GNU bug report logs - #71468
backref-multibyte-slow test fails on Cygwin

Previous Next

To reply to this bug, email your comments to 71468 AT debbugs.gnu.org.

Toggle the display of automated, internal messages from the tracker.

Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Bruno Haible <bruno <at> clisp.org>
To: bug-grep <at> gnu.org
Subject: backref-multibyte-slow test fails on Cygwin
Date: Mon, 10 Jun 2024 14:08:05 +0200

[Message part 1 (text/plain, inline)]

Building the current grep (git master) with the current Gnulib
on Cygwin 3.5.3, I see a test failure:

FAIL: backref-multibyte-slow

The detailed log is attached.

Bruno

[log.gz (application/gzip, attachment)]

Message #8 received at 71468 <at> debbugs.gnu.org (full text, mbox):

From: Collin Funk <collin.funk1 <at> gmail.com>
To: 71468 <at> debbugs.gnu.org, Bruno Haible <bruno <at> clisp.org>
Subject: backref-multibyte-slow test fails on Cygwin
Date: Sat, 01 Mar 2025 21:38:46 -0800

Hi Bruno,

You wrote:
> Building the current grep (git master) with the current Gnulib
> on Cygwin 3.5.3, I see a test failure:

> FAIL: backref-multibyte-slow

The issue here seems to be a carriage return emitted by Perl generated
in these lines:

+ gawk 'BEGIN {for (i=0; i<13000; i++) print "aba"}' /dev/null
++ LC_ALL=C
++ perl -le 'use Time::HiRes qw(time); my $s = time();
              system q,grep -E '\''^([a-z]).\1$'\'' in > junk,;
              my $elapsed = time() - $s; print int (1 + 10 * $elapsed)'
+ max_seconds=$'1\r'
+ test $'1\r' = 1

After that, the variable is used and causes an error to be emitted in
the output. Here is the diff with expected output:

+ diff -u out-en_US.UTF-8 in
--- out-en_US.UTF-8	2024-06-10 05:17:21.428471600 +0000
+++ in	2024-06-10 05:17:21.147110300 +0000
@@ -1,2 +1,13000 @@
-timeout: invalid time interval ‘1\rs’
-Try 'timeout --help' for more information.
+aba

Do you know what Perl is being used on the Cygwin CI machines?

As far as I can tell, you don't install the Cygwin Perl package in
many-platforms.yml [1]. GitHub says Windows CI machines come with Perl
installed [2], but I assume this would be some Windows distribution like
Strawberry Perl [3].

I haven't used Windows enough to know for sure, but I assume the Native
Windows Perl distributions emit a carriage return while the Cygwin
version does not. I believe that something similar occurs with Native
Windows Python vs. Cygwin Python.

Collin

[1] https://github.com/gnu-grep/ci-check/blob/36c4af9a80f2abf8aa67ab8d4c12e9a9c51c2595/.github/workflows/many-platforms.yml#L657
[2] https://github.com/actions/runner-images/blob/a88fc4db06ae2350cb3128627b7fe4b5ee847f8e/images/windows/Windows2022-Readme.md?plain=1#L23
[3] https://strawberryperl.com/

Message #11 received at 71468 <at> debbugs.gnu.org (full text, mbox):

From: Bruno Haible <bruno <at> clisp.org>
To: 71468 <at> debbugs.gnu.org, Collin Funk <collin.funk1 <at> gmail.com>
Subject: Re: backref-multibyte-slow test fails on Cygwin
Date: Sun, 02 Mar 2025 10:16:53 +0100

Hi Collin,

> The issue here seems to be a carriage return emitted by Perl generated
> in these lines:
> 
> + gawk 'BEGIN {for (i=0; i<13000; i++) print "aba"}' /dev/null
> ++ LC_ALL=C
> ++ perl -le 'use Time::HiRes qw(time); my $s = time();
>               system q,grep -E '\''^([a-z]).\1$'\'' in > junk,;
>               my $elapsed = time() - $s; print int (1 + 10 * $elapsed)'
> + max_seconds=$'1\r'
> + test $'1\r' = 1

I see.

> Do you know what Perl is being used on the Cygwin CI machines?
> 
> As far as I can tell, you don't install the Cygwin Perl package in
> many-platforms.yml [1]. GitHub says Windows CI machines come with Perl
> installed [2], but I assume this would be some Windows distribution like
> Strawberry Perl [3].

It's a native Windows Perl (I checked some time ago). The default PATH
contains these elements:
  C:\Strawberry\c\bin
  C:\Strawberry\perl\site\bin
  C:\Strawberry\perl\bin

> I haven't used Windows enough to know for sure, but I assume the Native
> Windows Perl distributions emit a carriage return while the Cygwin
> version does not.

Quite likely, yes.

Bruno

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.