bug-sed
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

bug#36128: Failed tests for sed 4.7 on Solaris 10 x86


From: Assaf Gordon
Subject: bug#36128: Failed tests for sed 4.7 on Solaris 10 x86
Date: Wed, 25 Sep 2019 11:57:13 -0600
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101 Thunderbird/60.8.0

Hello,

On 2019-09-24 6:15 a.m., Dagobert Michelsen wrote:

Am 07.06.2019 um 14:55 schrieb Dagobert Michelsen <address@hidden>:
I noticed two failed tests on Solaris 10 x86, please see the attached 
test-suite.log:

FAIL: testsuite/misc.pl
FAIL: testsuite/debug.pl

Please let me know if you need further information or if I should test anything.

The problem still exists and I would really like to release an updated sed.
Can someone please have a look?

From a cursory check,
I think this is an issue in the interplay between the old perl (5.8.4)
and the old shell (/bin/sh) on solaris 10.

---

For example, the "space" test (which is the first failure in the log)
is defined like so:
      ['space', q('s/_\S/XX/g;s/\s/_/g'),
          {IN=>  "Hello World\t!\nSecond_line_ of tests\n" },
          {OUT=> "Hello_World_!\nSecondXXine__of_tests\n" }],

When I run it interactively on OpenCSW's unstable10x, it works as expected:

   > printf 'Hello World\t!\nSecond_line_ of tests\n' > space.in
   > ./sed/sed 's/_\S/XX/g;s/\s/_/g' space.in
   Hello_World_!
   SecondXXine__of_tests

The failure in the log can be reproduced interactively if the backslashes are removed from the sed script:

   > ./sed/sed 's/_S/XX/g;s/s/_/g' space.in
   Hello World     !
   Second_line_ of te_t_

I take it as a hint that the perl quoting of the string using q('')
then going through /bin/sh somehow "loses" these backslashes.

---

Another example:
If I rebuild sed and forces the shell to be /usr/bin/bash,
all tests pass:

    ./configure SHELL=/usr/bin/bash
    make
    make check

---

As such, I think these are false positives, and the sed-4.7 binary
should work fine on Solaris 10 x86.

It might be worth fixing the tests, or detecting problematic shell
and working around it, but I can't get to that in the near future
(patches are of course welcomed).

Hope this helps,
 - assaf






reply via email to

[Prev in Thread] Current Thread [Next in Thread]