Bugzilla Bug 163527

Several people have reported that bash-3.2 doesn't handle quotes properly with
the =~ operator - here's my example:

$ V="alphabet"
$ [[ $V =~ "alphabet" ]] && echo yes

should say "yes" (which it does on 3.1), but it doesn't.  I queried this on the
bash-bugs mailing list, and the upstream maintainer shows how his 3.2 works
properly (http://lists.gnu.org/archive/html/bug-bash/2007-01/msg00090.html)
which puts the ball back in our court.

I've built bash on i686, ppc, ppc64, vanilla and hardened, with and without our
patches (but always to upstream patchlevel 9), and it always does the same for
me - i.e. it always fails to do the comparison properly; it'll only work if the
quotes are dropped.

Remains to be seen if upstream have any suggestions; but it occurred to me that
it could be an impact from something else (glibc, for example).

RedHat see the same behaviour that we do:
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=220087



Here's an emerge info; since it happens on x86 vanilla I figure that's most
relevant.
Portage 2.1.2 (default-linux/x86/2006.1, gcc-4.1.1, glibc-2.4-r4,
2.6.18-gentoo-r6 i686)
=================================================================
System uname: 2.6.18-gentoo-r6 i686 AMD Athlon(tm) XP 2500+
Gentoo Base System version 1.12.6
Timestamp of tree: Wed, 17 Jan 2007 23:30:01 +0000
dev-java/java-config: 1.3.7, 2.0.31
dev-lang/python:     2.4.3-r4
dev-python/pycrypto: 2.0.1-r5
sys-apps/sandbox:    1.2.17
sys-devel/autoconf:  2.13, 2.61
sys-devel/automake:  1.4_p6, 1.5, 1.6.3, 1.7.9-r1, 1.8.5-r3, 1.9.6-r2, 1.10
sys-devel/binutils:  2.16.1-r3
sys-devel/gcc-config: 1.3.14
sys-devel/libtool:   1.5.22
virtual/os-headers:  2.6.17-r2
ACCEPT_KEYWORDS="x86"
AUTOCLEAN="yes"
CBUILD="i686-pc-linux-gnu"
CFLAGS="-O2 -march=athlon-xp -pipe"
CHOST="i686-pc-linux-gnu"
CONFIG_PROTECT="/etc /usr/kde/3.5/env /usr/kde/3.5/share/config
/usr/kde/3.5/shutdown /usr/share/X11/xkb /usr/share/config"
CONFIG_PROTECT_MASK="/etc/env.d /etc/env.d/java/ /etc/gconf
/etc/java-config/vms/ /etc/terminfo /etc/texmf/web2c"
CXXFLAGS="-O2 -march=athlon-xp -pipe"
DISTDIR="/usr/portage/distfiles"
FEATURES="autoaddcvs autoconfig collision-protect cvs distlocks fixpackages
installsources metadata-transfer pax sandbox sfperms sign splitdebug strict
test userpriv"
GENTOO_MIRRORS="http://linux.rz.ruhr-uni-bochum.de/download/gentoo-mirror/
http://gentoo.osuosl.org/"
LC_ALL="en_GB.utf8"
LDFLAGS="-Wl,-O1"
LINGUAS="en_GB en it de es fr"
MAKEOPTS="-j1"
PKGDIR="/usr/portage/packages"
PORTAGE_RSYNC_OPTS="--recursive --links --safe-links --perms --times --compress
--force --whole-file --delete --delete-after --stats --timeout=180
--exclude=/distfiles --exclude=/local --exclude=/packages"
PORTAGE_TMPDIR="/var/tmp"
PORTDIR="/usr/portage"
SYNC="rsync://rsync.gentoo.org/gentoo-portage"
USE="3dnow 3dnowext X a52 aac acl acpi alsa apache2 arts audiofile berkdb
bitmap-fonts bzip2 cairo caps cdparanoia cdr cli cracklib crypt cups dbus
dlloader dri dts dv dvb dvd dvdr dvdread encode ffmpeg firefox flac flash
foomaticdb fortran gcj gdbm gif gpm gstreamer gtk gtk2 hal iconv imagemagick
imap imlib ipod ipv6 isdnlog jack java javascript jpeg jpeg2k kde ldap lesstif
libg++ mad mime ming mmap mmx mmxext mng mp3 mpeg mplayer nas ncurses nls nptl
nptlonly nsplugin offensive ogg opengl pam pcre pda pdf perl png ppds pppd
python qt3 qt4 quicktime readline reflection ruby samba session speex spell spl
sse sse2 ssl svg tcl tcltk tcpd test tetex theora threads tidy tiff tk truetype
truetype-fonts type1-fonts udev unicode usb v4l v4l2 vcd vdr vim-syntax vorbis
wifi win32codecs wmf x264 x86 xine xinerama xml xmlrpc xorg xv xvid zlib"
ALSA_CARDS="ali5451 als4000 atiixp atiixp-modem bt87x ca0106 cmipci emu10k1x
ens1370 ens1371 es1938 es1968 fm801 hda-intel intel8x0 intel8x0m maestro3
trident usb-audio via82xx via82xx-modem ymfpci" ALSA_PCM_PLUGINS="adpcm alaw
asym copy dmix dshare dsnoop empty extplug file hooks iec958 ioplug ladspa
lfloat linear meter mulaw multi null plug rate route share shm softvol"
ELIBC="glibc" INPUT_DEVICES="keyboard mouse" KERNEL="linux" LCD_DEVICES="bayrad
cfontz cfontz633 glk hd44780 lb216 lcdm001 mtxorb ncurses text" LINGUAS="en_GB
en it de es fr" USERLAND="GNU" VIDEO_CARDS="radeon vesa vga"
Unset:  CTARGET, EMERGE_DEFAULT_OPTS, INSTALL_MASK, LANG,
PORTAGE_RSYNC_EXTRA_OPTS, PORTDIR_OVERLAY

Chet is looking at it further as he's confirmed it doesnt work for him on Linux
but does work for him elsewhere

Yep.  Just want to be sure this blocks bash-3.2 from going stable.  We don't
use =~ much (java-vm.eclass, java-utils-2.eclass and my acroread ebuild in the
tree which all have no quotes now, and one commented-out use in baselayout
1.12.8 that does have single quotes) but it could have an impact on user
scripts.

Upstream response:
http://www.mail-archive.com/bug-bash@gnu.org/msg02445.html

Problem is that bash is relying on undefined behaviour in regex handling.  BSD
libc chooses to interpret a back-slashed non-special character in regexes as
just the character, however glibc chooses to interpret them as a back-slash and
the character.

I guess Chet is working to fix this, although he did say it would be awkward.

*** Bug 177755 has been marked as a duplicate of this bug. ***

See also the thread terminating in
http://www.mail-archive.com/bug-bash%40gnu.org/msg02590.html

3.2 behaviour is correct - upshot is that my bug is INVALID :)

FI Patch 10 (the one the above thread is a response to) fixed a separate bug in
the same operator, where bash internally was quoting all characters, rather
than just the ones specifically defined in the standard to be quotable. 
Unfortunately, the behaviour for other characters is undefined by the standard,
and glibc chooses different behaviour to BSD's libc.

*** Bug 178009 has been marked as a duplicate of this bug. ***

(In reply to comment #5)
> 3.2 behaviour is correct - upshot is that my bug is INVALID :)

Reopen to resolve properly.

*** Bug 191485 has been marked as a duplicate of this bug. ***

*** Bug 212346 has been marked as a duplicate of this bug. ***