On Mon, Oct 17, 2016 at 04:56:05PM -0500, Eric Blake wrote:
> Second, your claim that things are "spuriously bad if $RANDOM does not
> exist" is false. Look at the full context:
>
> tmpdir=${TMPDIR-/tmp}/ins$RANDOM-$$
> trap 'ret=$?; rmdir "$tmpdir/d" "$tmpdir" 2>/dev/null
On Mon, Oct 17, 2016 at 04:56:05PM -0500, Eric Blake wrote:
> You're not the first person to complain that $RANDOM is a bashism, and
> this is not the first time we've had to retort that our use of $RANDOM
> is a nicety, but not a necessity, and that the code is perfectly safe
> and tested on shell
On 10/17/2016 03:58 PM, co...@sdf.org wrote:
> Hi,
>
> in build-aux/install-sh scriptversion=2016-01-11.22 line 327
> tmpdir=${TMPDIR-/tmp}/ins$RANDOM-$$
> ... things which are spuriously bad if $RANDOM does not exist ...
Thanks for the report. However, you've made two mistakes.
First, gnulib d
Hi,
in build-aux/install-sh scriptversion=2016-01-11.22 line 327
tmpdir=${TMPDIR-/tmp}/ins$RANDOM-$$
... things which are spuriously bad if $RANDOM does not exist ...
Please use something like:
tmpdir=$(mktemp -d -p ${TMPDIR:-/tmp})
Note also :
NetBSD /bin/sh does not have $RANDOM.
I don't know
On 16/10/16 13:43, Bruno Haible wrote:
Can each of you please give a list of platforms to which you have access and
on which you frequently or occasionally test gnulib modules?
FWIW I run daily builds of all gnulib modules on these (obsolete) systems:
- Solaris 2.6, 7, 8, 9, 10 (SPARC)
- Sol
On 25/09/16 06:16, Paul Eggert wrote:
I attempted to fix the problem by installing the attached patch into
gnulib master. I can't easily test this, though, as I don't have macOS.
The patch did not include the substitution of HAVE_SYS_CDEFS_H needed
when creating sched.h.
Below is from a Cent
On Mon, 17 Oct 2016 11:45:43 +0900
Norihiro Tanaka wrote:
> When dfa builds a state, generates all next states. However, I believe
> most of them are not used.
>
> This patch changes as that when dfa builds a state, generates a next
> state including next input character only.
>
> The followi