don't recollect having seen a patch for this. It would be real
nice if this could be fixed in the upcoming release.
I'm happy to have a stab at this if someone can confirm that the atexit
mechanism is the correct way to go.
Alan Burlison
n Alan Burlison's Kstat module,
> should be included in the final version of Solaris 8.
Solaris 8 has how shipped, and it includes 5.005_03. the kstat(1)
utility is also included.
--
Alan Burlison
Solaris Kernel Development, Sun Microsystems
way people could keep track of bugs without necessarily
being subscribed to p5p. Don't forget that a goodly number of people wo
report bugs aren't on p5p. I don't in fact think that p5p should be on
the interest list at all.
Alan Burlison
wanting to go the
> toke.c route so I didn't mention it earlier.
Umm... it was suggested I started looking in toke.c, so that's what I
did. If you have any further suggestions or insights, please pipe up.
Assume I'm totally clueless (which is close enough to the truth anyway
;-)
Alan Burlison
ughts didn't include 'use namespace' having an effect at
run-time as is now being suggested, but I'm happy to listen to suggested
ways this might work - although personally I'm not convinced it is
needed.
Alan Burlison
think this test should also be restricted to SPARC. Alan?
This test is bobbins. It is confusing support for 64-bit integers (long
long) with support for a 64-bit address space (64 bit pointers). The
first version of Solaris that had a 64-bit address space was Solaris 7
on sparc, but Solaris 2.6 certainly had long long support.
This test I guess should go in the use64bitall segment, and the warning
should refer to 64-bit pointers rather than ints.
--
Alan Burlison
ersonally I'd be content just to see self-ties and tie loops
made illegal. However, at the moment the position WRT the supportedness
of self-ties is unclear.
--
Alan Burlison
y evil. At some level
the need to explicitly diddle refcounts in this manner reflects some
level of brokenness either in your module or in perl. Whilst it might
work for you, I really don't think you should recommend this practice to
others.
Alan Burlison
or the server during the
fork/bind step for both 2.5.1 and 2.8, and I'll have a look.
--
Alan Burlison
Solaris Kernel Development, Sun Microsystems
This is a success report for perl from [EMAIL PROTECTED],
generated with the help of perlbug 1.33 running under perl v5.7.0.
Perl reported to build OK on this system.
---
Flags:
category=install
severity=none
---
Site configuration information for perl v5.7.0:
Configured by alanbur at T
This is a success report for perl from [EMAIL PROTECTED],
generated with the help of perlbug 1.33 running under perl v5.7.0.
Perl reported to build OK on this system.
---
Flags:
category=install
severity=none
---
Site configuration information for perl v5.7.0:
Configured by alanbur at T
This is a success report for perl from [EMAIL PROTECTED],
generated with the help of perlbug 1.33 running under perl v5.7.0.
Perl reported to build OK on this system.
---
Flags:
category=install
severity=none
---
Site configuration information for perl v5.7.0:
Configured by alanbur at T
eads. I'm not sure how much of a restriction this would
turn out to be in practice - how many current perl platforms don't have
threads?
As for AIO - my guess is that faking it up with threads is a much better
bet. After all, what proportion of apps are MT vs AIO, and which is
most likely
This is a success report for perl from [EMAIL PROTECTED],
generated with the help of perlbug 1.33 running under perl v5.7.0.
Perl reported to build OK on this system.
---
Flags:
category=install
severity=none
---
Site configuration information for perl v5.7.0:
Configured by alanbur at S
This is a success report for perl from [EMAIL PROTECTED],
generated with the help of perlbug 1.33 running under perl v5.7.0.
Perl reported to build OK on this system.
---
Flags:
category=install
severity=none
---
Site configuration information for perl v5.7.0:
Configured by alanbur at S
This is a success report for perl from [EMAIL PROTECTED],
generated with the help of perlbug 1.33 running under perl v5.7.0.
Perl reported to build OK on this system.
---
Flags:
category=install
severity=none
---
Site configuration information for perl v5.7.0:
Configured by alanbur at S
This is a success report for perl from [EMAIL PROTECTED],
generated with the help of perlbug 1.33 running under perl v5.7.0.
Perl reported to build OK on this system.
---
Flags:
category=install
severity=none
---
Site configuration information for perl v5.7.0:
Configured by alanbur at S
g that leaks under eval is cause by eval' is fallacious - I've
already found one leak that is only *seen* under an eval, but isn't
*caused* by the eval itself. Again, I don't suppose anybody cares very
much...
Alan Burlison
Alan Burlison wrote:
> Gurusamy Sarathy wrote:
>
> > I'm not convinced this is correct. What if two parties own val
> > (i.e. refcount > 1)?
> >
> > Perhaps we should find out why sv_free() isn't doing the adequate
> > thing?
>
> I&
hopefully fix some of them. GC is
something for perl6.
Alan Burlison
which large lumbering beasts fall :-)
Alan Burlison
S_new_logop[op.c:3656]
Perl_newLOGOP [op.c:3617]
Perl_newCONDOP [op.c:3739]
Perl_yyparse [perly.y:212]
S_parse_body [perl.c:1322]
perl_parse [perl.c:900]
main [perlmain.c:50]
_start [crt1.o]
Alan Burlison
ay (or may not, aren't heisenbugs fun) be a red herring,
> since this
With a FMR it will be tickled differently by any sort of heap activity,
and the assignment to ENV will come into that category. I guess you'll
get all sorts of random strangeness, depending on exactly what has gon
BTW, it then goes on to write into the freed block. From that point
onwards all bets are off - it could dump core anywhere.
Alan Burlison
.
Thanks Sarathay!
Alan Burlison
ntext at - line 8.
GOT:
Useless use of time in void context at - line 4.
Useless use of length in void context at - line 8.
Scalars leaked: 1
PROG:
use warnings ;
time ;
{
use warnings FATAL => qw(void) ;
length "abc" ;
}
join "", 1,2,3 ;
print "done\n" ;
EXPECTED:
Useless use of time in void context at - line 4.
Useless use of length in void context at - line 8.
GOT:
Useless use of time in void context at - line 4.
Useless use of length in void context at - line 8.
Scalars leaked: 1
FAILED at test 145
You need to set PERL_DESTRUCT_LEVEL=5 to see them.
--
Alan Burlison
"H.Merijn Brand" wrote:
>
> Automated smoke report for patch 9171
Can you give me a pointer to the latest version of the smoke scripts?
Thanks,
Alan Burlison
the RE cannot be compiled
at the point of the qr//, but I'm probably missing something obvious.
--
Alan Burlison
--
$ head -1 /dev/bollocks
responsibly target omnipresent embedded e-commerce
I think it does. Perhaps B::Terse is lying.
Well, it gets it right for normal REs, why would it be lying? And, where
in pp_qr does the RE get compiled, or is it done somewhere else? And, if
it isn't done in pp_qr, why not?
--
Alan Burlison
--
$ head -1 /dev/bollocks
maximally grow heterogeneous e-services
ed to 32-bit
programs.
The practical effect is no change and no recompile is required - there is a
stub libthread for those old binaries that require it.
--
Alan Burlison
--
--- Begin Message ---
Automated smoke report for 5.8.1 patch 20558 on solaris - 2.9 (sun4/1 cpu)
(fubar) using /opt/SUNWspro/bin/cc version workshop
Report by Test::Smoke v1.18 (perl 5.6.1) [19 hours]
O = OK F = Failure(s), extended report at the bottom
? = still running or test results not (yet
Automated smoke report for 5.8.1 patch 20617 on solaris - 2.9 (sun4/1 cpu)
(fubar) using /opt/SUNWspro/bin/cc version workshop
Report by Test::Smoke v1.18.02 (perl 5.6.1) [19 hours 19 minutes]
O = OK F = Failure(s), extended report at the bottom
X = test(s) failed under TEST but not under harness
past myself. A question:
doesn't this overlap to some degree with what you can do with overload?
p.s. chocolateboy: who he?
--
Alan Burlison
--
Automated smoke report for 5.8.1 patch 20898 on solaris - 2.9 (sun4/1 cpu)
(fubar) using /opt/SUNWspro/bin/cc version workshop
Report by Test::Smoke v1.18.02 (perl 5.6.1) [18 hours 21 minutes]
O = OK F = Failure(s), extended report at the bottom
X = test(s) failed under TEST but not under harness
Automated smoke report for 5.8.1 patch 21021 on solaris - 2.7 (sun4/1 cpu)
(moe) using /opt/SUNWspro/bin/cc version workshop
Report by Test::Smoke v1.18.02 (perl 5.6.1) [1 day 7 hours 25 minutes]
O = OK F = Failure(s), extended report at the bottom
X = test(s) failed under TEST but not under harn
Automated smoke report for 5.8.1 patch 21021 on solaris - 2.7 (sun4/1 cpu)
(moe) using /opt/SUNWspro/bin/cc version workshop
Report by Test::Smoke v1.18.02 (perl 5.6.1) [1 day 7 hours 24 minutes]
O = OK F = Failure(s), extended report at the bottom
X = test(s) failed under TEST but not under harn
f someone was really keen they could grab the x86 binaries from
http://wwws.sun.com/software/solaris/solaris-express/sol_index.html, as we
have just started shipping free (hopefully monthly) snapshots of the Solaris
10 builds.
--
Alan Burlison
--
e) it isn't integrated into the Solaris build. Other than that, it is fine ;-)
--
Alan Burlison
--
can specify the RUNPATH used by the linker for a particular object file
relative to the location of the file itself. Look at the Makefile.PL
for Sun::Solaris::Exacct for more details.
--
Alan Burlison
--
Test =)
jipdjdhnvbcjqysagl
--
Test, yep.
<>
Test =)
ajfvepm
--
Test, yep.
<>
Test =)
bnxgbnorytrskhsgy
--
Test, yep.
<>
Test =)
ffkmrvh
--
Test, yep.
<>
Test =)
guorganjwmxnqukh
--
Test, yep.
<>
Test =)
bvyrbohftrmijp
--
Test, yep.
<>
e me best :-)
--
Alan Burlison
--
The message contains Unicode characters and has been sent as a binary attachment.
<>
t obviously not...
--
Alan Burlison
--
48 matches
Mail list logo