Bug #11753: Building ruby-git on Arch Linux eats all memory on ./configure

• Status: Third Party's Issue
• Assignee: -
• Target version: -
• Backport: 2.0.0: UNKNOWN, 2.1: UNKNOWN, 2.2: UNKNOWN

[bararchy (Bar Hofesh)]

Hi,

I'm trying to build Ruby from sources.

What I tried:

1. Build from git using git clone, autoconf, ./configure
2. Build with RVM versions 2.2.1, 2.2-head, 2.1.5, ruby-head

All have failed when reaching the "./configure" phase.
It seems that all memory is getting eat up and the system hangs unless a quick "Ctrl + C" is pressed to stop the process.

Notes:

• I had ruby installed and working until trying to re-build
• I had a system restore which left most of /home/ with wrong permissions (was taken care of but maybe Ruby's configure needs some files with +x which don't have them right now?)
• Other programs which relays on "./configure" scripts work and build without issues.

[bararchy (Bar Hofesh)]

Bar Hofesh wrote:

Hi,

I'm trying to build Ruby from sources.

What I tried:

1. Build from git using git clone, autoconf, ./configure
2. Build with RVM versions 2.2.1, 2.2-head, 2.1.5, ruby-head

All have failed when reaching the "./configure" phase.
It seems that all memory is getting eat up and the system hangs unless a quick "Ctrl + C" is pressed to stop the process.

Notes:

• I had ruby installed and working until trying to re-build
• I had a system restore which left most of /home/ with wrong permissions (was taken care of but maybe Ruby's configure needs some files with +x which don't have them right now?)
• Other programs which relays on "./configure" scripts work and build without issues.

Ok, using the "echo here and there" debug method I got to the part where it is getting stuck:

echo "Testing 1"
if test x"${_as_can_reexec}" != xno && test "x$CONFIG_SHELL" != x; then
  _as_can_reexec=no; export _as_can_reexec;
echo "Testing 2"

I see "Testing 1"
But not "Testing 2"

Removing those lines get me into an infinite loop

[nobu (Nobuyoshi Nakada)]

• Status changed from Open to Third Party's Issue

It's a part of standard AC_INIT().
Probably a bug of Arch Linux which autoconf does not support yet.

Please report it to Arch Linux and/or autoconf,
with exact versions of the OS and autoconf.