dist/Carp/t/arg_string.t: Test fails in blead #15682

p5pRT · 2016-10-24T13:13:24Z

Migrated from rt.perl.org#129954 (status was 'resolved')

Searchable as RT129954$

p5pRT · 2016-10-24T13:13:24Z

From @dcollinsn

Hello!

Starting somewhere between v5.25.6-35-ge7619cc and v5.25.6-47-g69c5d0e, I
began seeing the following test failures on linux with -Dusequadmath (
http://perl.develop-help.com/raw/?id=196087
http://perl.develop-help.com/raw/?id=196097) and on Windows with all
configurations (http://perl5.test-smoke.org/report/50770)

The Windows issue is the following test:

like lm(-3.5e30), qr/main::lm$-3\.5[eE]\+?30$/;

The test expects numbers to be printed like "-3.5e30" or "-3.5e+30". Both
reasonable choices. Windows and quadruple precision floats are here to make
your life more interesting.

On Windows, I get "030" for the exponent, instead of "30". I can include a
log of this when I get home, I just can't see it from here.

The Linux issue is the same test:

like lm(-3.5e30), qr/main::lm$-3\.5[eE]\+?30$/;

Under Quadmath, we get the following output:

main::lm(-3500000000000000000000000000000)

Not sure if the behavior is wrong or if the test is, but I remember Jarkko
fighting to get the same sort of thing - printf number formats - to test
cleanly while he was working on the hexfp stuff a few months ago.

I'll bisect the Linux issue.

p5pRT · 2016-10-24T13:26:43Z

From @dcollinsn

Here is the bisect:

HEAD is now at 0e25d86 vax-netbsd: do not tempt fp overflow, which will SIGFPE
bad - non-zero exit from sh -c cd t && ./perl TEST \.\.\/dist\/Carp\/t\/arg_string\.t
0e25d86 is the first bad commit
commit 0e25d86
Author: Jarkko Hietaniemi <jhi@iki.fi>
Date: Fri Oct 21 09:05:25 2016 -0400

vax-netbsd: do not tempt fp overflow, which will SIGFPE

The 3e100 seems to have no special meaning, except being in floating point.

:040000 040000 f9f6d9eb328f42f4573a324789155e44f5f23e6e 5585ec7f232fb5471a6de76756fd3588dbcdd489 M dist
bisect run success
That took 424 seconds.

Which modified the above test to change from a 3-digit exponent to a 2-digit exponent, and to significantly shorten it. I'd be happy to provide a patch to add an optional leading zero to the test, but I don't know how to handle the issue of Linux trying to print the value as a 37 digit integer. What's the correct behavior here?

--
Respectfully,
Dan Collins

p5pRT · 2016-10-24T13:26:43Z

From [Unknown Contact. See original ticket]

Here is the bisect:

HEAD is now at 0e25d86 vax-netbsd: do not tempt fp overflow, which will SIGFPE
bad - non-zero exit from sh -c cd t && ./perl TEST \.\.\/dist\/Carp\/t\/arg_string\.t
0e25d86 is the first bad commit
commit 0e25d86
Author: Jarkko Hietaniemi <jhi@iki.fi>
Date: Fri Oct 21 09:05:25 2016 -0400

vax-netbsd: do not tempt fp overflow, which will SIGFPE

The 3e100 seems to have no special meaning, except being in floating point.

:040000 040000 f9f6d9eb328f42f4573a324789155e44f5f23e6e 5585ec7f232fb5471a6de76756fd3588dbcdd489 M dist
bisect run success
That took 424 seconds.

Which modified the above test to change from a 3-digit exponent to a 2-digit exponent, and to significantly shorten it. I'd be happy to provide a patch to add an optional leading zero to the test, but I don't know how to handle the issue of Linux trying to print the value as a 37 digit integer. What's the correct behavior here?

--
Respectfully,
Dan Collins

p5pRT · 2016-10-25T01:01:21Z

From @jkeenan

On Mon Oct 24 06:13:24 2016, dcollinsn@gmail.com wrote:

Hello!

Starting somewhere between v5.25.6-35-ge7619cc and v5.25.6-47-g69c5d0e, I
began seeing the following test failures on linux with -Dusequadmath (
http://perl.develop-help.com/raw/?id=196087
http://perl.develop-help.com/raw/?id=196097) and on Windows with all
configurations (http://perl5.test-smoke.org/report/50770)

The Windows issue is the following test:
like lm$\-3\.5e30$\, qr/main&#8203;::lm\\$\-3\\\.5\[eE\]\\\+?30\\$/;
The test expects numbers to be printed like "-3.5e30" or "-3.5e+30". Both
reasonable choices. Windows and quadruple precision floats are here to make
your life more interesting.

On Windows, I get "030" for the exponent, instead of "30". I can include a
log of this when I get home, I just can't see it from here.

The Linux issue is the same test:
like lm$\-3\.5e30$\, qr/main&#8203;::lm\\$\-3\\\.5\[eE\]\\\+?30\\$/;
Under Quadmath, we get the following output:
main&#8203;::lm$\-3500000000000000000000000000000$
Not sure if the behavior is wrong or if the test is, but I remember Jarkko
fighting to get the same sort of thing - printf number formats - to test
cleanly while he was working on the hexfp stuff a few months ago.

I'll bisect the Linux issue.

Confirmation of test failure:

#####
$ cd t; ./perl harness -v ../dist/Carp/t/arg_string.t; cd -

ok 1
ok 2
ok 3
ok 4

not ok 5
# Failed test at t/arg_string.t line 25.
# 'x at t/arg_string.t line 9.
# main::lm(-3500000000000000000000000000000) called at t/arg_string.t line 25
# '
# doesn't match '(?^:main::lm$-3\.5[eE]\+?30$)'
#####

#####
$ ./perl -Ilib -V
Summary of my perl5 (revision 5 version 25 subversion 7) configuration:
Commit id: a1b60c8
Platform:
osname=linux
osvers=4.4.0-45-generic
archname=x86_64-linux-quadmath
uname='linux zareason 4.4.0-45-generic #66-ubuntu smp wed oct 19 14:12:37 utc 2016 x86_64 x86_64 x86_64 gnulinux '
config_args='-des -Dusedevel -DDEBUGGING -Dusequadmath'
hint=recommended
useposix=true
d_sigaction=define
useithreads=undef
usemultiplicity=undef
use64bitint=define
use64bitall=define
uselongdouble=undef
usemymalloc=n
bincompat5005=undef
Compiler:
cc='cc'
ccflags ='-fwrapv -DDEBUGGING -fno-strict-aliasing -pipe -fstack-protector-strong -I/usr/local/include -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64'
optimize='-O2 -g'
cppflags='-fwrapv -DDEBUGGING -fno-strict-aliasing -pipe -fstack-protector-strong -I/usr/local/include'
ccversion=''
gccversion='5.4.0 20160609'
gccosandvers=''
intsize=4
longsize=8
ptrsize=8
doublesize=8
byteorder=12345678
doublekind=3
d_longlong=define
longlongsize=8
d_longdbl=define
longdblsize=16
longdblkind=3
ivtype='long'
ivsize=8
nvtype='__float128'
nvsize=16
Off_t='off_t'
lseeksize=8
alignbytes=16
prototype=define
Linker and Libraries:
ld='cc'
ldflags =' -fstack-protector-strong -L/usr/local/lib'
libpth=/usr/local/lib /usr/lib/gcc/x86_64-linux-gnu/5/include-fixed /usr/include/x86_64-linux-gnu /usr/lib /lib/x86_64-linux-gnu /lib/../lib /usr/lib/x86_64-linux-gnu /usr/lib/../lib /lib /lib64 /usr/lib64
libs=-lpthread -lnsl -ldb -ldl -lm -lcrypt -lutil -lc -lquadmath
perllibs=-lpthread -lnsl -ldl -lm -lcrypt -lutil -lc -lquadmath
libc=libc-2.23.so
so=so
useshrplib=false
libperl=libperl.a
gnulibc_version='2.23'
Dynamic Linking:
dlsrc=dl_dlopen.xs
dlext=so
d_dlsymun=undef
ccdlflags='-Wl,-E'
cccdlflags='-fPIC'
lddlflags='-shared -O2 -g -L/usr/local/lib -fstack-protector-strong'

Characteristics of this binary (from libperl):
Compile-time options:
DEBUGGING
HAS_TIMES
PERLIO_LAYERS
PERL_COPY_ON_WRITE
PERL_DONT_CREATE_GVSV
PERL_HASH_FUNC_ONE_AT_A_TIME_HARD
PERL_MALLOC_WRAP
PERL_OP_PARENT
PERL_PRESERVE_IVUV
PERL_USE_DEVEL
USE_64_BIT_ALL
USE_64_BIT_INT
USE_LARGE_FILES
USE_LOCALE
USE_LOCALE_COLLATE
USE_LOCALE_CTYPE
USE_LOCALE_NUMERIC
USE_LOCALE_TIME
USE_PERLIO
USE_PERL_ATOF
USE_QUADMATH
Built under linux
#####
--
James E Keenan (jkeenan@cpan.org)

p5pRT · 2016-10-25T01:01:21Z

The RT System itself - Status changed from 'new' to 'open'

p5pRT · 2016-10-27T08:19:21Z

From @iabyn

On Mon, Oct 24, 2016 at 06:26:43AM -0700, Dan Collins via RT wrote:

Here is the bisect:

HEAD is now at 0e25d86 vax-netbsd: do not tempt fp overflow, which will SIGFPE
bad - non-zero exit from sh -c cd t && ./perl TEST \.\.\/dist\/Carp\/t\/arg_string\.t
0e25d86 is the first bad commit
commit 0e25d86
Author: Jarkko Hietaniemi <jhi@iki.fi>
Date: Fri Oct 21 09:05:25 2016 -0400
vax\-netbsd&#8203;: do not tempt fp overflow\, which will SIGFPE

The 3e100 seems to have no special meaning\, except being in floating point\.
:040000 040000 f9f6d9eb328f42f4573a324789155e44f5f23e6e 5585ec7f232fb5471a6de76756fd3588dbcdd489 M dist
bisect run success
That took 424 seconds.

Which modified the above test to change from a 3-digit exponent to a 2-digit exponent, and to significantly shorten it. I'd be happy to provide a patch to add an optional leading zero to the test, but I don't know how to handle the issue of Linux trying to print the value as a 37 digit integer. What's the correct behavior here?

I think the test was being overly picky.

I've just pushed the following:

commit 13596ab
Author: David Mitchell <davem@iabyn.com>
AuthorDate: Thu Oct 27 09:09:24 2016 +0100
Commit: David Mitchell <davem@iabyn.com>
CommitDate: Thu Oct 27 09:09:24 2016 +0100

Carp/t/arg_string.t: be liberal in f/p formats

[perl #129954] dist/Carp/t/arg_string.t: Test fails

This test script checks that args are displayed sensibly in longmess()
output, but floating-point numbers can be displayed in various formats
depending on platform, so make the regex more forgiving.

Also add a comment to the top of the script explaining its purpose.

Affected files ...
M dist/Carp/t/arg_string.t

Differences ...

Inline Patch

diff --git a/dist/Carp/t/arg_string.t b/dist/Carp/t/arg_string.t
index c94ec48..4c754c5 100644
--- a/dist/Carp/t/arg_string.t
+++ b/dist/Carp/t/arg_string.t
@@ -1,6 +1,8 @@
 use warnings;
 use strict;
 
+# confirm that stack args are displayed correctly by longmess()
+
 use Test::More tests => 32;
 
 use Carp ();
@@ -22,7 +24,11 @@ like lm(3), qr/main::lm\(3\)/;
 like lm(substr("3\x{2603}", 0, 1)), qr/main::lm\(3\)/;
 like lm(-3), qr/main::lm\(-3\)/;
 like lm(-3.5), qr/main::lm\(-3\.5\)/;
-like lm(-3.5e30), qr/main::lm\(-3\.5[eE]\+?30\)/;
+like lm(-3.5e30),
+            qr/main::lm\(
+                -3500000000000000000000000000000
+              | -3\.5[eE]\+?0?30\)
+            /x;
 like lm(""), qr/main::lm\(""\)/;
 like lm("foo"), qr/main::lm\("foo"\)/;
 like lm("a\$b\@c\\d\"e"), qr/main::lm\("a\\\$b\\\@c\\\\d\\\"e"\)/;


--

I thought I was wrong once, but I was mistaken.

p5pRT · 2016-10-27T08:27:56Z

@iabyn - Status changed from 'open' to 'resolved'

p5pRT · 2016-10-27T10:21:42Z

From @iabyn

On Thu, Oct 27, 2016 at 09:19:02AM +0100, Dave Mitchell wrote:

I think the test was being overly picky.

I've just pushed the following:

And that was being underly picky, so I just pushed the following too:

commit fd609c8
Author: David Mitchell <davem@iabyn.com>
AuthorDate: Thu Oct 27 09:29:33 2016 +0100
Commit: David Mitchell <davem@iabyn.com>
CommitDate: Thu Oct 27 09:32:48 2016 +0100

Carp/t/arg_string.t: fixup to previous fixup

I added alternates to a regex for matching a f/p number, but forgot
to put parentheses around them. So it was being ridiculously over-liberal

Affected files ...
M dist/Carp/t/arg_string.t

Differences ...

Inline Patch

diff --git a/dist/Carp/t/arg_string.t b/dist/Carp/t/arg_string.t
index 4c754c5..dc70f43 100644
--- a/dist/Carp/t/arg_string.t
+++ b/dist/Carp/t/arg_string.t
@@ -26,9 +26,11 @@ like lm(-3), qr/main::lm\(-3\)/;
 like lm(-3.5), qr/main::lm\(-3\.5\)/;
 like lm(-3.5e30),
             qr/main::lm\(
+              (
                 -3500000000000000000000000000000
-              | -3\.5[eE]\+?0?30\)
-            /x;
+              | -3\.5[eE]\+?0?30
+              )
+              \) /x;
 like lm(""), qr/main::lm\(""\)/;
 like lm("foo"), qr/main::lm\("foo"\)/;
 like lm("a\$b\@c\\d\"e"), qr/main::lm\("a\\\$b\\\@c\\\\d\\\"e"\)/;


--

All wight. I will give you one more chance. This time, I want to hear
no Wubens. No Weginalds. No Wudolf the wed-nosed weindeers.
-- Life of Brian

p5pRT closed this as completed Oct 27, 2016

p5pRT added the Severity Low label Oct 19, 2019

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

dist/Carp/t/arg_string.t: Test fails in blead #15682

dist/Carp/t/arg_string.t: Test fails in blead #15682

p5pRT commented Oct 24, 2016

p5pRT commented Oct 24, 2016

p5pRT commented Oct 24, 2016

p5pRT commented Oct 24, 2016

p5pRT commented Oct 25, 2016

p5pRT commented Oct 25, 2016

p5pRT commented Oct 27, 2016

p5pRT commented Oct 27, 2016

p5pRT commented Oct 27, 2016

dist/Carp/t/arg_string.t: Test fails in blead #15682

dist/Carp/t/arg_string.t: Test fails in blead #15682

Comments

p5pRT commented Oct 24, 2016

p5pRT commented Oct 24, 2016

From @dcollinsn

p5pRT commented Oct 24, 2016

From @dcollinsn

p5pRT commented Oct 24, 2016

From [Unknown Contact. See original ticket]

p5pRT commented Oct 25, 2016

From @jkeenan

p5pRT commented Oct 25, 2016

p5pRT commented Oct 27, 2016

From @iabyn

p5pRT commented Oct 27, 2016

p5pRT commented Oct 27, 2016

From @iabyn