Re: $PS0 patch

[Dan Stromberg]

Hello?  I sent the message I'm replying to 5 days ago, and no one has replied.

Is there anyone on the list with commit access for bash?

The reason I don't just use trap DEBUG is that trap debug can be prohibitively verbose at times:

$ for i in $(seq 5); do echo $((i*3)); done
cmd started 2015 Thu Oct 29 07:14:50 AM PDT
cmd started 2015 Thu Oct 29 07:14:50 AM PDT
3
cmd started 2015 Thu Oct 29 07:14:50 AM PDT
cmd started 2015 Thu Oct 29 07:14:50 AM PDT
6
cmd started 2015 Thu Oct 29 07:14:50 AM PDT
cmd started 2015 Thu Oct 29 07:14:50 AM PDT
9
cmd started 2015 Thu Oct 29 07:14:50 AM PDT
cmd started 2015 Thu Oct 29 07:14:50 AM PDT
12
cmd started 2015 Thu Oct 29 07:14:50 AM PDT
cmd started 2015 Thu Oct 29 07:14:50 AM PDT
15
dstromberg@benchbox:~ x86_64-pc-linux-gnu 28861 - above cmd done 2015 Thu Oct 29 07:14:50 AM PDT

$

It's also noisy for other one-liners that involve pipes, semicolons, and especially: the following CPU-busy loop is turned into an I/O heavy loop:

$ while : ; do : ; done
cmd started 2015 Thu Oct 29 01:57:45 PM PDT
cmd started 2015 Thu Oct 29 01:57:45 PM PDT
cmd started 2015 Thu Oct 29 01:57:45 PM PDT
cmd started 2015 Thu Oct 29 01:57:45 PM PDT
cmd started 2015 Thu Oct 29 01:57:46 PM PDT
cmd started 2015 Thu Oct 29 01:57:46 PM PDT
cmd started 2015 Thu Oct 29 01:57:46 PM PDT
cmd started 2015 Thu Oct 29 01:57:46 PM PDT
cmd started 2015 Thu Oct 29 01:57:46 PM PDT
...

Are there any changes that need to be made before the patch can be merged?

I'd like to be able to log my commands' start and stop times (for good post mortem analysis), without overwhelming a disk or increasing my cut-and-paste spruce-up time.

Thanks.

On Sat, Oct 24, 2015 at 9:34 PM, Dan Stromberg <dstromberglists@gmail.com> wrote:

I just noticed there's an fprintf buglet in there.

I've put a fixed version of at http://stromberg.dnsalias.org/~strombrg/diffs/ps0-diffs - rather than inflict another shar on the list :)

Thanks again.