QUOTE CL-GLUT));

> > ; compilation unit aborted

> > ; caught 1 fatal ERROR condition

> > ;

> > ; compilation aborted because of fatal error:

> > ; SB-INT:SIMPLE-READER-PACKAGE-ERROR at 231 (line 6, column 32) on #
> > ; for "file /Users/kevinsmith/intro.lisp"

> > ; {100381E971}>:

> > ; package "GLUT" not found

> > ;

> > ; compilation aborted after 0:00:00.005

> > NIL

> > T

> >

> > an excerpt of the source file looks like this:

> >

> > (require :asdf) ; need ASDF to load other things

> > (require 'cl-opengl) ; load OpenGL bindings

> > (require 'cl-glu) ; load GLU bindings

> > (require 'cl-glut) ; load GLUT bindings

> >

> > (defclass my-window (glut:window)

> > ()

> > (:default-initargs :width 400 :height 300

> > :title "My Window Title"

> > :x 100 :y 100

> > :mode '(:double :rgb :depth)))

> >

> Put all your requires inside (eval-when  
> (:compile-toplevel :load-toplevel :execute) ...)



> > Question 2

> >

> > What are my options for the sbcl shell/evaluation loop ? With no arrow  
> keys for history and the annoying way it handles errors, requiring you to  
> type in a different number to go back to the previous state, it makes it  
> a bit rough for me. Running everything through emacs seems to be the only  
> alternative, but a nicer shell (similar to haskell shell or others) would  
> be nice if available. (or maybe something like the inspector window in  
> other lisps I've seen).

> Try rlwrap (http://utopia.knoware.nl/~hlub/rlwrap/ )



> --

> With Best Regards, Stas.

Jerry James <loganjerry@...> writes:

> I'm working my way through the compiler notes on a small project,
> trying to understand what each one means (and fix those that are
> fixable).  I've encountered some notes relating to the use of the
> system-area-pointer type, which is used to interface with some C code.
>  I can distill my question down to this code:
>
> (proclaim '(optimize (debug 0) (safety 0) (space 1) (speed 3)
> 	    (compilation-speed 3)))
>
> ;; Gets a compiler note about SAP to pointer coercion
> (defun sap-test1 (ptr fun)
>   (declare (type system-area-pointer ptr)
> 	   (type (function (system-area-pointer) fixnum) fun))
>   (funcall fun (sb-sys:sap+ ptr 10)))
>
> ;; Doesn't get a compiler note
> (defun sap-test2 (ptr fun)
>   (declare (type system-area-pointer ptr)
> 	   (type (function (system-area-pointer) fixnum) fun))
>   (funcall fun ptr))
>
> which produces this output when compiled:
>  [...]
> Why is the SAP to pointer coercion taking place in the first version?  Thanks,

In general, no matter what the safety/speed/etc optimization qualities,
Lisp functions must receive their arguments as "boxed" -- in a form
which includes type information -- so that they can detect when things
go wrong, and so that function redefinition is not prevented from
working by some non-boxed argument being constructed at a call site.  So
both SAP-TEST1 and -TEST2 receive their arguments as pointers.

SAP-TEST2 merely passes those pointers through to FUN; so no extra work
needs to be done at all.  SAP-TEST1 constructs a new SAP.  If something
were done with that SAP locally, such as derefing it, there would be no
need to convert it into a "pointer" (which here means heap-allocating
the SAP data and constructing an appropriately tagged pointer to that
heap allocation); however, since the SAP gets passed to a function which
cannot accept an unboxed argument, that pointer construction must take
place.

If FUN were an inline function, or a local function, rather than an
external function (however arbitrarily tightly type-declared), this
pointer coercion (heap allocation) wouldn't need to happen.

Cheers,

Christophe

2010/6/24 Christophe Rhodes <csr21@...>:
> If FUN were an inline function, or a local function, rather than an
> external function (however arbitrarily tightly type-declared), this
> pointer coercion (heap allocation) wouldn't need to happen.

What might not be obvious from Christophe's reply is the most common
way (in my experience) for FUN to be a local function. If SAP-TEST1 is
inlined and FUN is a local function for the call site.

The trick is to put all your mapping functions (which is what I assume
this was derived from) in one file, all declaimed inline. Their
compilation will generate lots of notes, but that's okay. What you
care about is the call sites, and making sure you get the
local-function optimizations you need there.

In article 
<AANLkTin9a_FGQ_xcdtu6hdhRhAx_zmGRaSWg1CSfz1qG@...>,
 "Thomas F. Burdick" <tburdick@...> wrote:

> 2010/6/24 Christophe Rhodes <csr21@...>:
> > If FUN were an inline function, or a local function, rather than an
> > external function (however arbitrarily tightly type-declared), this
> > pointer coercion (heap allocation) wouldn't need to happen.
> 
> What might not be obvious from Christophe's reply is the most common
> way (in my experience) for FUN to be a local function. If SAP-TEST1 is
> inlined and FUN is a local function for the call site.
> 
> The trick is to put all your mapping functions (which is what I assume
> this was derived from) in one file, all declaimed inline. Their
> compilation will generate lots of notes, but that's okay. What you
> care about is the call sites, and making sure you get the
> local-function optimizations you need there.

I had the idea two days ago to make SBCL NOT emit the
note about having to box a return value (be it machine
word, or SAP) in case the user declared the function's
return value accordingly.

(It's not directly related to the OP's example code
but another common annoyance.) 

  -T.

On Thu, Jun 24, 2010 at 3:17 AM, Thomas F. Burdick <tburdick@...> wrote:
> 2010/6/24 Christophe Rhodes <csr21@...>:
>> If FUN were an inline function, or a local function, rather than an
>> external function (however arbitrarily tightly type-declared), this
>> pointer coercion (heap allocation) wouldn't need to happen.
>
> What might not be obvious from Christophe's reply is the most common
> way (in my experience) for FUN to be a local function. If SAP-TEST1 is
> inlined and FUN is a local function for the call site.
>
> The trick is to put all your mapping functions (which is what I assume
> this was derived from) in one file, all declaimed inline. Their
> compilation will generate lots of notes, but that's okay. What you
> care about is the call sites, and making sure you get the
> local-function optimizations you need there.

Thank you, Thomas and Christophe.  I appreciate the explanation.  Now
that I understand what's going on a little better, I've been able to
eliminate nearly half of the compiler notes we were getting.  It looks
like we'll probably have to live with the other half.

Regards,
-- 
Jerry James
http://www.jamezone.org/

* 2010-06-26 21:33 (+0300), Teemu Likonen wrote:

> The following expression with SIMPLE-ERROR condition doesn't work. My
> restart is not active any more and I don't understand why.
>
>     $ sbcl-eval "(error 'simple-error)"
>     [See the backtrace below]
>
> There is some "unbound condition slot" error in the backtrace. What is
> happening?

OK, I figured it out that the condition object has no slot named
CONDITION and therefore the object can't be pretty-printed with FORMAT.
It's unclear to me why the slot doesn't exist, though.

Anyway, I added a SLOT-EXISTS-P test and now SIMPLE-ERROR condition is
handled too:

--8<---------------cut here---------------start------------->8---
#!/home/dtw/local/bin/sbcl --script

(defun handler-eof (condition)
  (declare (ignore condition))
  (invoke-restart 'continue-eval "No value"))

(defun handler-t (condition)
  (invoke-restart 'continue-eval (if (slot-exists-p condition 'condition)
                                     condition
                                     "Unknown condition")))

(handler-bind ((end-of-file #'handler-eof)
               (t #'handler-t))
  (dolist (arg (rest sb-ext:*posix-argv*))
    (restart-case (format *standard-output* "~&~S~%"
                          (eval (read-from-string arg)))
      (continue-eval (msg)
        (format *error-output* "~&; ~A~%" msg)))))
--8<---------------cut here---------------end--------------->8---

In article <87lja1hcee.fsf@...>,
 Teemu Likonen <tlikonen@...> wrote:

> There is something in Common Lisp's (or SBCL's) conditions and restarts
> which I don't understand. Here's some background and an example code. I
> made a simple tool "sbcl-eval" which evaluates its command-line
> arguments as Lisp expressions and prints the results:
> 
>     $ sbcl-eval "" '(* 123 2)' blah pi '(foo)'
>     ; No value
>     246
>     ; The variable BLAH is unbound.
>     3.141592653589793d0
>     ; undefined function: FOO
> 
> The code is here:
> 
> --8<---------------cut here---------------start------------->8---
> #!/home/dtw/local/bin/sbcl --script
> 
> (defun handler-eof (condition)
>   (declare (ignore condition))
>   (invoke-restart 'continue-eval "No value"))
> 
> (defun handler-t (condition)
>   (invoke-restart 'continue-eval condition))
> 
> (handler-bind ((end-of-file #'handler-eof)
>                (t #'handler-t))
>   (dolist (arg (rest sb-ext:*posix-argv*))
>     (restart-case (format *standard-output* "~&~S~%"
>                           (eval (read-from-string arg)))
>       (continue-eval (msg)
>         (format *error-output* "~&; ~A~%" msg)))))
> --8<---------------cut here---------------end--------------->8---
> 
> The code does what I want but then I tried to test more its error
> handling. At least this works:
> 
>     $ sbcl-eval "(error 'error)"
>     ; Condition ERROR was signalled.
> 
> The following expression with SIMPLE-ERROR condition doesn't work. My
> restart is not active any more and I don't understand why.
> 
>     $ sbcl-eval "(error 'simple-error)"
>     [See the backtrace below]
> 
> There is some "unbound condition slot" error in the backtrace. What is
> happening?

If you M-. on simple-error in Slime, you'll be brought
into src/code/condition.lisp; unfortunately not to the right
toplevel-form, the definition of SIMPLE-ERROR, but the macro
definition of DEFINE-CONDITION -- my memory currently fails
for the exact reason for that mismatch, but iirc it's due to
some ordering issues during compilation of SBCL itself.

Anyway, you can just search for "simple-error" in that file,
and you'll see: SIMPLE-ERROR is a SIMPLE-CONDITION which consists
of two slots, in particular its FORMAT-CONTROL slot has no initform.
If you know look into its reporter, SIMPLE-CONDITION-PRINTER, you'll
see that printing SIMPLE-ERRORs depends on the presence of a
FORMAT-CONTROL value.

Indeed, if you try (ERROR 'SIMPLE-ERROR) at the REPL, you'll see
SBCL complain for

  Unable to display error condition: unbound condition slot:
  SB-KERNEL:FORMAT-CONTROL

As SIMPLE-ERROR and SIMPLE-CONDITION are condition specified by
the standard (they're external symbols of the CL package), you
can actually look up their specifications via C-c C-d h in Slime.

Arguably, the definition of SIMPLE-CONDITION should have a
:INITFORM (SB-INT:MISSING-ARGUMENT :FORMAT-CONTROL) in the
slot definition of FORMAT-CONTROL to bail out early.

  -T.

In article <87hbkphav1.fsf@...>,
 Teemu Likonen <tlikonen@...> wrote:

> * 2010-06-26 21:33 (+0300), Teemu Likonen wrote:
> 
> > The following expression with SIMPLE-ERROR condition doesn't work. My
> > restart is not active any more and I don't understand why.
> >
> >     $ sbcl-eval "(error 'simple-error)"
> >     [See the backtrace below]
> >
> > There is some "unbound condition slot" error in the backtrace. What is
> > happening?
> 
> OK, I figured it out that the condition object has no slot named
> CONDITION and therefore the object can't be pretty-printed with FORMAT.
> It's unclear to me why the slot doesn't exist, though.

Are you sure the name of the slot is CONDITION? It should complain
about a missing slot named FORMAT-CONTROL.
 
> Anyway, I added a SLOT-EXISTS-P test and now SIMPLE-ERROR condition is
> handled too:
> 
> --8<---------------cut here---------------start------------->8---
> #!/home/dtw/local/bin/sbcl --script
> 
> (defun handler-eof (condition)
>   (declare (ignore condition))
>   (invoke-restart 'continue-eval "No value"))
> 
> (defun handler-t (condition)
>   (invoke-restart 'continue-eval (if (slot-exists-p condition 'condition)
>                                      condition
>                                      "Unknown condition")))
> 
> (handler-bind ((end-of-file #'handler-eof)
>                (t #'handler-t))
>   (dolist (arg (rest sb-ext:*posix-argv*))
>     (restart-case (format *standard-output* "~&~S~%"
>                           (eval (read-from-string arg)))
>       (continue-eval (msg)
>         (format *error-output* "~&; ~A~%" msg)))))
> --8<---------------cut here---------------end--------------->8---

N.B. This will /seem/ to work even in case I'm right and the slot in
question is FORMAT-CONTROL and not CONDITION.

  -T.

* 2010-06-27 11:48 (+0800), Tobias C. Rittweiler wrote:

> If you M-. on simple-error in Slime, you'll be brought into
> src/code/condition.lisp; unfortunately not to the right toplevel-form,
> the definition of SIMPLE-ERROR, but the macro definition of
> DEFINE-CONDITION -- my memory currently fails for the exact reason for
> that mismatch, but iirc it's due to some ordering issues during
> compilation of SBCL itself.
>
> Anyway, you can just search for "simple-error" in that file, and
> you'll see: SIMPLE-ERROR is a SIMPLE-CONDITION which consists of two
> slots, in particular its FORMAT-CONTROL slot has no initform. If you
> know look into its reporter, SIMPLE-CONDITION-PRINTER, you'll see that
> printing SIMPLE-ERRORs depends on the presence of a FORMAT-CONTROL
> value.

Thanks. It seems that I presumed wrong things and thought that
PRINT-OBJECT method (called by FORMAT "~A" in my code) should always
print something when given a condition. The HyperSpec says that
PRINT-OBJECT method is required for class STANDARD-OBJECT but it seems
that conditions are not inherited from that.

And the second message I wrote was just some more misunderstanding on
the top of misunderstanding. :-)

One of the goals with my simple "sbcl-eval" tool is to learn to handle
errors without the debugger and ugly backtraces. Now I think I
understand that the handler wasn't active because the error was
signalled in the handler+restart code itself (because of the missing
PRINT-OBJECT method called by FORMAT "~A"). I added HANDLER-CASE form in
the handler function and now my code catches those SIMPLE-ERRORs and
-CONDITIONs too.

--8<---------------cut here---------------start------------->8---
#!/home/dtw/local/bin/sbcl --script

(defun handler-eof (c)
  (declare (ignore c))
  (invoke-restart 'continue-eval "No value"))

(defun handler-condition (c)
  (invoke-restart 'continue-eval (handler-case (princ-to-string c)
                                   (condition () (type-of c)))))

(defun interrupt (c)
  (declare (ignore c))
  (terpri *error-output*)
  (sb-ext:quit :unix-status 1))

(handler-bind ((sb-sys:interactive-interrupt #'interrupt)
               (end-of-file #'handler-eof)
               (condition #'handler-condition))
  (dolist (arg (rest sb-ext:*posix-argv*))
    (restart-case (format *standard-output* "~&~S~%"
                          (eval (read-from-string arg)))
      (continue-eval (msg)
        (format *error-output* "~&; ~A~%" msg)))))
--8<---------------cut here---------------end--------------->8---

On Tue, 15 Jun 2010, Brett Hoerner wrote:

> While installing CLSQL I get the following,
>
>   ; loading system definition from /home/brett/.sbcl/systems/clsql.asd into
>   ; #<PACKAGE "ASDF0">
>
>   debugger invoked on a MISSING-COMPONENT in thread #<THREAD
>                                                       "initial thread" RUNNING
>                                                       {1002ABBEF1}>:
>     component "cffi-uffi-compat" not found
>
> I have installed UFFI, but this error still occurs.

cffi-uffi-compat lives in the cffi source tree; it is a compatibility 
layer meant to replace uffi.

Check that you have
/home/brett/.sbcl/systems/cffi-uffi-compat.asd
pointing to the actual file in the cffi source tree.

- Daniel

Hello Didier,

 There is the bug report on this:

https://bugs.launchpad.net/sbcl/+bug/338924

I am not sure about the exact match, but the issue seems to be related to
what you have encountered.


Regards,
Roman


2010/6/14 Didier Verna <didier@...>

>
>  Hi, two questions about getenv:
>
> 1/ should I use sb-posix:getenv or sb-ext:posix-getenv ?
>
> 2/ also, getenv seems to behave differently on Mac and on Linux
>   (probably because of differences at a the OS level):
>
>   (getenv nil) used to return nil which I kinda liked, but now on Linux
>   (1.0.39.15, although I can't tell exactly when it started) it throws
>   an error (see below). Is this intentional ? On my Mac on the other
>   hand, with 1.0.39.9, it still returns nil.
>
>
> Thanks a lot !
>
>
> * (sb-ext:posix-getenv nil)
> CORRUPTION WARNING in SBCL pid 13919(tid 3085067968):
> Memory fault at 0 (pc=0xb7e9a561, sp=0xb78eeb4c)
> The integrity of this image is possibly compromised.
> Continuing with fingers crossed.
>
> debugger invoked on a SB-SYS:MEMORY-FAULT-ERROR in thread #<THREAD
>                                                            "initial thread"
> RUNNING
>                                                            {A9F5AF1}>:
>  Unhandled memory fault at #x0.
>
> Type HELP for debugger help, or (SB-EXT:QUIT) to exit from SBCL.
>
> restarts (invokable by number or by possibly-abbreviated name):
>  0: [ABORT] Exit debugger, returning to top level.
>
> (SB-SYS:MEMORY-FAULT-ERROR)
> 0] 0
>
> --
> Resistance is futile. You will be jazzimilated.
>
> Scientific site:   http://www.lrde.epita.fr/~didier<http://www.lrde.epita.fr/%7Edidier>
> Music (Jazz) site: http://www.didierverna.com
>
>
> ------------------------------------------------------------------------------
> ThinkGeek and WIRED's GeekDad team up for the Ultimate
> GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the
> lucky parental unit.  See the prize list and enter to win:
> http://p.sf.net/sfu/thinkgeek-promo
> _______________________________________________
> Sbcl-help mailing list
> Sbcl-help@...
> https://lists.sourceforge.net/lists/listinfo/sbcl-help
>

Sorry, it is the completely different issue.

2010/6/14 Roman Marynchak <roman.marynchak@...>

> Hello Didier,
>
>  There is the bug report on this:
>
> https://bugs.launchpad.net/sbcl/+bug/338924
>
> I am not sure about the exact match, but the issue seems to be related to
> what you have encountered.
>
>
> Regards,
> Roman
>
>
> 2010/6/14 Didier Verna <didier@...>
>
>
>>  Hi, two questions about getenv:
>>
>> 1/ should I use sb-posix:getenv or sb-ext:posix-getenv ?
>>
>> 2/ also, getenv seems to behave differently on Mac and on Linux
>>   (probably because of differences at a the OS level):
>>
>>   (getenv nil) used to return nil which I kinda liked, but now on Linux
>>   (1.0.39.15, although I can't tell exactly when it started) it throws
>>   an error (see below). Is this intentional ? On my Mac on the other
>>   hand, with 1.0.39.9, it still returns nil.
>>
>>
>> Thanks a lot !
>>
>>
>> * (sb-ext:posix-getenv nil)
>> CORRUPTION WARNING in SBCL pid 13919(tid 3085067968):
>> Memory fault at 0 (pc=0xb7e9a561, sp=0xb78eeb4c)
>> The integrity of this image is possibly compromised.
>> Continuing with fingers crossed.
>>
>> debugger invoked on a SB-SYS:MEMORY-FAULT-ERROR in thread #<THREAD
>>                                                            "initial
>> thread" RUNNING
>>                                                            {A9F5AF1}>:
>>  Unhandled memory fault at #x0.
>>
>> Type HELP for debugger help, or (SB-EXT:QUIT) to exit from SBCL.
>>
>> restarts (invokable by number or by possibly-abbreviated name):
>>  0: [ABORT] Exit debugger, returning to top level.
>>
>> (SB-SYS:MEMORY-FAULT-ERROR)
>> 0] 0
>>
>> --
>> Resistance is futile. You will be jazzimilated.
>>
>> Scientific site:   http://www.lrde.epita.fr/~didier<http://www.lrde.epita.fr/%7Edidier>
>> Music (Jazz) site: http://www.didierverna.com
>>
>>
>> ------------------------------------------------------------------------------
>> ThinkGeek and WIRED's GeekDad team up for the Ultimate
>> GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the
>> lucky parental unit.  See the prize list and enter to win:
>> http://p.sf.net/sfu/thinkgeek-promo
>> _______________________________________________
>> Sbcl-help mailing list
>> Sbcl-help@...
>> https://lists.sourceforge.net/lists/listinfo/sbcl-help
>>
>
>

I wrote:

>    On my Mac on the other hand, with 1.0.39.9, it still returns nil.

     I've upgraded to 1.0.39.15[1] this morning on my new Mac Book Pro
(x86_64 / Snow Leopard), and (getenv nil) still gives nil here.

uname -a:
Darwin scofield.lrde.epita.fr 10.3.1 Darwin Kernel Version 10.3.1: Mon Mar 22 15:13:15 PDT 2010; root:xnu-1504.3.52~1/RELEASE_I386 i386 i386 MacBookPro6,2 Darwin


Footnotes: 
[1]  BTW, compilation still fails on sb-posix's readdir.1 test without
the "encoding" patch.

-- 
Resistance is futile. You will be jazzimilated.

Scientific site:   http://www.lrde.epita.fr/~didier
Music (Jazz) site: http://www.didierverna.com

Hi Didier,

Do you have some non-standard characters in your root directory? This test passes for me.

thanks,

Cyrus

On Jun 14, 2010, at 4:54 AM, Didier Verna wrote:

> I wrote:
> 
>>   On my Mac on the other hand, with 1.0.39.9, it still returns nil.
> 
>     I've upgraded to 1.0.39.15[1] this morning on my new Mac Book Pro
> (x86_64 / Snow Leopard), and (getenv nil) still gives nil here.
> 
> uname -a:
> Darwin scofield.lrde.epita.fr 10.3.1 Darwin Kernel Version 10.3.1: Mon Mar 22 15:13:15 PDT 2010; root:xnu-1504.3.52~1/RELEASE_I386 i386 i386 MacBookPro6,2 Darwin
> 
> 
> Footnotes: 
> [1]  BTW, compilation still fails on sb-posix's readdir.1 test without
> the "encoding" patch.
> 
> -- 
> Resistance is futile. You will be jazzimilated.
> 
> Scientific site:   http://www.lrde.epita.fr/~didier
> Music (Jazz) site: http://www.didierverna.com
> 
> ------------------------------------------------------------------------------
> ThinkGeek and WIRED's GeekDad team up for the Ultimate 
> GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the 
> lucky parental unit.  See the prize list and enter to win: 
> http://p.sf.net/sfu/thinkgeek-promo
> _______________________________________________
> Sbcl-help mailing list
> Sbcl-help@...
> https://lists.sourceforge.net/lists/listinfo/sbcl-help

Cyrus Harmon <ch-sbcl@...> wrote:

> Hi Didier,
>
> Do you have some non-standard characters in your root directory? This
> test passes for me.

  Yes. There's a directory name which contains a small (TM) character.

-- 
Resistance is futile. You will be jazzimilated.

Scientific site:   http://www.lrde.epita.fr/~didier
Music (Jazz) site: http://www.didierverna.com

Yes, indeed! Works perfectly.

Stefan

On 06/08/2010 06:10 PM, karol skocik wrote:
> something like: (sb-thread:interrupt-thread <thread>
> #'sb-debug:backtrace) should do it
> 
> On Tue, Jun 8, 2010 at 5:09 PM, Stefan Neumann
> <stefan.neumann@...> wrote:
>> Hey guys,
>>
>> I am currently facing a not responding webserver (hunchentoot). Each
>> request does not timeout, so I am in the glad situation to inspect our
>> test-system, why the threads are not answering. I can reattach to the
>> server and list the threads[1]. Is there any possibility to have a kind
>> of stacktrace, what the threads are doing?
>>
>> Appreciate your help,
>>
>> Stefan
>>
>> [1]
>> * (sb-thread:list-all-threads)
>>
>> (#<SB-THREAD:THREAD "Hunchentoot worker (client: 127.0.0.1:46639)" RUNNING
>>   {10052FEE41}>
>>  #<SB-THREAD:THREAD "Hunchentoot worker (client: 127.0.0.1:42568)" RUNNING
>>   {10139B7911}>
>>  #<SB-THREAD:THREAD "Hunchentoot worker (client: 127.0.0.1:57279)" RUNNING
>>   {1013CA9E41}>
>>  #<SB-THREAD:THREAD "Hunchentoot worker (client: 127.0.0.1:49259)" RUNNING
>>   {100545B691}>
>>  #<SB-THREAD:THREAD "Hunchentoot worker (client: 127.0.0.1:36498)" RUNNING
>>   {1005376E41}>
>>  #<SB-THREAD:THREAD "Hunchentoot worker (client: 127.0.0.1:54980)" RUNNING
>>   {1011ED9471}>
>>  #<SB-THREAD:THREAD "Hunchentoot worker (client: 127.0.0.1:42381)" RUNNING
>>   {1011E71901}>
>> ....
>>
>> --
>>
>> Stefan Neumann
>> Dipl.-Ing.
>>
>> freiheit.com technologies gmbh
>> Straßenbahnring 22 / 20251 Hamburg, Germany
>> fon       +49 (0)40 / 890584-0
>> fax       +49 (0)40 / 890584-20
>> HRB Hamburg 70814
>>
>> 1CB2 BA3C 168F 0C2B 6005 FC5E 3EBA BCE2 1BF0 21D3
>> Geschäftsführer: Claudia Dietze, Stefan Richter, Jörg Kirchhof
>>
>>
>> ------------------------------------------------------------------------------
>> ThinkGeek and WIRED's GeekDad team up for the Ultimate
>> GeekDad Father's Day Giveaway. ONE MASSIVE PRIZE to the
>> lucky parental unit.  See the prize list and enter to win:
>> http://p.sf.net/sfu/thinkgeek-promo
>> _______________________________________________
>> Sbcl-help mailing list
>> Sbcl-help@...
>> https://lists.sourceforge.net/lists/listinfo/sbcl-help
>>
> 

-- 

Stefan Neumann
Dipl.-Ing.

freiheit.com technologies gmbh
Straßenbahnring 22 / 20251 Hamburg, Germany
fon       +49 (0)40 / 890584-0
fax       +49 (0)40 / 890584-20
HRB Hamburg 70814

1CB2 BA3C 168F 0C2B 6005 FC5E 3EBA BCE2 1BF0 21D3
Geschäftsführer: Claudia Dietze, Stefan Richter, Jörg Kirchhof