1.8 branch test failure

1.8 branch test failure

[Greg Troxel]

[-- Attachment #1: Type: text/plain, Size: 528 bytes --]


Git tree... release_1-8-6-53-g66f3b6c
GCC version... gcc (GCC) 4.1.3 20080704 prerelease (NetBSD nb1 20080202)
Differences:
End differnces.

[snip]

FAIL: popen.test: open-input-pipe: open-input-pipe process gets (current-input-port) as stdin

[snip]

Totals for this test run:
passes:                 12022
failures:               1
unexpected passes:      0
expected failures:      25
unresolved test cases:  9
untested test cases:    0
unsupported test cases: 13
errors:                 0


This is on netbsd/amd64, 5.0ish.

[-- Attachment #2: Type: application/pgp-signature, Size: 193 bytes --]

Re: 1.8 branch test failure

[Neil Jerram]

Greg Troxel <gdt@ir.bbn.com> writes:

> Git tree... release_1-8-6-53-g66f3b6c
> GCC version... gcc (GCC) 4.1.3 20080704 prerelease (NetBSD nb1 20080202)
> Differences:
> End differnces.
>
> [snip]
>
> FAIL: popen.test: open-input-pipe: open-input-pipe process gets (current-input-port) as stdin

That looks to be caused by a `fix' that I did.  I'll have a
look/think.  If you are able to provide any more detail on the cause
of the failure, that would be great.

     Neil

Re: 1.8 branch test failure

[Neil Jerram]

Neil Jerram <neil@ossau.uklinux.net> writes:

> Greg Troxel <gdt@ir.bbn.com> writes:
>
>> Git tree... release_1-8-6-53-g66f3b6c
>> GCC version... gcc (GCC) 4.1.3 20080704 prerelease (NetBSD nb1 20080202)
>> Differences:
>> End differnces.
>>
>> [snip]
>>
>> FAIL: popen.test: open-input-pipe: open-input-pipe process gets (current-input-port) as stdin
>
> That looks to be caused by a `fix' that I did.  I'll have a
> look/think.  If you are able to provide any more detail on the cause
> of the failure, that would be great.

Could you try the patch below, to show what the result of the (read
...) is?  I'm assuming that there must be a result, otherwise the test
would hang.

Thanks,
        Neil

diff --git a/test-suite/tests/popen.test b/test-suite/tests/popen.test
index 08bfa7c..54ab5a8 100644
--- a/test-suite/tests/popen.test
+++ b/test-suite/tests/popen.test
@@ -80,7 +80,12 @@
                      (open-input-pipe "read && echo $REPLY")))))
       (display "hello\n" (cdr p2c))
       (force-output (cdr p2c))
-      (let ((result (eq? (read port) 'hello)))
+      (let* ((x (read port))
+	     (result (eq? x 'hello)))
+	(newline)
+	(display "read: ")
+	(write x)
+	(newline)
 	(close-port (cdr p2c))
 	(close-pipe port)
 	result)))

Re: 1.8 branch test failure

[Barry Fishman]

Neil Jerram <neil@ossau.uklinux.net> writes:

> Greg Troxel <gdt@ir.bbn.com> writes:
>
>> Git tree... release_1-8-6-53-g66f3b6c
>> GCC version... gcc (GCC) 4.1.3 20080704 prerelease (NetBSD nb1 20080202)
>> Differences:
>> End differnces.
>>
>> [snip]
>>
>> FAIL: popen.test: open-input-pipe: open-input-pipe process gets (current-input-port) as stdin
>
> That looks to be caused by a `fix' that I did.  I'll have a
> look/think.  If you are able to provide any more detail on the cause
> of the failure, that would be great.

The same thing happens with Ubuntu Jaunty.  I Don't think it is Guile
failing.  It the test that caused problems:

    (open-input-pipe "read && echo $REPLY")

In Ubuntu and NetBSD, Bash is not the default shell, so "read" requires
an argument. From the /bin/sh shell (linked to /bin/dash):

$ read && echo $REPLY
read: 1: arg count

It works when popen.test is changed to do:

    (open-input-pipe "read line && echo $line")

-- 
Barry Fishman

Re: 1.8 branch test failure

[Neil Jerram]

Barry Fishman <barry_fishman@acm.org> writes:

> The same thing happens with Ubuntu Jaunty.  I Don't think it is Guile
> failing.  It the test that caused problems:
>
>     (open-input-pipe "read && echo $REPLY")
>
> In Ubuntu and NetBSD, Bash is not the default shell, so "read" requires
> an argument. From the /bin/sh shell (linked to /bin/dash):
>
> $ read && echo $REPLY
> read: 1: arg count
>
> It works when popen.test is changed to do:
>
>     (open-input-pipe "read line && echo $line")

That's fantastic, thanks.  I've committed this change to
branch_release-1-8, and will shortly to master also.

      Neil