[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [sed-devel] build feedback for sed-4.2.2.177-a348
From: |
Jim Meyering |
Subject: |
Re: [sed-devel] build feedback for sed-4.2.2.177-a348 |
Date: |
Tue, 13 Dec 2016 08:51:38 -0800 |
On Mon, Dec 12, 2016 at 5:52 PM, Jim Meyering <address@hidden> wrote:
> On Tue, Dec 6, 2016 at 8:31 AM, Nelson H. F. Beebe <address@hidden> wrote:
>> I've just completed builds of sed-4.2.2.177-a348 on 126 different
>> systems in our test lab. Of those, 82 passed all of the validation
>> suite.
>>
>> Here is a summary of the test failures, sorted first by test name, and
>> then by system name:
>>
> ...
>> borel testsuite/y-zero
>> borel testsuite/zero-anchor
>
> Thanks again for all the testing.
> This exposed a portability problem with the test suite: it assumed
> that "tr" in the C locale would not elide NUL bytes. But Solaris
> 11.3's tr would do precisely that: remove any NUL byte:
>
> $ printf '\0'|wc -c
> 1
> $ printf '\0'|tr -d \\r |wc -c
> 0
>
> Here's the patch I've just pushed:
As Nelson suggested privately, this is not really a portability
problem with the test, but rather simply a bug in Solaris 11.3's tr.
tr has always been specified to process input streams containing NUL
bytes.