From: Junio C Hamano <gitster@pobox.com>
To: "Gábor Bernát" <bernat@primeranks.net>
Cc: git@vger.kernel.org, peff@peff.net, sunshine@sunshineco.com,
mikachu@gmail.com, cbailey32@bloomberg.net,
Lee.Carver@servicenow.com, mfwitten@gmail.com,
Gabor Bernat <gabor.bernat@gravityrd.com>
Subject: Re: [PATCH] filter-branch: add passed/remaining seconds on progress
Date: Sun, 06 Sep 2015 11:13:58 -0700 [thread overview]
Message-ID: <xmqqpp1va00p.fsf@gitster.mtv.corp.google.com> (raw)
In-Reply-To: <1441545064-3126-1-git-send-email-bernat@primeranks.net> ("Gábor Bernát"'s message of "Sun, 6 Sep 2015 15:11:04 +0200")
Gábor Bernát <bernat@primeranks.net> writes:
> +if date '+%s' 2>/dev/null | grep -q '^[0-9][0-9]*$'
> +then
> + show_seconds=t
> +else
> + show_seconds=
> +fi
> +
> +case "$show_seconds" in
> +t)
> + start_timestamp=$(date +%s)
> + next_sample_at=0
> + ;;
> +'')
> + progress=""
> + ;;
> +esac
Why the case statement here?
For that matter, you probably do not need $show_seconds and use the
fact that start_timestamp and progress are only set to a non-empty
string when we measure progress, i.e. making all of the above
something more like this:
progress= start_timestamp=
if date '+%s' 2>/dev/null | grep -q '^[0-9][0-9]*$'
then
next_sample_at=0
progress="dummy to ensure this is not empty"
start_timestamp=$(date '+%s')
fi
If you are more efficiency-minded, I think you could even do with a
single call to /bin/date, perhaps like so:
next_sample_at= progress=
start_timestamp=$(date '+%s' 2>/dev/null)
case "$start_timestamp" in
*[!0-9]* | "")
# not a "digit only" output
;;
?*)
progress="dummy to ensure this is not empty"
next_sample_at=0
;;
esac
but I suspect that may be going a bit too far ;-).
> while read commit parents; do
> git_filter_branch__commit_count=$(($git_filter_branch__commit_count+1))
> - printf "\rRewrite $commit ($git_filter_branch__commit_count/$commits)"
> +
> + case "$show_seconds" in
> + t)
> + if test $git_filter_branch__commit_count -gt $next_sample_at
> + then
> + now_timestamp=$(date +%s)
> + elapsed_seconds=$(($now_timestamp - $start_timestamp))
> + remaining_second=$(( ($commits - $git_filter_branch__commit_count) * $elapsed_seconds / $git_filter_branch__commit_count ))
> + if test $elapsed_seconds -gt 0
> + then
> + next_sample_at=$(( ($elapsed_seconds + 1) * $git_filter_branch__commit_count / $elapsed_seconds ))
> + else
> + next_sample_at=$(($next_sample_at + 1))
> + fi
> + progress=" ($elapsed_seconds seconds passed, remaining $remaining_second predicted)"
> + fi
> + ;;
> + '')
> + progress=""
> + ;;
> + esac
> +
> + printf "\rRewrite $commit ($git_filter_branch__commit_count/$commits)$progress"
>
> case "$filter_subdir" in
> "")
It would be easier to follow the logic of this loop whose _primary_
point is to rewrite one commit if you moved this part into a helper
function. As written above, the deeply nested case statement whose
purpose is only to show the progress is very distracting.
Then the loop would read like:
while read commit parents
do
: $(( $git_filter_branch__commit_count++ ))
report_progress
case "$filter_subdir" in
...
# all the work that is about rewriting this commit
# comes here.
done
which would be much less noisy and helpful to people who are
interested in learning what it is doing.
As I said earlier in this message, the report_progress helper
function can use the fact that $progress or $start_timestamp are
non-empty if and only if you need to compute and tack $progress at
the end of the output, i.e.
report_progress ()
{
if test -n "$progress"
then
... do rate computation here ...
progress=" ($elapsed_seconds seconds passed,..."
fi
printf "\rRewrite $commit (...)$progress"
}
next prev parent reply other threads:[~2015-09-06 18:14 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <xmqq4mjgyrl9.fsf () gitster ! mtv ! corp ! google ! com>
2015-09-04 15:16 ` [PATCH] filter-branch: add passed/remaining seconds on progress Gábor Bernát
2015-09-04 18:34 ` Junio C Hamano
2015-09-04 20:15 ` Eric Sunshine
2015-09-06 9:49 ` Gabor Bernat
2015-09-06 10:05 ` Eric Sunshine
2015-09-06 13:11 ` Gábor Bernát
2015-09-06 18:13 ` Junio C Hamano [this message]
2015-09-07 12:31 ` Gábor Bernát
2015-09-07 13:44 ` Ramsay Jones
2015-09-07 13:52 ` Gábor Bernát
2015-09-07 20:54 ` Eric Sunshine
2015-09-08 17:32 ` Junio C Hamano
2015-09-08 17:59 ` Eric Sunshine
2015-09-21 19:52 ` Junio C Hamano
2015-09-21 21:22 ` Eric Sunshine
[not found] ` <CALYJoz3xoiB2pVT+r0Nz+EYdE91WX6ypdmieMs1uubg=Vs4bog@mail.gmail.com>
2015-09-22 15:53 ` Gabor Bernat
2015-09-22 16:08 ` Junio C Hamano
2015-09-23 20:48 ` Gabor Bernat
2015-09-08 21:44 ` Jeff King
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=xmqqpp1va00p.fsf@gitster.mtv.corp.google.com \
--to=gitster@pobox.com \
--cc=Lee.Carver@servicenow.com \
--cc=bernat@primeranks.net \
--cc=cbailey32@bloomberg.net \
--cc=gabor.bernat@gravityrd.com \
--cc=git@vger.kernel.org \
--cc=mfwitten@gmail.com \
--cc=mikachu@gmail.com \
--cc=peff@peff.net \
--cc=sunshine@sunshineco.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.