clisp-devel — closed list for developers and testers

Re: screen -> module

[Bruno H. <br...@cl...>]

Hi Sam,

> > The F() macro had made an assumption that the subr_t at runtime and
> > the subr_t that contains its initialization data are identical...
> 
> I wonder if this has bearing on modules creating new stream types
> https://sourceforge.net/p/clisp/feature-requests/37/
> https://sourceforge.net/p/clisp/mailman/message/14629441/
> and thus a possibility to moving sockets (including secure sockets) to a
> module.

It has no relation. The fixed issue is about the management of the table
of SUBRs. The other issue, which we discussed in 2007, is that the table
of Pseudofuns is not extensible.

Moving the sockets to a separate module is very low priority (no benefit
I can see).

Bruno

Re: DBus module foreign function warnings

[Sam S. <sd...@gn...>]

> * Sam Steingold <fq...@ta...t> [2018-01-26 09:17:03 -0500]:
>
>> * Jerry James <ybtnawreel@tznvy.pbz> [2018-01-25 19:56:43 -0700]:
>>
>> While compiling mercurial head (34f69a3fbf80) on an x86_64 Fedora
>> Rawhide machine with dbus 1.12.0,

Are you actually using this module for something or is it just a
maintenance exercise?

> Let me look into this...

Fixed, thanks.

-- 
Sam Steingold (http://sds.podval.org/) on darwin Ns 10.3.1561
http://steingoldpsychology.com http://www.childpsy.net http://mideasttruth.com
http://camera.org http://www.memritv.org http://memri.org https://ffii.org
A bullet affects the way the brain functions even when it hits the butt.

Re: screen -> module

[Sam S. <sd...@gn...>]

Hi Bruno,

> * Bruno Haible <oe...@py...t> [2018-01-24 19:29:02 +0100]:
>
> The F() macro had made an assumption that the subr_t at runtime and
> the subr_t that contains its initialization data are identical. This
> assumption is no longer generally valid. In fact, the address of a
> SUBR such as REGEXP::REGEXP-FREE is no longer predictable: it depends
> on the number of SUBRs defined by subr.d and of the number of SUBRs
> defined by all prior modules.

I wonder if this has bearing on modules creating new stream types
https://sourceforge.net/p/clisp/feature-requests/37/
https://sourceforge.net/p/clisp/mailman/message/14629441/
and thus a possibility to moving sockets (including secure sockets) to a
module.

Thanks.

-- 
Sam Steingold (http://sds.podval.org/) on darwin Ns 10.3.1561
http://steingoldpsychology.com http://www.childpsy.net http://think-israel.org
https://jihadwatch.org http://honestreporting.com https://ffii.org
There are no answers, only cross references.

Re: DBus module foreign function warnings

[Sam S. <sd...@gn...>]

Hi Jerry,

> * Jerry James <ybtnawreel@tznvy.pbz> [2018-01-25 19:56:43 -0700]:
>
> While compiling mercurial head (34f69a3fbf80) on an x86_64 Fedora
> Rawhide machine with dbus 1.12.0, I get the following warnings when
> building the dbus module:
>
> ;; Loading file /builddir/build/BUILD/clisp-2.49.60/build/dbus/dbus.fas ...
> WARNING: FFI::FIND-FOREIGN-FUNCTION: foreign function
>          "dbus_message_set_serial" does not exist
>          Skip foreign function creation
> WARNING: FFI::FIND-FOREIGN-FUNCTION: foreign function "dbus_message_lock" does
>          not exist
>          Skip foreign function creation
> WARNING: FFI::FIND-FOREIGN-FUNCTION: foreign function
>          "dbus_message_demarshal_bytes_needed" does not exist
>          Skip foreign function creation
> WARNING: FFI::FIND-FOREIGN-FUNCTION: foreign function
>          "dbus_connection_get_adt_audit_session_data" does not exist
>          Skip foreign function creation
> ;; Loaded file /builddir/build/BUILD/clisp-2.49.60/build/dbus/dbus.fas
>
> All 4 functions do exist in libdbus-1.so, according to readelf.  There
> are prototypes for the first 3 in dbus-message.h, and in
> dbus-connection.h for the last one.  Is this something I should worry
> about?  Is there some way of determining why clisp thinks those
> functions don't exist?

Take a look at build/dbus/dbus.c and search for, say,
dbus_connection_get_adt_audit_session_data.
You will see that it is surrounded with
#if defined(HAVE_DBUS_CONNECTION_GET_ADT_AUDIT_SESSION_DATA)
and this is why you get the warning.

Now you need to figure out why configure does not define these constant.
It appears that they are not even mentioned in configure.in.

This is weird.
Let me look into this...

-- 
Sam Steingold (http://sds.podval.org/) on darwin Ns 10.3.1561
http://steingoldpsychology.com http://www.childpsy.net https://jihadwatch.org
http://thereligionofpeace.com http://camera.org http://islamexposedonline.com
The difference between theory and practice is that in theory there isn't any.

DBus module foreign function warnings

[Jerry J. <log...@gm...>]

Greetings,

While compiling mercurial head (34f69a3fbf80) on an x86_64 Fedora
Rawhide machine with dbus 1.12.0, I get the following warnings when
building the dbus module:

;; Loading file /builddir/build/BUILD/clisp-2.49.60/build/dbus/dbus.fas ...
WARNING: FFI::FIND-FOREIGN-FUNCTION: foreign function
         "dbus_message_set_serial" does not exist
         Skip foreign function creation
WARNING: FFI::FIND-FOREIGN-FUNCTION: foreign function "dbus_message_lock" does
         not exist
         Skip foreign function creation
WARNING: FFI::FIND-FOREIGN-FUNCTION: foreign function
         "dbus_message_demarshal_bytes_needed" does not exist
         Skip foreign function creation
WARNING: FFI::FIND-FOREIGN-FUNCTION: foreign function
         "dbus_connection_get_adt_audit_session_data" does not exist
         Skip foreign function creation
;; Loaded file /builddir/build/BUILD/clisp-2.49.60/build/dbus/dbus.fas

All 4 functions do exist in libdbus-1.so, according to readelf.  There
are prototypes for the first 3 in dbus-message.h, and in
dbus-connection.h for the last one.   Is this something I should worry
about?  Is there some way of determining why clisp thinks those
functions don't exist?

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

Re: Pari module build failure

[Jerry J. <log...@gm...>]

On Tue, Jan 23, 2018 at 10:01 PM, Jerry James <log...@gm...> wrote:
> Excellent!  Thanks for the quick fix.  I don't have time tonight, but
> will try to build again tomorrow and see how it goes.  Regards,

Sorry for the delay.  The problem is indeed fixed.  Thanks again.  Regards,
-- 
Jerry James
http://www.jamezone.org/

latest nightly build result

[<don...@is...>]

  ...
  ===> /bin/sh ../src/configure --disable-option-checking --disable-maintainer-mode --cache-file=config.cache SUCCEEDED
  ...
  ===> make SUCCEEDED

  ==> cbcstep3.log <==

  Type :h and hit Enter for context help.

  ;; connecting to "http://clisp.org/beta/impnotes/id-href.map"...connected...HTTP/1.1 301 Moved Permanently --> "https://clisp.sourceforge.io/beta/impnotes/id-href.map"
  *** - OPEN-HTTP("https://clisp.sourceforge.io/beta/impnotes/id-href.map"):
        HTTPS protocol is not supported yet
  Bye.
  Makefile:2247: recipe for target 'check-doc' failed
  make: *** [check-doc] Error 1
  ===> make check FAILED

So this problem is now back.
It looks to me like step 3 is doing more than my "make check" did.
After reporting 
  Test passed.
(where my previous make check reported failure and stopped)
it continues with 
  make[1]: Leaving directory '/home/don/hg/clisp/build-dir/tests'
  mkdir sacla-tests
  cd sacla-tests && ln -s ../../sacla-tests/*.lisp .

So now, how can I skip the check-doc?
Can I do it still with configure --cbcx or do I have to unpack that? 
(In which case I have to add whatever generates all this nice output.)

  ==> cbcstep4.log <==
  ...
  ===> make extracheck SUCCEEDED

  ==> cbcstep5.log <==
  (READ READLINE:*READLINE-INPUT-STREAM*)
  EQL-OK: 5
  RUN-TEST: finished #P"/home/don/hg/clisp/modules/readline/test.tst" (0 errors out of 37 tests)
  finished   4 files:               0 errors out of     632 tests
    1             i18n/test.tst:    0 errors out of      11 tests
    2         syscalls/test.tst:    0 errors out of     264 tests
    3           regexp/test.tst:    0 errors out of     320 tests
    4         readline/test.tst:    0 errors out of      37 tests
  Bye.
  ===> make base-mod-check SUCCEEDED

Excellent - finally succeeded

  ==> cbcstep6.log <==
  NIL
  EQL-OK: T
  *** - READ from
        #<INPUT BUFFERED FILE-STREAM CHARACTER #P"/home/don/hg/clisp/modules/rawsock/test.tst" @24>
        : #<PACKAGE RAWSOCK> has no external symbol with name "CONVERT-ADDRESS"

  Bye.
  Makefile:2407: recipe for target 'full-mod-check' failed
  make: *** [full-mod-check] Error 1
  ===> make full-mod-check FAILED

as before, a problem to be fixed, I guess

  ... cat /home/don/hg/clisp/build-dir/tests/*.erg
  cat: '/home/don/hg/clisp/build-dir/tests/*.erg': No such file or directory

And now I see the problem with unconditionally showing the test failures.
Well, as problems go, this is a good one to have.

Re: nightly build make check etc

[<don...@is...>]

Bruno Haible writes:
 > Don Cohen writes:
 > >  > the only reliable indicator is the `make` exist status.
 > > 
 > > perhaps that result could be printed?
 > 
 > When you run "../configure --cbcx", it creates log files
 > cbcstep1.log ... cbcstep6.log, and each of these files ends in a line
 > such as
 >   ===> make extracheck SUCCEEDED
 > or
 >   ===> make extracheck FAILED
 > 
 > It's easy to grep for.

I see it also ends with a summary:

  Bye.
  Makefile:2407: recipe for target 'full-mod-check' failed
  make: *** [full-mod-check] Error 1
  ===> make full-mod-check FAILED
  Step 1 ===> /bin/sh ../src/configure --disable-option-checking --disable-maintainer-mode --cache-file=config.cache SUCCEEDED
  Step 2 ===> make SUCCEEDED
  Step 3 ===> make check FAILED
  Step 4 ===> make extracheck SUCCEEDED
  Step 5 ===> make base-mod-check FAILED
  Step 6 ===> make full-mod-check FAILED

But tail cbcstep* seems better for showing what's wrong in each case:
  $ tail cbcstep*
  ==> cbcstep1.log <==
  config.status: creating po/Makefile.in
  config.status: creating gllib/Makefile
  config.status: creating makemake
  config.status: creating config.h
  config.status: executing depfiles commands
  config.status: executing libtool commands
  config.status: executing po-directories commands
  config.status: creating po/POTFILES
  config.status: creating po/Makefile
  ===> /bin/sh ../src/configure --disable-option-checking --disable-maintainer-mode --cache-file=config.cache SUCCEEDED

  ==> cbcstep2.log <==
  1815980 ;
  167664 ;
  2 ;
  356064 ;
  31679
  grep "^REPLACE_.*='1'" `find . -name config.log` || true
  ./config.log:REPLACE_IOCTL='1'
  ./config.log:REPLACE_MBRTOWC='1'
  ./config.log:REPLACE_STRERROR_R='1'
  ===> make SUCCEEDED

  ==> cbcstep3.log <==
  Test failed:
  -rw-r--r--. 1 don don 188 Jan 24 14:33 streams.erg
  To see which tests failed, type
      cat /home/don/hg/clisp/build-dir/tests/*.erg

I suppose I could just do that cat unconditionally at the end.


  Makefile:34: recipe for target 'compare' failed
  make[1]: *** [compare] Error 1
  make[1]: Leaving directory '/home/don/hg/clisp/build-dir/tests'
  Makefile:2218: recipe for target 'check-tests' failed
  make: *** [check-tests] Error 2
  ===> make check FAILED

It's interesting that this doesn't show the same error I got before
from make check.  I wonder why not.

  ==> cbcstep4.log <==
  if ./image; then exit 1; fi
  STACK size: 98238 [0x180000bfe00 0x18000000010]
  *** - myerror
  ls -l lisp.run lispinit.mem image image.mem
  -rwxrwxr-x. 1 don don 9890592 Jan 24 14:33 image
  -rw-r--r--. 1 don don 3353376 Jan 24 14:33 image.mem
  -rw-r--r--. 1 don don 3357376 Jan 24 14:32 lispinit.mem
  -rwxrwxr-x. 1 don don 6532472 Jan 24 14:31 lisp.run
  rm -f image image.mem
  ===> make extracheck SUCCEEDED

Ah, that was what I wanted to know.  Thanks.

  ==> cbcstep5.log <==
  RUN-TEST: finished #P"/home/don/hg/clisp/modules/readline/test.tst" (0 errors out of 37 tests)
  finished   4 files:               3 errors out of     632 tests
    1             i18n/test.tst:    0 errors out of      11 tests
    2         syscalls/test.tst:    3 errors out of     264 tests
    3           regexp/test.tst:    0 errors out of     320 tests
    4         readline/test.tst:    0 errors out of      37 tests
  Bye.
  Makefile:2403: recipe for target 'base-mod-check' failed
  make: *** [base-mod-check] Error 1
  ===> make base-mod-check FAILED

So now I'd really like to cat syscalls/test.erg


  ==> cbcstep6.log <==
  NIL
  EQL-OK: T
  *** - READ from
        #<INPUT BUFFERED FILE-STREAM CHARACTER #P"/home/don/hg/clisp/modules/rawsock/test.tst" @24>
        : #<PACKAGE RAWSOCK> has no external symbol with name "CONVERT-ADDRESS"

  Bye.
  Makefile:2407: recipe for target 'full-mod-check' failed
  make: *** [full-mod-check] Error 1
  ===> make full-mod-check FAILED

I guess I've never done the full-mod-check.  Sam only asked for mod-check.

Re: nightly build make check etc

[Bruno H. <br...@cl...>]

Don Cohen writes:
>  > the only reliable indicator is the `make` exist status.
> 
> perhaps that result could be printed?

When you run "../configure --cbcx", it creates log files
cbcstep1.log ... cbcstep6.log, and each of these files ends in a line
such as
  ===> make extracheck SUCCEEDED
or
  ===> make extracheck FAILED

It's easy to grep for.

Bruno

Re: [clisp:clisp] New commit [53a51b] by Bruno Haible

[Bruno H. <br...@cl...>]

Hi Sam,

> > Fix initialization of modules in SINGLEMAP_MEMORY model, part 2.
> 
> it's weird that you need such an elaborate fix for the problem triggered
> by optimizations.

The commit that you found by bisecting enabled various kinds of optimizations;
in particular it disabled NO_ADDRESS_SPACE_ASSUMPTIONS. This made it possible
to use GENERATIONAL_GC, with either TRIVIALMAP_MEMORY or SINGLEMAP_MEMORY.
On Linux/x86_64, the heuristics choose SINGLEMAP_MEMORY.

SINGLEMAP_MEMORY implies that all objects contain the typecode already in
the address, and that this typecode does not need to be masked out.
But this means that all "preallocated" objects, like the symbols from
constsym.d and the SUBRs need to be allocated in mmap()ed memory.
So, the subr_t records that we build up in subr.d and through the modprep-
generated tables are not used at runtime; they only serve to _initialize_
the subr_t records that are using at runtime.

The F() macro had made an assumption that the subr_t at runtime and the
subr_t that contains its initialization data are identical. This assumption
is no longer generally valid. In fact, the address of a SUBR such as
REGEXP::REGEXP-FREE is no longer predictable: it depends on the number
of SUBRs defined by subr.d and of the number of SUBRs defined by all
prior modules.

So spvw.d has to inform the module about the base address of the SUBRs
of the module. This has to be done in init_function_2 and (maybe -
better safe than sorry) also in init_function_1.

And this means that the default implementation of the init_function_2
is no longer empty. Which required changes to all modules that define
an init_function_2 of their own.

Bruno

Re: nightly build make check etc

[<don...@is...>]

Sam Steingold writes:
 > disable check-doc or run it as
how does one disable it?
You mean edit the make file or can you do something like
 make check --exclude check-doc

 > > 2. tail of output of make extracheck:
 > > (Is this output considered "correct"?)
 > probably.
 > the only reliable indicator is the `make` exist status.

perhaps that result could be printed?

How about #3 mod-check ?

Re: nightly build make check etc

[Sam S. <sd...@gn...>]

> * Don Cohen <qba...@vf...3-vap.pbz> [2018-01-24 16:22:15 +0000]:
>
> 3. tail of output of make mod-check:
>    2         syscalls/test.tst:    3 errors out of     264 tests
>
> Form: (LET ((DIR "syscalls-tests-dir/") (DEST "qwer/adsf") COPY) (WHEN (NTH-VALUE 1 (ENSURE-DIRECTORIES-EXIST DIR :VERBOSE T)) (MAPC #'DELETE-FILE (DIRECTORY (CONCATENATE 'STRING DIR "**") :IF-DOES-NOT-EXIST :KEEP))) (SETQ COPY (COPY-FILE DEST DIR :METHOD :SYMLINK)) (LIST (STRING= DEST (CAAR COPY)) (STRING= (PATHNAME-NAME DEST) (PATHNAME-NAME (DELETE-FILE (CADAR COPY)))) (DELETE-DIRECTORY DIR) (PROBE-PATHNAME DIR)))
> CORRECT: (T T T NIL)
> CLISP  : ERROR
> OS-FILE-ERROR(ENOENT): No such file or directory
> OUT:
> "Creating directory: /home/don/hg/clisp/build-dir/syscalls-tests-dir/
> [OS-FILE-ERROR]: OS-FILE-ERROR(ENOENT): No such file or directory
> "

I fixed this.
Fascinating bug.
It turned out that

./clisp -C -E UTF-8 -Emisc 1:1 -q -norc -i tests/tests -x \
 '(run-some-tests :dirlist (list "syscalls/") :srcdir "../modules/" :outdir "./")'

failed while

./clisp -E 1:1 -q -norc -i tests/tests -x \
 '(run-test "../modules/syscalls/test" :logname "syscalls/test")'

worked, because RUN-SOME-TESTS bound *MERGE-PATHNAMES-ANSI* to T.

Thanks.

-- 
Sam Steingold (http://sds.podval.org/) on darwin Ns 10.3.1561
http://steingoldpsychology.com http://www.childpsy.net https://ffii.org
http://islamexposedonline.com http://thereligionofpeace.com
If you need to keep a secret, keep it a secret that you have a secret to keep.

Re: nightly build make check etc

[Sam S. <sd...@gn...>]

> * Don Cohen <qba...@vf...3-vap.pbz> [2018-01-24 16:22:15 +0000]:
>
> 1. tail of output of make check:
> (seems a bad way for make check to end)
>
>  Copyright (c) Sam Steingold, Bruno Haible 2001-2018
>
>  Type :h and hit Enter for context help.
>
>  ;; connecting to "http://clisp.org/beta/impnotes/id-href.map"...connected...HTTP/1.1 301 Moved Permanently --> "https://clisp.sourceforge.io/beta/impnotes/id-href.map"
>  *** - OPEN-HTTP("https://clisp.sourceforge.io/beta/impnotes/id-href.map"):
>        HTTPS protocol is not supported yet
>  Bye.
>  Makefile:2247: recipe for target 'check-doc' failed
>  make: *** [check-doc] Error 1

disable check-doc or run it as

hg clone http://hg.code.sf.net/p/clisp/www /tmp/clisp-www
make IMPNOTES=/tmp/clisp-www/impnotes check-doc

>
> 2. tail of output of make extracheck:
> (Is this output considered "correct"?)
>
>  STACK size: 98238 [0x180000bfe00 0x18000000010]
>  if ./image; then exit 1; fi
>  STACK size: 98238 [0x180000bfe00 0x18000000010]
>  *** - myerror
>  ls -l lisp.run lispinit.mem image image.mem
>  -rwxr-xr-x. 1 don don 9894688 Jan 24 04:09 image
>  -rw-r--r--. 1 don don 3357472 Jan 24 04:09 image.mem
>  -rw-r--r--. 1 don don 3361472 Jan 24 03:09 lispinit.mem
>  -rwxr-xr-x. 1 don don 6532472 Jan 24 03:08 lisp.run
>  rm -f image image.mem

probably.
the only reliable indicator is the `make` exist status.

Thanks for testing!

-- 
Sam Steingold (http://sds.podval.org/) on darwin Ns 10.3.1561
http://steingoldpsychology.com http://www.childpsy.net
http://jij.org http://think-israel.org http://no2bds.org
Microsoft: announce yesterday, code today, think tomorrow.

nightly build make check etc

[<don...@is...>]

1. tail of output of make check:
(seems a bad way for make check to end)

 Copyright (c) Sam Steingold, Bruno Haible 2001-2018

 Type :h and hit Enter for context help.

 ;; connecting to "http://clisp.org/beta/impnotes/id-href.map"...connected...HTTP/1.1 301 Moved Permanently --> "https://clisp.sourceforge.io/beta/impnotes/id-href.map"
 *** - OPEN-HTTP("https://clisp.sourceforge.io/beta/impnotes/id-href.map"):
       HTTPS protocol is not supported yet
 Bye.
 Makefile:2247: recipe for target 'check-doc' failed
 make: *** [check-doc] Error 1


2. tail of output of make extracheck:
(Is this output considered "correct"?)

 STACK size: 98238 [0x180000bfe00 0x18000000010]
 if ./image; then exit 1; fi
 STACK size: 98238 [0x180000bfe00 0x18000000010]
 *** - myerror
 ls -l lisp.run lispinit.mem image image.mem
 -rwxr-xr-x. 1 don don 9894688 Jan 24 04:09 image
 -rw-r--r--. 1 don don 3357472 Jan 24 04:09 image.mem
 -rw-r--r--. 1 don don 3361472 Jan 24 03:09 lispinit.mem
 -rwxr-xr-x. 1 don don 6532472 Jan 24 03:08 lisp.run
 rm -f image image.mem


3. tail of output of make mod-check:
This looks like a big improvement.

 EQL-OK: 5
 RUN-TEST: finished #P"/home/don/hg/clisp/modules/readline/test.tst" (0 errors out of 37 tests)
 finished   4 files:               3 errors out of     632 tests
   1             i18n/test.tst:    0 errors out of      11 tests
   2         syscalls/test.tst:    3 errors out of     264 tests
   3           regexp/test.tst:    0 errors out of     320 tests
   4         readline/test.tst:    0 errors out of      37 tests
 Bye.
 Makefile:2403: recipe for target 'base-mod-check' failed
 make: *** [base-mod-check] Error 1


syscalls/test.erg:

Form: (LET ((DIR "syscalls-tests-dir/") (DEST "qwer/adsf") COPY) (WHEN (NTH-VALUE 1 (ENSURE-DIRECTORIES-EXIST DIR :VERBOSE T)) (MAPC #'DELETE-FILE (DIRECTORY (CONCATENATE 'STRING DIR "**") :IF-DOES-NOT-EXIST :KEEP))) (SETQ COPY (COPY-FILE DEST DIR :METHOD :SYMLINK)) (LIST (STRING= DEST (CAAR COPY)) (STRING= (PATHNAME-NAME DEST) (PATHNAME-NAME (DELETE-FILE (CADAR COPY)))) (DELETE-DIRECTORY DIR) (PROBE-PATHNAME DIR)))
CORRECT: (T T T NIL)
CLISP  : ERROR
OS-FILE-ERROR(ENOENT): No such file or directory
OUT:
"Creating directory: /home/don/hg/clisp/build-dir/syscalls-tests-dir/
[OS-FILE-ERROR]: OS-FILE-ERROR(ENOENT): No such file or directory
"


Form: (LET ((DIR "syscalls-tests-dir/") (LINK #P"syscalls-tests-symlink/")) (LIST (MAKE-DIRECTORY DIR) (STRING= DIR (CAAR (COPY-FILE DIR LINK :METHOD :SYMLINK))) (DELETE-DIRECTORY DIR) (DELETE-DIRECTORY LINK)))
CORRECT: (T T T T)
CLISP  : ERROR
OS-FILE-ERROR(EEXIST): File exists
OUT:
"[OS-FILE-ERROR]: OS-FILE-ERROR(EEXIST): File exists
"


Form: (LET ((DIR "syscalls-tests-dir/") (DEST "foo/bar")) (MAKE-DIRECTORY DIR) (COPY-FILE DEST DIR :METHOD :SYMLINK) (HANDLER-CASE (OR (DELETE-DIRECTORY DIR) (ERROR "deleted non-empty directory")) (OS-ERROR (E) (OR (EQ (OS-ERROR-CODE E) :ENOTEMPTY) (INTEGERP (OS-ERROR-CODE E)) (ERROR "wrong error code: ~s" (OS-ERROR-CODE E))))) (MAPC #'DELETE-FILE (DIRECTORY (CONCATENATE 'STRING DIR "**") :IF-DOES-NOT-EXIST :KEEP)) (DELETE-DIRECTORY DIR))
CORRECT: T
CLISP  : ERROR
OS-FILE-ERROR(EEXIST): File exists
OUT:
"[OS-FILE-ERROR]: OS-FILE-ERROR(EEXIST): File exists
"

Re: [clisp:clisp] New commit [53a51b] by Bruno Haible

[Sam S. <sd...@gn...>]

> * CLISP - an ANSI Common Lisp Mercurial repository <ab...@py...g> [2018-01-23 21:52:45 +0000]:
>
> ## Branch: default  
>
> Fix initialization of modules in SINGLEMAP_MEMORY model, part 2.
>
> By Bruno Haible on 01/23/2018 21:52
> [**View
> Changes**](https://sourceforge.net/p/clisp/clisp/ci/53a51b7cce4b0bad900608e463bb040640c5dedd/)

it's weird that you need such an elaborate fix for the problem triggered
by optimizations.
care to say more?

thanks!

-- 
Sam Steingold (http://sds.podval.org/) on darwin Ns 10.3.1561
http://steingoldpsychology.com http://www.childpsy.net
https://jihadwatch.org http://islamexposedonline.com https://ffii.org
Failure is not an option. It comes bundled with your Microsoft product.

Re: Pari module build failure

[Jerry J. <log...@gm...>]

On Tue, Jan 23, 2018 at 2:55 PM, Bruno Haible <br...@cl...> wrote:
> This should be fixed now. Thanks a lot for the report!

Excellent!  Thanks for the quick fix.  I don't have time tonight, but
will try to build again tomorrow and see how it goes.  Regards,
-- 
Jerry James
http://www.jamezone.org/

Re: supporting older compilers

[Sam S. <sd...@gn...>]

Hi Bruno,

> * Bruno Haible <oe...@py...t> [2018-01-23 22:43:58 +0100]:
>
> These mentions of old compilers in lispbibl.d don't really cost many
> lines of code.

The issue is a clear picture - what we are supporting (and why).

> But it seems safe to remove support for versions < 2.95.2.

Please do!

-- 
Sam Steingold (http://sds.podval.org/) on darwin Ns 10.3.1561
http://steingoldpsychology.com http://www.childpsy.net http://memri.org
http://iris.org.il https://jihadwatch.org http://camera.org
If I had a $ for every time I was rejected, I would no longer be rejected.

Re: Pari module build failure

[Bruno H. <br...@cl...>]

Hi Jerry and Sam,

> > ;; Loading file /builddir/build/BUILD/clisp-2.49.60/build/pari/desc2lisp.fas ...
> > *** - EXT:FINALIZE: #<address #x0000000000859E30> is not a function
> > make[1]: *** [Makefile:35: pari.c] Error 1
> >
> > What does that mean?
> 
> This stems from
> 
> --8<---------------cut here---------------start------------->8---
> > (regexp:regexp-compile "abc")
> *** - FINALIZE: #<ADDRESS #x0000000107B75918> is not a function
> --8<---------------cut here---------------end--------------->8---

This should be fixed now. Thanks a lot for the report!

Bruno

Re: supporting older compilers

[Bruno H. <br...@cl...>]

Hi Sam,

> lispbibl mentions
> "gcc-2.8.1 on hppa and egcs-1.0.3a on i386..."
> "gcc-2.7.2 has a bug..."
> 
> What is the minimal compiler version CLISP supports?

Usually I'd say 2.95.2 is the theoretical minimum. There are still
a couple of machines around with only a gcc 3.3 or 3.4.6.

A few groups (Haiku, OS/2 porters) used to stick with old compilers for a
lnng time.
- On Haiku, gcc 2.95.3 is the default, but people can also use a gcc-5.4 port [1].
- On OS/2, there is a gcc 7.2 meanwhile. [2]

These mentions of old compilers in lispbibl.d don't really cost many
lines of code. But it seems safe to remove support for versions < 2.95.2.

Bruno

[1] https://lists.gnu.org/archive/html/bug-gnulib/2017-08/msg00177.html
[2] http://os2ports.smedley.id.au/index.php?page=copy-of-gcc-v6.x

Re: which build is the fastest?

[Bruno H. <br...@cl...>]

Hi Sam,

> > If you give a man a toy, he will play with it.
> > Now that we have all those multibuild targets in Makefile.devel, we can
> > find out which build is the fastest (e.g., if I make small
> > architecture-agnostic change, which build do I test it in?)

This is something you can easily find out, by putting a 'time' command
in Makefile.devel and then running a 'multibuild-<platform>' target with '-k'.

In my experience, the following factors are important (in decreasing order
of importance):
  1) Put the GCC binaries on a local disk, preferably on a SSD.
     Do not put them on a CIFS mount - this gives terrible compilation times.
  2) Use an older GCC version. GCC 3.x is way faster than 7.x.
  3) Revisit the optimization options: -O is a good compromise, -O2 increases
     the compilation times significantly, and -O2 -fexpensive-optimizations is
     just excessive.

> > Bruno, do you mind if I add "time" to every multibuild target?
> 
> Also, does the configure script have to support platforms without
> /usr/bin/time?

I do mind, yes. And you guessed the reason: We haven't tested portability.
The POSIX 'time' command produced ugly output. 'time' as a bash built-in is OK.

> I noticed that you use sed instead of tail

Yes: Solaris 10 'tail' does not understand the '-n' option.
See also the chapter "Portable Shell Programming" [1].

> so I wonder what the assumptions are.

The assumptions are those listed by the GNU coding standards [2].

Bruno

[1] https://www.gnu.org/savannah-checkouts/gnu/autoconf/manual/autoconf-2.69/html_node/Portable-Shell.html
[2] https://www.gnu.org/prep/standards/html_node/Utilities-in-Makefiles.html

Re: supporting older compilers

[Sam S. <sd...@gn...>]

Hi Bruno,

> * Bruno Haible <oe...@py...t> [2018-01-22 18:44:51 +0100]:
>
> My g++ version is 5.4. And yours?

lispbibl mentions
"gcc-2.8.1 on hppa and egcs-1.0.3a on i386..."
"gcc-2.7.2 has a bug..."

What is the minimal compiler version CLISP supports?

thanks.
-- 
Sam Steingold (http://sds.podval.org/) on darwin Ns 10.3.1561
http://steingoldpsychology.com http://www.childpsy.net
http://www.dhimmitude.org http://mideasttruth.com http://iris.org.il
XML is like violence. If it doesn't solve the problem, use more.

Re: which build is the fastest?

[Sam S. <sd...@gn...>]

> * Sam Steingold <fq...@ta...t> [2018-01-23 09:37:59 -0500]:
>
> If you give a man a toy, he will play with it.
> Now that we have all those multibuild targets in Makefile.devel, we can
> find out which build is the fastest (e.g., if I make small
> architecture-agnostic change, which build do I test it in?)
>
> Bruno, do you mind if I add "time" to every multibuild target?

Also, does the configure script have to support platforms without
/usr/bin/time?

I noticed that you use sed instead of tail, so I wonder what the
assumptions are.

Thanks.

-- 
Sam Steingold (http://sds.podval.org/) on darwin Ns 10.3.1561
http://steingoldpsychology.com http://www.childpsy.net http://no2bds.org
https://jihadwatch.org http://www.dhimmitude.org http://americancensorship.org
What was there first: the Compiler or its Source code?

which build is the fastest?

[Sam S. <sd...@gn...>]

Hi,

If you give a man a toy, he will play with it.
Now that we have all those multibuild targets in Makefile.devel, we can
find out which build is the fastest (e.g., if I make small
architecture-agnostic change, which build do I test it in?)

Bruno, do you mind if I add "time" to every multibuild target?

Thanks.

-- 
Sam Steingold (http://sds.podval.org/) on darwin Ns 10.3.1561
http://steingoldpsychology.com http://www.childpsy.net http://memri.org
http://americancensorship.org http://iris.org.il http://islamexposedonline.com
I am worried - am I paranoid enough?

Re: hg pull, hg update

[Sam S. <sd...@gn...>]

> * Bruno Haible <oe...@py...t> [2018-01-23 08:24:02 +0100]:
>
>> Looks like I managed mess up "hg up".
>
> The distinction between "hg pull && hg update" and "hg pull --update"
> is something that I find very confusing in 'hg'.

I use rebase: in ~/.hgrc:

--8<---------------cut here---------------start------------->8---
[extensions]
hgext.rebase =

[defaults]
pull = --rebase
--8<---------------cut here---------------end--------------->8---



-- 
Sam Steingold (http://sds.podval.org/) on darwin Ns 10.3.1561
http://steingoldpsychology.com http://www.childpsy.net http://camera.org
https://ffii.org http://memri.org http://www.dhimmitude.org
History doesn't repeat itself, but historians do repeat each other.

Re: hg pull, hg update

[Bruno H. <br...@cl...>]

> Looks like I managed mess up "hg up".

The distinction between "hg pull && hg update" and "hg pull --update" [1]
is something that I find very confusing in 'hg'. So, in order to never
use "hg pull" alone, I'm using this script in my $PATH:

================================== hg ==================================
#!/bin/sh
# "hg pull" without --update is always a mistake.
HG=/usr/bin/hg
if test $# = 1 -a "x$1" = xpull; then 
  exec ${HG} pull --update
else
  exec ${HG} "$@"
fi
========================================================================

Hope this helps.

Bruno

[1] https://stackoverflow.com/questions/15241394/mercurial-command-to-combine-hg-pull-hg-up