commit 15899733e changes SKIP to FAIL on Win32 in op/sprintf.t

[p5pRT]

Migrated from rt.perl.org#128630 (status was 'resolved')

Searchable as RT128630$

[p5pRT]

From @bulk88

Created by @bulk88

Blead perl on WinXP 32 bits with VC 2003. Git reverting commit "VAX​:
test changes for VAX floats" causes op/sprintf.t to pass.

------------------------------------------------------
ok 228 - >12345.7<
ok 229 - >+12345.7<
ok 230 - >12345.7<
ok 231 - >0 MISSING<
ok 232 - > 0 MISSING<
ok 233 - >0 MISSING<
ok 234 - >0C MISSING<
not ok 235 - >%.0g< >-0.0< >-0< >0< C99 standard mandates minus sign but
C89 does not
ok 236 - >1e+04< >1e+004< three-digit exponent accepted
ok 237 - >1.e+04< >1.e+004< three-digit exponent accepted
ok 238 - >1.2e+04< >1.2e+004< three-digit exponent accepted
ok 239 - >1.2e+04< >1.2e+004< three-digit exponent accepted
ok 240 - >12345.6789<
ok 241 - > 12345.6789<
ok 242 - >0012345.6789<
------------------------------------------------------
ok 548 - >030​:031​:032< perl \#83194​: vector flag + custom separator +
dynamic precision
ok 549 - >118.101.114.115.105.111.110< perl \#102586​: vector flag +
"version"
ok 550 - > 011/ 022/ 033< four reordered args
Failed 1/550 subtests

Test Summary Report
-------------------
op/sprintf.t (Wstat​: 0 Tests​: 550 Failed​: 1)
  Failed test​: 235
Files=1, Tests=550, 0 wallclock secs ( 0.11 usr + 0.00 sys = 0.11 CPU)
Result​: FAIL
------------------------------------------------------

Perl Info

Flags:
      category=library
      severity=medium
      module=locale

Site configuration information for perl 5.25.3:

Configured by Owner at Wed Jul 13 21:11:05 2016.

Summary of my perl5 (revision 5 version 25 subversion 3) configuration:

    Platform:
      osname=MSWin32
      osvers=5.1
      archname=MSWin32-x86-multi-thread
      uname=''
      config_args='undef'
      hint=recommended
      useposix=true
      d_sigaction=undef
      useithreads=define
      usemultiplicity=define
      use64bitint=undef
      use64bitall=undef
      uselongdouble=undef
      usemymalloc=n
      bincompat5005=undef
    Compiler:
      cc='cl'
      ccflags ='-nologo -GF -W3 -Od -MD -Zi -DDEBUGGING -DWIN32 -D_CONSOLE
-DNO_STRICT  -DPERL_TEXTMODE_SCRIPTS -DPERL_IMPLICIT_CONTEXT
-DPERL_IMPLICIT_SYS -DWIN32_NO_REGISTRY -DUSE_PERLIO'
      optimize='-Od -MD -Zi -DDEBUGGING'
      cppflags='-DWIN32'
      ccversion='13.10.6030'
      gccversion=''
      gccosandvers=''
      intsize=4
      longsize=4
      ptrsize=4
      doublesize=8
      byteorder=1234
      doublekind=3
      d_longlong=undef
      longlongsize=8
      d_longdbl=define
      longdblsize=8
      longdblkind=0
      ivtype='long'
      ivsize=4
      nvtype='double'
      nvsize=8
      Off_t='__int64'
      lseeksize=8
      alignbytes=8
      prototype=define
    Linker and Libraries:
      ld='link'
      ldflags ='-nologo -nodefaultlib -debug -libpath:"c:\perl\lib\CORE"
-machine:x86'
      libpth=\lib
      libs=oldnames.lib kernel32.lib user32.lib gdi32.lib winspool.lib
comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib
netapi32.lib uuid.lib ws2_32.lib mpr.lib winmm.lib version.lib
odbc32.lib odbccp32.lib comctl32.lib msvcrt.lib
      perllibs=oldnames.lib kernel32.lib user32.lib gdi32.lib winspool.lib
comdlg32.lib advapi32.lib shell32.lib ole32.lib oleaut32.lib
netapi32.lib uuid.lib ws2_32.lib mpr.lib winmm.lib version.lib
odbc32.lib odbccp32.lib comctl32.lib msvcrt.lib
      libc=msvcrt.lib
      so=dll
      useshrplib=true
      libperl=perl525.lib
      gnulibc_version=''
    Dynamic Linking:
      dlsrc=dl_win32.xs
      dlext=dll
      d_dlsymun=undef
      ccdlflags=' '
      cccdlflags=' '
      lddlflags='-dll -nologo -nodefaultlib -debug
-libpath:"c:\perl\lib\CORE" -machine:x86'



@INC for perl 5.25.3:
      lib
      C:/perl521/src/lib
      .


Environment for perl 5.25.3:
      HOME (unset)
      LANG (unset)
      LANGUAGE (unset)
      LD_LIBRARY_PATH (unset)
      LOGDIR (unset)
      PATH=C:\sperl\c\bin;C:\WINDOWS\system32;C:\Program Files\Microsoft
Visual Studio .NET 2003\Vc7\bin;C:\Program Files\Microsoft Visual Studio
.NET 2003\Common7\IDE;C:\WINDOWS;C:\Program Files\Git\cmd;C:\Program
Files\Microsoft Visual Studio .NET 2003\Common7\Tools\bin;C:\perl\bin
      PERL_BADLANG (unset)
      PERL_JSON_BACKEND=Cpanel::JSON::XS
      PERL_YAML_BACKEND=YAML
      SHELL (unset)

[p5pRT]

From @dcollinsn

This refers to commit 1589973, specifically to this test​:

->%.0g< >-0.0< >-0< >C99 standard mandates minus sign but C89 does not skip​: MSWin32 VMS hpux​:10.20 openbsd netbsd​:1.5 irix darwin freebsd​:4.9 android<
+>%.0g< >-0.0< >-0< >C99 standard mandates minus sign but C89 does not skip​: MSWin32 VMS netbsd​:vax-netbsd hpux​:10.20 openbsd netbsd​:1.5 irix darwin freebsd​:4.9 android<

But really this change​:

- } elsif ($os =~ /\b$^O(?​::(\S+))?\b/i) {
+ } elsif ($os =~ /\b$^O(?​::(\S+))\b/i) {

Which makes the skip-by-os only work with a :version or :archname. The patch which I will attach momentarily restores the check for plain "OS".

There's another issue as well, though. The problem commit tried to add support for skip-by-archname, and now skips certain tests for 'netbsd​:vax-netbsd' - but what about 'netbsd​:vax-netbsd-thread-multi'? I'll attach a second patch that fixes that.

--
Respectfully,
Dan Collins

[p5pRT]

From @dcollinsn

0001-RT-128630-t-op-sprintf-Fix-skip-by-os.patch

From ee000efb35d066dfc1b0dba0798f6151d3768733 Mon Sep 17 00:00:00 2001
From: Dan Collins <dcollinsn@gmail.com>
Date: Fri, 15 Jul 2016 20:29:20 -0400
Subject: [PATCH 1/2] [RT #128630] t/op/sprintf: Fix skip by os

15899733e modified the search for skip by OS. Previously it would
accept 'OS:ver' or 'OS', where 'ver' was numeric. That commit
added, for example, 'netbsd:vax-netbsd', but removed support for
plain 'OS'. This restores that.
---
 t/op/sprintf.t | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/t/op/sprintf.t b/t/op/sprintf.t
index 4aef466..8cd32ea 100644
--- a/t/op/sprintf.t
+++ b/t/op/sprintf.t
@@ -115,6 +115,8 @@ for (@tests) {
 	# >comment skip: all<
 	if ($os =~ /\ball\b/i) {
 	    $skip = 1;
+	} elsif ($os =~ /\b$^O\b/i) {
+	    $skip = 1;
 	} elsif ($os =~ /\b$^O(?::(\S+))\b/i) {
             my $cond = $1;
             if ($cond =~ m{^/(.+)/$}) {
-- 
2.8.1

[p5pRT]

From @dcollinsn

0002-t-op-sprintf.t-Further-fix-os-archname-skip.patch

From 3e12e4fe409a784a2b3534d231430d4d8f4efb9c Mon Sep 17 00:00:00 2001
From: Dan Collins <dcollinsn@gmail.com>
Date: Fri, 15 Jul 2016 20:42:10 -0400
Subject: [PATCH 2/2] t/op/sprintf.t: Further fix os:archname skip

The tests in t/op/sprintf.t can be skipped by os, os:ver, or
os:archname. However, if it is desired to skip a test on
netbsd:vax-netbsd, then that should also be skipped on the
other perl-specific "flavors" of that archname, such as
netbsd:vax-netbsd-thread-multi, or -64all, or -ld. This commit
checks if the skip command is the prefix of the archname in
config.sh rather than comparing strict equality.
---
 t/op/sprintf.t | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/t/op/sprintf.t b/t/op/sprintf.t
index 8cd32ea..4c77f6b 100644
--- a/t/op/sprintf.t
+++ b/t/op/sprintf.t
@@ -132,7 +132,7 @@ for (@tests) {
                 $skip = $vsn ? ($osv <= $vsn ? 1 : 0) : 1;
             } else {
                 # >comment skip: netbsd:vax-netbsd<
-                $skip = $cond eq $archname;
+                $skip = $archname =~ /^$cond/;
             }
 	}
 	$skip and $comment =~ s/$/, failure expected on $^O $osv $archname/;
-- 
2.8.1

[p5pRT]

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

[p5pRT]

From @tonycoz

On Fri Jul 15 15​:59​:53 2016, bulk88 wrote​:

Blead perl on WinXP 32 bits with VC 2003. Git reverting commit "VAX​:
test changes for VAX floats" causes op/sprintf.t to pass.

This appears to have been fixed by 6151d43.

Tony

[p5pRT]

From @tonycoz

On Fri Jul 15 17​:49​:18 2016, dcollinsn@​gmail.com wrote​:

There's another issue as well, though. The problem commit tried to add
support for skip-by-archname, and now skips certain tests for
'netbsd​:vax-netbsd' - but what about 'netbsd​:vax-netbsd-thread-multi'?
I'll attach a second patch that fixes that.

Should the comment above that, and before the __END__ also be updated?

Tony

[p5pRT]

From @bulk88

On Sun Jul 17 17​:52​:43 2016, tonyc wrote​:

On Fri Jul 15 15​:59​:53 2016, bulk88 wrote​:

Blead perl on WinXP 32 bits with VC 2003. Git reverting commit "VAX​:
test changes for VAX floats" causes op/sprintf.t to pass.

This appears to have been fixed by 6151d43.

Tony

It wasn't fixed by 6151d43

On, SHA-1​: 9648eab

* Add epigraphs for 5.22.3-RC2 and 5.24.1-RC2

I get

---

C​:\perl521\src\win32>cd ..\t & perl harness -v op/sprintf.t & cd ..\win32
# Failed test 235 - >%.0g< >-0.0< >-0< >0< C99 standard mandates minus sign but C89 does not at op/sprintf.t line 168
op/sprintf.t ..
1..550
ok 1 - >%6. 6s INVALID REDUNDANT< (See use of $w in code above)
ok 2 - >%6 .6s INVALID REDUNDANT<
ok 3 - >%6.6 s INVALID REDUNDANT<
ok 4 - skip %A tested in sprintf2.t, failure expected on MSWin32 5.1 MSWin32-x86-multi-thread
ok 5 - >11111111111111111111111111111111<
ok 6 - >11111111111111111111111111111111<
ok 7 - >0B11111111111111111111111111111111<
ok 8 - >%C INVALID REDUNDANT<
ok 9 - >2147483647< Synonym for %ld
ok 10 - >1.234568E+05< >1.234568E+005< three-digit exponent accepted
ok 11 - >123456.789000< Synonym for %f
ok 12 - >1.23457E+06< >1.23457E+006< three-digit exponent accepted
ok 13 - >1.23457E+102<
ok 14 - >1.23457E-102<
ok 15 - >12345.7<
ok 16 - >1.23457E+102< exponent too big
ok 17 - >1.23457E-102< exponent too small
ok 18 - >%H INVALID REDUNDANT<
ok 19 - >%I INVALID REDUNDANT<
ok 20 - >%J INVALID REDUNDANT<
ok 21 - >%K INVALID REDUNDANT<
ok 22 - >%L INVALID REDUNDANT<
ok 23 - >%M INVALID REDUNDANT<
ok 24 - >%N INVALID REDUNDANT<
ok 25 - >37777777777< Synonym for %lo
ok 26 - >%P INVALID REDUNDANT<
ok 27 - >%Q INVALID REDUNDANT<
****cut*****
ok 229 - >+12345.7<
ok 230 - >12345.7<
ok 231 - >0 MISSING<
ok 232 - > 0 MISSING<
ok 233 - >0 MISSING<
ok 234 - >0C MISSING<
not ok 235 - >%.0g< >-0.0< >-0< >0< C99 standard mandates minus sign but C89 does not
ok 236 - >1e+04< >1e+004< three-digit exponent accepted
ok 237 - >1.e+04< >1.e+004< three-digit exponent accepted
ok 238 - >1.2e+04< >1.2e+004< three-digit exponent accepted
ok 239 - >1.2e+04< >1.2e+004< three-digit exponent accepted
ok 240 - >12345.6789<
ok 241 - > 12345.6789<
ok 242 - >0012345.6789<
ok 243 - >12345.6789 <
ok 244 - >12345.6789 <
ok 245 - >12345.6789 <
ok 246 - >-12345.7<
ok 247 - >-12345.7<
ok 248 - >1.23457e+06< >1.23457e+006< three-digit exponent accepted
ok 249 - >+1.23457e+06< >+1.23457e+006< three-digit exponent accepted
ok 250 - >1.23457e+06< >1.23457e+006< three-digit exponent accepted
ok 251 - >-1.23457e+06< >-1.23457e+006< three-digit exponent accepted
ok 252 - >-1.23457e+06< >-1.23457e+006< three-digit exponent accepted
ok 253 - >-1.23457e+06< >-1.23457e+006< three-digit exponent accepted
ok 254 - >0.00012345<
ok 255 - >1.2345e-05< >1.2345e-005< three-digit exponent accepted
ok 256 - >1.23457e+102<
ok 257 - >1.23457e-102<
ok 258 - >0<
ok 259 - > 1.23457e+06< > 1.23457e+006< three-digit exponent accepted
ok 260 - > +1.23457e+06< >+1.23457e+006< three-digit exponent accepted
******cut******
ok 544 - >30​:31​:32< perl \#83194​: vector flag + custom separator
ok 545 - >030.031.032< perl \#83194​: vector flag + static precision
ok 546 - >030.031.032< perl \#83194​: vector flag + dynamic precision
ok 547 - >030​:031​:032< perl \#83194​: vector flag + custom separator + static precision
ok 548 - >030​:031​:032< perl \#83194​: vector flag + custom separator + dynamic precision
ok 549 - >118.101.114.115.105.111.110< perl \#102586​: vector flag + "version"
ok 550 - > 011/ 022/ 033< four reordered args
Failed 1/550 subtests

Test Summary Report

---

op/sprintf.t (Wstat​: 0 Tests​: 550 Failed​: 1)
  Failed test​: 235
Files=1, Tests=550, 1 wallclock secs ( 0.08 usr + 0.00 sys = 0.08 CPU)
Result​: FAIL

---

--
bulk88 ~ bulk88 at hotmail.com

[p5pRT]

From @dcollinsn

On Mon Jul 25 16​:50​:22 2016, bulk88 wrote​:

It wasn't fixed by 6151d43

I agree with bulk88. These lines handle the skip logic for plain win32​:

140 # >comment skip​: netbsd​:vax-netbsd<
141 $skip = $cond eq $archname;
142 }

That test has a skip condition "MSWin32", but my windows perls have archnames like "archname=MSWin32-x64-multi-thread". I suggest that any fix regex match `$archname =~ /\b$cond\b/` or `$archname =~ /^$cond/` rather than relying on string equality.

Copying/assigning to Jarkko, as he committed 6151d43

--
Respectfully,
Dan Collins

[p5pRT]

From @jhi

On Mon Jul 25 17​:07​:03 2016, dcollinsn@​gmail.com wrote​:

On Mon Jul 25 16​:50​:22 2016, bulk88 wrote​:

It wasn't fixed by 6151d43

I agree with bulk88. These lines handle the skip logic for plain
win32​:

140 # >comment skip​: netbsd​:vax-netbsd<
141 $skip = $cond eq $archname;
142 }

That test has a skip condition "MSWin32", but my windows perls have
archnames like "archname=MSWin32-x64-multi-thread". I suggest that any
fix regex match `$archname =~ /\b$cond\b/` or `$archname =~ /^$cond/`
rather than relying on string equality.

Copying/assigning to Jarkko, as he committed 6151d43

Before tweaking any further I would like to understand how the mswin32 case matched before the 1589973.

[p5pRT]

From @dcollinsn

Before, it used $^O instead of $archname. Come to think of it, that is also
probably fine - though I don't know if there are any unusual ways to build
perl on Win32 that might end up with a strange $^O...

W​:\buildbot\windows1\win64-mingww64-4_9_3\build>.\perl -Ilib -MConfig -wle
"print $^O; print $Config{archname}"
MSWin32
MSWin32-x64-multi-thread

On Mon, Jul 25, 2016 at 8​:54 PM, Jarkko Hietaniemi via RT <
perlbug-followup@​perl.org> wrote​:

On Mon Jul 25 17​:07​:03 2016, dcollinsn@​gmail.com wrote​:

On Mon Jul 25 16​:50​:22 2016, bulk88 wrote​:

It wasn't fixed by 6151d43

I agree with bulk88. These lines handle the skip logic for plain
win32​:

140 # >comment skip​: netbsd​:vax-netbsd<
141 $skip = $cond eq $archname;
142 }

That test has a skip condition "MSWin32", but my windows perls have
archnames like "archname=MSWin32-x64-multi-thread". I suggest that any
fix regex match `$archname =~ /\b$cond\b/` or `$archname =~ /^$cond/`
rather than relying on string equality.

Copying/assigning to Jarkko, as he committed 6151d43

Before tweaking any further I would like to understand how the mswin32
case matched before the 1589973.

---
via perlbug​: queue​: perl5 status​: open
https://rt-archive.perl.org/perl5/Ticket/Display.html?id=128630

[p5pRT]

From @cpansprout

Does dbdc20d fix this?

--

Father Chrysostomos

[p5pRT]

From @bulk88

On Fri Jul 29 00​:27​:33 2016, sprout wrote​:

Does dbdc20d fix this?

Probably that commit fixed it. Blead at Revision​: c7202de
Author​: Karl Williamson <khw@​cpan.org>
Date​: 7/31/2016 5​:34​:47 AM
Message​: locale.c​: Add some DEBUG statements

passes in sprintf.t for me now so I think this ticket can be closed.

--
bulk88 ~ bulk88 at hotmail.com

[p5pRT]

From @khwilliamson

Closed by OP request
--
Karl Williamson

[p5pRT]

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