From sven.strickroth@tu-clausthal.de

perl -v
This is perl, v5.8.8 built for msys

Perl readline creates "sys$command" files when no STDIN is connected to
the perl process.

I noticed this issue on using msysgit with TortoiseGit. I tracked down
which msysgit versions are affected (<=1.7.7 work, but >=1.7.7.1 has the
issue), but both use the same perl versions (5.8.8) and
Termin/Readline.pl (1.02). I don't know how this issue was introduced,
but the reason seems to be that there is no "con"-file and $^O is
'msys', but a check is missing for this (in Term/ReadLine.pm, sub
findConsole, line 214 - see attached patch).

--
Best regards,
Sven Strickroth

From sven.strickroth@tu-clausthal.de

--- "a/ReadLine.pm"	
+++ "b/ReadLine.pm"	
@@ -214,7 +214,7 @@ sub findConsole {
         $console = "Dev:Console";
     } elsif (-e "/dev/tty") {
 	$console = "/dev/tty";
-    } elsif (-e "con" or $^O eq 'MSWin32') {
+    } elsif (-e "con" or $^O eq 'MSWin32' or $^O eq 'msys') {
 	$console = "con";
     } else {
 	$console = "sys\$command";

From @craigberry

On Sat, Nov 24, 2012 at 12​:26 PM, Sven Strickroth
<perlbug-followup@​perl.org> wrote​:

# New Ticket Created by Sven Strickroth
# Please include the string​: [perl #115900]
# in the subject line of all future correspondence about this issue.
# <URL​: https://rt-archive.perl.org/perl5/Ticket/Display.html?id=115900 >

perl -v
This is perl, v5.8.8 built for msys

Perl readline creates "sys$command" files when no STDIN is connected to
the perl process.

I noticed this issue on using msysgit with TortoiseGit. I tracked down
which msysgit versions are affected (<=1.7.7 work, but >=1.7.7.1 has the
issue), but both use the same perl versions (5.8.8) and
Termin/Readline.pl (1.02). I don't know how this issue was introduced,
but the reason seems to be that there is no "con"-file and $^O is
'msys', but a check is missing for this (in Term/ReadLine.pm, sub
findConsole, line 214 - see attached patch).

Thanks, I've applied your patch (manually, as GNU patch couldn't parse it) as​:

<http​://perl5.git.perl.org/perl.git/commitdiff/0a616d17c63ccea3e07a9370becf1d236ce89558>

Then I went a bit further and did​:

commit c0788ef
Author​: Craig A. Berry <craigberry@​mac.com>
Date​: Sun Nov 25 08​:54​:32 2012 -0600

  Rational findConsole dispatch for Term​::ReadLine.

  Back in 5.002 or so, if we didn't find /dev/tty and weren't on
  Windows, the console was assumed to be sys$command, which only
  makes sense on VMS (possibly $^O didn't work yet on VMS?).

  Later accretions have assumed that the sys$command default meant
  something other than laziness and a second if block with various
  specific overrides was added, some of which set the console back
  to undef after its having been set to sys$command.

  That can all be avoided by simply checking we're on VMS before
  setting the console to sys$command and letting it default to
  STDIN for cases where we don't know of something else specific
  that it should be.

diff --git a/dist/Term-ReadLine/lib/Term/ReadLine.pm
b/dist/Term-ReadLine/lib/Term/ReadLine.pm
index 6f94981..115ce15 100644
--- a/dist/Term-ReadLine/lib/Term/ReadLine.pm
+++ b/dist/Term-ReadLine/lib/Term/ReadLine.pm
@​@​ -238,19 +238,12 @​@​ sub findConsole {
  } elsif (-e "con" or $^O eq 'MSWin32' or $^O eq 'msys') {
  $console = 'CONIN$';
  $consoleOUT = 'CONOUT$';
- } else {
+ } elsif ($^O eq 'VMS') {
  $console = "sys\$command";
- }
-
- if (($^O eq 'amigaos') || ($^O eq 'beos') || ($^O eq 'epoc')) {
- $console = undef;
- }
- elsif ($^O eq 'os2') {
- if ($DB​::emacs) {
- $console = undef;
- } else {
+ } elsif ($^O eq 'os2' && !$DB​::emacs) {
  $console = "/dev/con";
- }
+ } else {
+ $console = undef;
  }

  $consoleOUT = $console unless defined $consoleOUT;
[end]

So I think the ticket can be closed. This is unlikely to make it into
the wild before 5.18, but it would be nice if the OP could verify that
all is well on msys before then.

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

From email@cs-ware.de

On So. 25. Nov. 2012, 08​:16​:12, craig.a.berry@​gmail.com wrote​:

Then I went a bit further and did​:

commit c0788ef
Author​: Craig A. Berry <craigberry@​mac.com>
Date​: Sun Nov 25 08​:54​:32 2012 -0600

So I think the ticket can be closed. This is unlikely to make it into
the wild before 5.18, but it would be nice if the OP could verify that
all is well on msys before then.

Works for me.

Sven

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

From sven.strickroth@tu-clausthal.de

Am 25.11.2012 17​:16 schrieb Craig Berry via RT​:

Then I went a bit further and did​:

commit c0788ef
Author​: Craig A. Berry <craigberry@​mac.com>
Date​: Sun Nov 25 08​:54​:32 2012 -0600

So I think the ticket can be closed. This is unlikely to make it into
the wild before 5.18, but it would be nice if the OP could verify that
all is well on msys before then.

Works for me.

--
Best regards,
Sven Strickroth