1.5.8 "make check" on powerpc

1.5.8 "make check" on powerpc

[Wolfgang Jährling]

Hi!

On powerpc-unknown-linux-gnu (a Debian 2.2 installation), a "make check"
after compiling Guile 1.5.8 gives me:

UNRESOLVED: environments.test: leaf-environments: observe weak: weak
observer gets collected
UNRESOLVED: environments.test: leaf-environment based eval-environments:
observe weak: weak observer gets collected

Everything else works.

Cheers,
GNU/Wolfgang

-- 
GMX - Die Kommunikationsplattform im Internet.
http://www.gmx.net



_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/guile-devel

Re: 1.5.8 "make check" on powerpc

[Rob Browning]

Wolfgang Jährling <pro-linux@gmx.de> writes:

> On powerpc-unknown-linux-gnu (a Debian 2.2 installation), a "make check"
> after compiling Guile 1.5.8 gives me:
>
> UNRESOLVED: environments.test: leaf-environments: observe weak: weak
> observer gets collected
> UNRESOLVED: environments.test: leaf-environment based eval-environments:
> observe weak: weak observer gets collected

Did you compile with --enable-debug-malloc or similar?  If so, do you
get the same problem if you try compiling without those options?

I've seen something similar, though not the same, (that I think we
discussed here in the past), and as I recall, it's not certain whether
its actually a bug in regular guile, or a bug in the debugging code --
I don't think I've ever seen it happen without an --enable-debug*
option.

Thanks

-- 
Rob Browning
rlb @defaultvalue.org, @linuxdevel.com, and @debian.org
Previously @cs.utexas.edu
GPG=1C58 8B2C FB5E 3F64 EA5C  64AE 78FE E5FE F0CB A0AD


_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/guile-devel

Re: 1.5.8 "make check" on powerpc

[Wolfgang Jährling]

> > UNRESOLVED: environments.test: leaf-environments: observe weak: weak
> > observer gets collected
> > UNRESOLVED: environments.test: leaf-environment based eval-environments:
> > observe weak: weak observer gets collected
> 
> Did you compile with --enable-debug-malloc or similar?

No, I just did

./configure --prefix=/usr
make
make check

Cheers,
GNU/Wolfgang

-- 
GMX - Die Kommunikationsplattform im Internet.
http://www.gmx.net



_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/guile-devel

Re: 1.5.8 "make check" on powerpc

[Rob Browning]

Wolfgang Jährling <pro-linux@gmx.de> writes:

> On powerpc-unknown-linux-gnu (a Debian 2.2 installation), a "make check"
> after compiling Guile 1.5.8 gives me:
>
> UNRESOLVED: environments.test: leaf-environments: observe weak: weak
> observer gets collected
> UNRESOLVED: environments.test: leaf-environment based eval-environments:
> observe weak: weak observer gets collected

Oh, wait -- memory temporarily vacant.  This is probably the same pair
of issues that are still unresolved on the alpha and ia64.  From NEWS:

  Guile now compiles and passes "make check" with only two UNRESOLVED
  GC cases on Alpha and ia64 based machines now.  Thanks to John
  Goerzen for the use of a test machine, and thanks to Stefan Jahn for
  ia64 patches.

If I recall correctly, the two cases where the same as above.  It's
interesting that it happens on your powerpc too.  Is that a 32 or 64
bit machine?

-- 
Rob Browning
rlb @defaultvalue.org, @linuxdevel.com, and @debian.org
Previously @cs.utexas.edu
GPG=1C58 8B2C FB5E 3F64 EA5C  64AE 78FE E5FE F0CB A0AD


_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/guile-devel

Re: 1.5.8 "make check" on powerpc

[Wolfgang Jährling]

> If I recall correctly, the two cases where the same as above.  It's
> interesting that it happens on your powerpc too.  Is that a 32 or 64
> bit machine?

It is a second-generation iBook, i.e. a 32-bit machine.

Cheers,
GNU/Wolfgang

-- 
GMX - Die Kommunikationsplattform im Internet.
http://www.gmx.net



_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/guile-devel

Re: 1.5.8 "make check" on powerpc

[Rob Browning]

Wolfgang Jährling <pro-linux@gmx.de> writes:

>> If I recall correctly, the two cases where the same as above.  It's
>> interesting that it happens on your powerpc too.  Is that a 32 or 64
>> bit machine?
>
> It is a second-generation iBook, i.e. a 32-bit machine.

Interesting.  I had presumed it was a 32-bit vs 64-bit issue, but
perhaps not.  Anyone know of some other arch-specific (likely C
related) trait those three architectures have in common that they
don't share with x86?

If I get time, I may go run make check on an ia64 and an alpha again
to see if it really is the exact same message.

-- 
Rob Browning
rlb @defaultvalue.org, @linuxdevel.com, and @debian.org
Previously @cs.utexas.edu
GPG=1C58 8B2C FB5E 3F64 EA5C  64AE 78FE E5FE F0CB A0AD


_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/guile-devel

Re: 1.5.8 "make check" on powerpc

[Neil Jerram]

>>>>> "Rob" == Rob Browning <rlb@defaultvalue.org> writes:

    Rob> Wolfgang Jährling <pro-linux@gmx.de> writes:
    >>> If I recall correctly, the two cases where the same as above.  It's
    >>> interesting that it happens on your powerpc too.  Is that a 32 or 64
    >>> bit machine?
    >> 
    >> It is a second-generation iBook, i.e. a 32-bit machine.

    Rob> Interesting.  I had presumed it was a 32-bit vs 64-bit issue, but
    Rob> perhaps not.  Anyone know of some other arch-specific (likely C
    Rob> related) trait those three architectures have in common that they
    Rob> don't share with x86?

Endianness?

        Neil



_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/guile-devel

Re: 1.5.8 "make check" on powerpc

[Rob Browning]

Neil Jerram <neil@ossau.uklinux.net> writes:

> Endianness?

Could be, but I was under the impression that the ia64 is
little-endian.

-- 
Rob Browning
rlb @defaultvalue.org, @linuxdevel.com, and @debian.org
Previously @cs.utexas.edu
GPG=1C58 8B2C FB5E 3F64 EA5C  64AE 78FE E5FE F0CB A0AD


_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/guile-devel

Re: 1.5.8 "make check" on powerpc

[Neil Jerram]

>>>>> "Rob" == Rob Browning <rlb@defaultvalue.org> writes:

    Rob> Neil Jerram <neil@ossau.uklinux.net> writes:
    >> Endianness?

    Rob> Could be, but I was under the impression that the ia64 is
    Rob> little-endian.

Me too, but (i) you didn't seem 100% that it was the same symptom on
all 3 platforms (powerpc, alpha, ia64), and (ii) there could
(conceivably :-) be a bug that manifests on any platform that is
_either_ big-endian _or_ 64-bit.

Regards,
        Neil



_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/guile-devel

Re: 1.5.8 "make check" on powerpc

[Rob Browning]

Neil Jerram <neil@ossau.uklinux.net> writes:

> Me too, but (i) you didn't seem 100% that it was the same symptom on
> all 3 platforms (powerpc, alpha, ia64), and (ii) there could
> (conceivably :-) be a bug that manifests on any platform that is
> _either_ big-endian _or_ 64-bit.

Yeah -- that's a possibility I was thinking about.

-- 
Rob Browning
rlb @defaultvalue.org, @linuxdevel.com, and @debian.org
Previously @cs.utexas.edu
GPG=1C58 8B2C FB5E 3F64 EA5C  64AE 78FE E5FE F0CB A0AD


_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/guile-devel

Re: 1.5.8 "make check" on powerpc

[Dirk Herrmann]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: TEXT/PLAIN; charset=X-UNKNOWN, Size: 1646 bytes --]

Hello,

although I have to admit that I have not looked into those specific test
cases myself I would like to point out the following:  Some UNRESOLVED
issues may be due to guile's garbage collection:  You can not guarantee
that something gets collected, although you expect it to.  The reason is
that guile scans the C execution stack for potential references to
objects.  Some non-SCM values may irritate the gc and keep objects from
being collected.  The two cases below might belong to such a category.
Probably we should clarify this issue in the names of the test cases.

Best regards, 
Dirk Herrmann

--
This message is best viewed with ISO 8859/1 (latin-1) character encoding.
              Microsoft .. what do you want to boot today?

On Thu, 29 Aug 2002, Wolfgang [ISO-8859-1] Jährling wrote:

> Hi!
> 
> On powerpc-unknown-linux-gnu (a Debian 2.2 installation), a "make check"
> after compiling Guile 1.5.8 gives me:
> 
> UNRESOLVED: environments.test: leaf-environments: observe weak: weak
> observer gets collected
> UNRESOLVED: environments.test: leaf-environment based eval-environments:
> observe weak: weak observer gets collected
> 
> Everything else works.
> 
> Cheers,
> GNU/Wolfgang
> 
> -- 
> GMX - Die Kommunikationsplattform im Internet.
> http://www.gmx.net
> 
> 
> 
> _______________________________________________
> Guile-devel mailing list
> Guile-devel@gnu.org
> http://mail.gnu.org/mailman/listinfo/guile-devel
> 



_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/guile-devel

Re: 1.5.8 "make check" on powerpc

[Han-Wen Nienhuys]

dirk@sallust.ida.ing.tu-bs.de writes:
> Hello,
> 
> although I have to admit that I have not looked into those specific test
> cases myself I would like to point out the following:  Some UNRESOLVED
> issues may be due to guile's garbage collection:  You can not guarantee
> that something gets collected, although you expect it to.  The reason is
> that guile scans the C execution stack for potential references to
> objects.  Some non-SCM values may irritate the gc and keep objects from
> being collected.  The two cases below might belong to such a category.
> Probably we should clarify this issue in the names of the test cases.

I believe that the error is that the object is in fact collected, when
it should not have been. At least, that's what I gather from the message.


-- 

Han-Wen Nienhuys   |   hanwen@cs.uu.nl   |   http://www.cs.uu.nl/~hanwen 


_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/guile-devel

Re: 1.5.8 "make check" on powerpc

[Dirk Herrmann]

> I believe that the error is that the object is in fact collected, when
> it should not have been. At least, that's what I gather from the message.

No, the test names are fomulated to describe the successfull case:
  "(+ 1 2) gives 2",
which sounds right when read as:
  PASS: "(+ 1 2) gives 2"
or
  FAIL: "(+ 1 2) gives 2"

Best regards,
Dirk



_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/guile-devel

Re: 1.5.8 "make check" on powerpc

[Dirk Herrmann]

On Fri, 6 Sep 2002, Dirk Herrmann wrote:

> > I believe that the error is that the object is in fact collected, when
> > it should not have been. At least, that's what I gather from the message.
> 
> No, the test names are fomulated to describe the successfull case:
>   "(+ 1 2) gives 2",
> which sounds right when read as:
>   PASS: "(+ 1 2) gives 2"
> or
>   FAIL: "(+ 1 2) gives 2"
> 
> Best regards,
> Dirk

Well, funny.  Who is that guy that wrote this last message?  Could we ban
him from the list for writing up such nonsense ?  ;-)

Please, stop rolling on the floor and simply replace
  "(+ 1 2) gives 2"
by
  "(+ 1 1) gives 2"

Thanks for being mercyful with me.
Dirk



_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://mail.gnu.org/mailman/listinfo/guile-devel