Mike Sullivan wrote:
On 11/29/11 10:59, Stefan Teleman wrote:
Alan Coopersmith wrote:
On 11/29/11 10:30, Stefan Teleman wrote:
https://cr.opensolaris.org/action/browse/userland/steleman/bash-4.2-7111705/webrev/
One patch is gone (support/signames.c was accepted upstream), a few
patches had to be refreshed because they no longer applied cleanly.
And you're changing the behavior of the builtin echo command to no
longer be the standard conforming version?
-CONFIGURE_OPTIONS += --enable-usg-echo-default
-CONFIGURE_OPTIONS += --enable-xpg-echo-default
+CONFIGURE_OPTIONS += --enable-usg-echo-default=no
+CONFIGURE_OPTIONS += --enable-xpg-echo-default=no
Yes, I am changing it, because setting the builtin echo to default to
POSIX causes a very large number of regressions in the 4.2 test
harness. Not just on Solaris, but on Linux too.
POSIX can be re-enabled with set -o posix.
So I'm confused. this looks like:
6899504 Default builtin echo behavior should be -E (like before) instead
of current -e
which was recently closed as not-a-bug. are you fixing it, or am I wrong
that this is
related?
That bug was filed against bash 4.1. bash 4.1 did not exhibit
regressions in its test harness, even when the POSIX echo builtin is
enabled as default. Therefore it was not a bug. That was the intended
and expected behavior, according to 4.1's test harness.
bash 4.2 exhibits a very large number of regressions when the builtin
echo is set to default to POSIX. This did not happen in 4.1.
--Stefan
--
Stefan Teleman
Oracle USA Corporation
stefan.tele...@oracle.com
_______________________________________________
userland-discuss mailing list
userland-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/userland-discuss
