Warning if -F given but not -a #12669
Comments
From @epaCreated by @epa% perl -F: -nE 'say $F[0]' Here the user has forgotten the -a flag to split the string into @F. Perl Info |
From @jkeenanOn Mon Dec 24 08:19:49 2012, eda@waniasset.com wrote:
Would it be possible for you to compose a test file for this -- Thank you very much. |
|
The RT System itself - Status changed from 'new' to 'open' |
From @epaHere is a test file for the behaviour suggested: #!perl -- ______________________________________________________________________ |
From @jkeenanOn Wed Jan 02 09:15:47 2013, eda@waniasset.com wrote:
Thanks for the patch. I had to rework it into a form suitable for P5P: please review the patch attached. Thank you very much. |
From @jkeenanFrom e701c5fa5e168ec52e39b9d38d4f4bfc8a5ecb58 Mon Sep 17 00:00:00 2001 Responding to bug report by Ed Avis++. Enforce requirement For RT #116190. MANIFEST | 1 + Inline Patchdiff --git a/MANIFEST b/MANIFEST
index ffb79ac..c5d22db 100644
--- a/MANIFEST
+++ b/MANIFEST
@@ -5522,6 +5522,7 @@ t/run/switchd-78586.t See whether bug 78586 is fixed
t/run/switchd.t Test the -d switch
t/run/switches.t Tests for the other switches (-0, -l, -c, -s, -M, -m, -V, -v, -h, -z, -i)
t/run/switchF1.t Pathological tests for the -F switch
+t/run/switchF2.t Test the -F -a switch combination
t/run/switchF.t Test the -F switch
t/run/switchI.t Test the -I switch
t/run/switchn.t Test the -n switch
diff --git a/perl.c b/perl.c
index 0cfb73c..f079b77 100644
--- a/perl.c
+++ b/perl.c
@@ -3121,16 +3121,18 @@ Perl_moreswitches(pTHX_ const char *s)
if (PL_unicode & PERL_UNICODE_UTF8CACHEASSERT_FLAG)
PL_utf8cache = -1;
return s;
- case 'F':
- PL_minus_F = TRUE;
- PL_splitstr = ++s;
- while (*s && !isSPACE(*s)) ++s;
- PL_splitstr = savepvn(PL_splitstr, s - PL_splitstr);
- return s;
case 'a':
- PL_minus_a = TRUE;
- s++;
- return s;
+ PL_minus_a = TRUE;
+ s++;
+ return s;
+ case 'F':
+ if (!PL_minus_a)
+ Perl_croak(aTHX_ "-F given without -a");
+ PL_minus_F = TRUE;
+ PL_splitstr = ++s;
+ while (*s && !isSPACE(*s)) ++s;
+ PL_splitstr = savepvn(PL_splitstr, s - PL_splitstr);
+ return s;
case 'c':
PL_minus_c = TRUE;
s++;
diff --git a/pod/perldiag.pod b/pod/perldiag.pod
index 600436f..bcef569 100644
--- a/pod/perldiag.pod
+++ b/pod/perldiag.pod
@@ -1882,6 +1882,13 @@ e.g. bless($ref, $p || 'MyPackage');
(A) You've accidentally run your script through B<csh> instead of Perl.
Check the #! line, or manually feed your script into Perl yourself.
+=item -F given without -a
+
+(F) The C<-F> option was passed on the command line but it was not accompanied
+by a C<-a> option. The C<-F> option specifies the pattern to split on if and
+only if C<-a>, which turns on the autosplit mode, is also in effect. See
+L<perlrun> for more details.
+
=item %s failed--call queue aborted
(F) An untrapped exception was raised while executing a UNITCHECK,
diff --git a/t/run/switchF2.t b/t/run/switchF2.t
new file mode 100644
index 0000000..aadc205
--- /dev/null
+++ b/t/run/switchF2.t
@@ -0,0 +1,14 @@
+#!./perl
+
+BEGIN {
+ chdir 't' if -d 't';
+ @INC = '../lib';
+ require './test.pl';
+}
+plan(tests => 1);
+
+my $cmd = "./perl -F: -e 0";
+my $got = `$cmd 2>&1` || '';
+my $failure = $?;
+ok( ($failure and $got =~ /-F given without -a/),
+ "passing -F without -a gives error" );
--
1.6.3.2 |
From @tonycozOn Wed, Jan 02, 2013 at 05:03:49PM -0800, James E Keenan via RT wrote:
As written: tony@mars:.../git/perl$ ./perl -Ilib -F: -ae1 Tony |
From @jkeenanOn Wed Jan 02 19:57:34 2013, tonyc wrote:
Well, so much for that bright idea. :-) Can anyone more familiar with perl.c suggest the correct approach? Thank you very much. |
From @TuxOn Wed, 02 Jan 2013 17:03:49 -0800, "James E Keenan via RT"
Is it too late to make -F *imply* -a ? -- |
From @ap* H.Merijn Brand <h.m.brand@xs4all.nl> [2013-01-07 15:50]:
++ It’s spiteful for no good reason to tell the user that you know what The code to do that is usually much simpler too, as is the case here |
From @ap0001-perl-116190-Make-the-F-switch-imply-a.patchFrom 092e84e616d9cb3cd5a6f8a591bf010ede6f0b30 Mon Sep 17 00:00:00 2001
Message-Id: <092e84e616d9cb3cd5a6f8a591bf010ede6f0b30.1357727233.git.pagaltzis@gmx.de>
From: Aristotle Pagaltzis <pagaltzis@gmx.de>
Date: Wed, 9 Jan 2013 11:26:56 +0100
Subject: [PATCH] [perl #116190] Make the -F switch imply -a
---
MANIFEST | 1 +
perl.c | 1 +
t/run/switchF2.t | 15 +++++++++++++++
3 files changed, 17 insertions(+), 0 deletions(-)
create mode 100644 t/run/switchF2.t
diff --git a/MANIFEST b/MANIFEST
index 2dd04ec..009c0ef 100644
--- a/MANIFEST
+++ b/MANIFEST
@@ -5465,6 +5465,7 @@ t/run/switchd-78586.t See whether bug 78586 is fixed
t/run/switchd.t Test the -d switch
t/run/switches.t Tests for the other switches (-0, -l, -c, -s, -M, -m, -V, -v, -h, -z, -i)
t/run/switchF1.t Pathological tests for the -F switch
+t/run/switchF2.t Test the -F -a switch combination
t/run/switchF.t Test the -F switch
t/run/switchI.t Test the -I switch
t/run/switchn.t Test the -n switch
diff --git a/perl.c b/perl.c
index 65b0a1c..6e7f88d 100644
--- a/perl.c
+++ b/perl.c
@@ -3083,6 +3083,7 @@ Perl_moreswitches(pTHX_ const char *s)
PL_utf8cache = -1;
return s;
case 'F':
+ PL_minus_a = TRUE;
PL_minus_F = TRUE;
PL_splitstr = ++s;
while (*s && !isSPACE(*s)) ++s;
diff --git a/t/run/switchF2.t b/t/run/switchF2.t
new file mode 100644
index 0000000..51136d2
--- /dev/null
+++ b/t/run/switchF2.t
@@ -0,0 +1,15 @@
+#!./perl
+
+BEGIN {
+ chdir 't' if -d 't';
+ @INC = '../lib';
+ require './test.pl';
+}
+plan(tests => 1);
+
+my $cmd = "echo 1 | ./perl -n -F: -e print+\\\@F";
+my $got = `$cmd` || '';
+my $ok = 0 == $?;
+chomp $got;
+ok( ($ok and $got eq 1),
+ "passing -F implies -a" );
--
1.7.4.4
|
From @epaIf -F implies -a, then -a should also imply -n. -- ______________________________________________________________________ |
From @ap* Aristotle Pagaltzis <pagaltzis@gmx.de> [2013-01-09 11:29]:
And I know not how Unix-ish a shell it requires to work correctly. Someone who knows better than I, please review. Regards, |
From @TuxOn Wed, 9 Jan 2013 10:31:20 +0000, Ed Avis <eda@waniasset.com> wrote:
I'm not sure, I remember to have used -paF if a implies -n, -p is negated after it has been set. Be sure not to -- |
From @epa
-paF is equivalent to -pnaF, isn't it? (ISTR that in the old days you had to give -pn, but in recent -- ______________________________________________________________________ |
From @ap* H.Merijn Brand <h.m.brand@xs4all.nl> [2013-01-09 11:45]:
That was my first thought but it turns out to be wrong. If you pass -p |
From @khwilliamsonI agree that -F should imply -a Any patch should include updates to perlrun -- |
From @rjbs* Karl Williamson via RT <perlbug-followup@perl.org> [2013-01-11T16:43:15]
Agreed. I wanted to have a look at whether I could see any weird way in which Also, is there a reason that -a should not imply -n? As near as I can tell, using -n does not preclude -p being added, which means perl -an -F: -E 'say $F[4]' could become: perl -F: -E 'say $F[4]' I don't see a reason. (There's some chance that someone has set PERL5OPT to -- |
From @tonycozOn Wed Jan 09 02:50:19 2013, aristotle wrote:
Have you had another chance to cook another patch? For the test, you might want to consider fresh_perl_is() from test.pl to Tony |
From @tonycozOn Sun Jun 30 22:45:24 2013, tonyc wrote:
Here's 3 changes in addition to aristotle's change: 0002 - change to use fresh_perl_is() 0003 - make -a and -F also imply -n 0004 - alter run_multiple_progs() to supply an empty stdin, two warnings Tony |
From @tonycoz0004-perl-116190-feed-an-empty-stdin-to-run_multiple_prog.patchFrom fc640d8b9d9baaed0ce6de80f25d1ab2c1148452 Mon Sep 17 00:00:00 2001
From: Tony Cook <tony@develop-help.com>
Date: Tue, 16 Jul 2013 14:57:20 +1000
Subject: [PATCH 4/4] [perl #116190] feed an empty stdin to
run_multiple_progs() programs
Two tests for -a were attempting to read stdin and blocking with the -a
implies -n change.
---
t/test.pl | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/t/test.pl b/t/test.pl
index e141b91..9f8ed07 100644
--- a/t/test.pl
+++ b/t/test.pl
@@ -1059,7 +1059,8 @@ sub run_multiple_progs {
print $fh "\n#line 1\n"; # So the line numbers don't get messed up.
print $fh $prog,"\n";
close $fh or die "Cannot close $tmpfile: $!";
- my $results = runperl( stderr => 1, progfile => $tmpfile, $up
+ my $results = runperl( stderr => 1, progfile => $tmpfile,
+ stdin => '', $up
? (switches => ["-I$up/lib", $switch], nolib => 1)
: (switches => [$switch])
);
--
1.7.10.4
|
From @tonycoz0003-perl-116190-F-and-a-now-imply-n.patchFrom 9b2f2102fc3fda3a3013588dae18a938397911d7 Mon Sep 17 00:00:00 2001
From: Tony Cook <tony@develop-help.com>
Date: Tue, 16 Jul 2013 12:11:55 +1000
Subject: [PATCH 3/4] [perl #116190] -F and -a now imply -n
---
perl.c | 2 ++
pod/perlrun.pod | 10 +++++++---
t/run/switchF2.t | 12 +++++++++++-
3 files changed, 20 insertions(+), 4 deletions(-)
diff --git a/perl.c b/perl.c
index 41a62b7..5458c1d 100644
--- a/perl.c
+++ b/perl.c
@@ -3215,12 +3215,14 @@ Perl_moreswitches(pTHX_ const char *s)
case 'F':
PL_minus_a = TRUE;
PL_minus_F = TRUE;
+ PL_minus_n = TRUE;
PL_splitstr = ++s;
while (*s && !isSPACE(*s)) ++s;
PL_splitstr = savepvn(PL_splitstr, s - PL_splitstr);
return s;
case 'a':
PL_minus_a = TRUE;
+ PL_minus_n = TRUE;
s++;
return s;
case 'c':
diff --git a/pod/perlrun.pod b/pod/perlrun.pod
index 05dea4e..dbaa12c 100644
--- a/pod/perlrun.pod
+++ b/pod/perlrun.pod
@@ -265,6 +265,8 @@ is equivalent to
An alternate delimiter may be specified using B<-F>.
+B<-a> implicitly sets B<-n>.
+
=item B<-C [I<number/list>]>
X<-C>
@@ -487,9 +489,11 @@ perl, you can check the value of C<$Config{usesitecustomize}>.
=item B<-F>I<pattern>
X<-F>
-specifies the pattern to split on if B<-a> is also in effect. The
-pattern may be surrounded by C<//>, C<"">, or C<''>, otherwise it will be
-put in single quotes. You can't use literal whitespace in the pattern.
+specifies the pattern to split on for B<-a>. The pattern may be
+surrounded by C<//>, C<"">, or C<''>, otherwise it will be put in single
+quotes. You can't use literal whitespace in the pattern.
+
+B<-F> implicitly sets both B<-a> and B<-n>.
=item B<-h>
X<-h>
diff --git a/t/run/switchF2.t b/t/run/switchF2.t
index 5e99e1e..a411711 100644
--- a/t/run/switchF2.t
+++ b/t/run/switchF2.t
@@ -5,7 +5,7 @@ BEGIN {
@INC = '../lib';
require './test.pl';
}
-plan(tests => 1);
+plan(tests => 3);
{ # perl #116190
fresh_perl_is('print qq!@F!', '1 2',
@@ -13,4 +13,14 @@ plan(tests => 1);
stdin => "1:2",
switches => [ '-n', '-F:' ],
}, "passing -F implies -a");
+ fresh_perl_is('print qq!@F!', '1 2',
+ {
+ stdin => "1:2",
+ switches => [ '-F:' ],
+ }, "passing -F implies -an");
+ fresh_perl_is('print join q!,!, @F', '1,2',
+ {
+ stdin => "1 2",
+ switches => [ '-a' ],
+ }, "passing -a implies -n");
}
--
1.7.10.4
|
From @tonycoz0002-perl-116190-use-the-true-and-trusted-fresh_perl_is.patchFrom 2c8f357c7607c454281c069e555784dde27cd71a Mon Sep 17 00:00:00 2001
From: Tony Cook <tony@develop-help.com>
Date: Tue, 16 Jul 2013 12:00:41 +1000
Subject: [PATCH 2/4] [perl #116190] use the true and trusted fresh_perl_is()
instead of re-inventing it yet again
---
t/run/switchF2.t | 13 +++++++------
1 file changed, 7 insertions(+), 6 deletions(-)
diff --git a/t/run/switchF2.t b/t/run/switchF2.t
index 51136d2..5e99e1e 100644
--- a/t/run/switchF2.t
+++ b/t/run/switchF2.t
@@ -7,9 +7,10 @@ BEGIN {
}
plan(tests => 1);
-my $cmd = "echo 1 | ./perl -n -F: -e print+\\\@F";
-my $got = `$cmd` || '';
-my $ok = 0 == $?;
-chomp $got;
-ok( ($ok and $got eq 1),
- "passing -F implies -a" );
+{ # perl #116190
+ fresh_perl_is('print qq!@F!', '1 2',
+ {
+ stdin => "1:2",
+ switches => [ '-n', '-F:' ],
+ }, "passing -F implies -a");
+}
--
1.7.10.4
|
From @ap* Tony Cook via RT <perlbug-followup@perl.org> [2013-07-16 07:05]:
Thanks for taking this off my todo list. :-) |
From @TuxOn Mon, 15 Jul 2013 22:02:31 -0700, "Tony Cook via RT" to allow -paF, wouldn't it be better to not imply -n if -p was already
-- |
From @ap* H.Merijn Brand <h.m.brand@xs4all.nl> [2013-07-16 09:05]:
Haven’t we gone over this before? No. It would make zero difference. This is the code for parsing the -n and -p switches: case 'n': So if you pass both -n and -p, both flags get set. This is what happens when PL_minus_n is set: $ perl -MO=Deparse -n -e1 And this is what happens when both are set: $ perl -MO=Deparse -p -n -n -n -n -n -n -n -n -n -n -e1 In other words, when both are set, you get PL_minus_p behaviour regardless of You seem to have thought otherwise, but this is how it works. So the guard clauses you suggest are no-ops. Regards, |
From @tonycozOn Mon Jul 15 22:02:31 2013, tonyc wrote:
All four applied as merge commit 1d5bb6b. Tony |
|
@tonycoz - Status changed from 'open' to 'resolved' |
Migrated from rt.perl.org#116190 (status was 'resolved')
Searchable as RT116190$
The text was updated successfully, but these errors were encountered: