From: Junio C Hamano <gitster@pobox.com>
To: Karthik Nayak <karthik.188@gmail.com>
Cc: git@vger.kernel.org, christian.couder@gmail.com,
Matthieu.Moy@grenoble-inp.fr
Subject: Re: [PATCH v4 01/10] ref-filter: add option to align atoms to the left
Date: Fri, 24 Jul 2015 16:00:56 -0700 [thread overview]
Message-ID: <xmqqpp3hjh3r.fsf@gitster.dls.corp.google.com> (raw)
In-Reply-To: <xmqqwpxpjk6p.fsf@gitster.dls.corp.google.com> (Junio C. Hamano's message of "Fri, 24 Jul 2015 14:54:22 -0700")
Junio C Hamano <gitster@pobox.com> writes:
> Karthik Nayak <karthik.188@gmail.com> writes:
>
>> - if (!ref->value) {
>> - populate_value(ref);
>> + /*
>> + * If the atom is a pseudo_atom then we re-populate the value
>> + * into the ref_formatting_state stucture.
>> + */
>> + if (!ref->value || ref->value[atom].pseudo_atom) {
>> + populate_value(state, ref);
>> fill_missing_values(ref->value);
>
> I am not sure why you need to do this. populate_value() and
> fill_missing_values() are fairly heavy-weight operations that are
> expected to grab everything necessary from scratch, and that is why
> we ensure that we do not call them more than once for each "ref"
> with by guarding the calls with "if (!ref->value)".
>
> This change is breaking that basic arrangement, and worse yet, it
> forces us re-read everything about that ref, leaking old ref->value.
>
> Why could this be a good idea?
I think populate_value() should not take state; that is the root
cause of this mistake.
The flow should be:
- verify_format() looks at the format string and enumerates all
atoms that will ever be used in the output by calling
parse_atom() and letting it to fill used_atom[];
- when ref->value is not yet populated, populate_value() is
called, just once. This uses the enumeration in used_atom[]
and stores computed value to refs->value[atom];
- show_ref() repeatedly calls find_next() to find the next
reference to %(...), emits everything before it, and then
uses the atom value (i.e. ref->value[atom]).
I would expect that the atom value for pseudos like color and align
to be parsed and stored in ref->value in populate_value() when it is
called for the first time for each ref _just once_.
"color:blue" may choose to store "blue" as v->s, and "align:4" may
choose to do "v->ul = 4".
And the code that uses these values should look more like:
for (cp = format; *cp && (sp = find_next(cp)); cp = ep + 1) {
struct atom_value *atomv;
ep = strchr(sp, ')');
if (cp < sp)
emit(cp, sp);
get_ref_atom_value(info, parse_ref_filter_atom(sp + 2, ep), &atomv);
if (atomv->is_pseudo)
apply_pseudo_state(&state, atomv);
else
print_value(&state, atomv);
}
where apply_pseudo_state() would switch on what kind of pseudo the
atom is and update the state accordingly, i.e. the "state" munging
code you added to populate_value() in this patch.
next prev parent reply other threads:[~2015-07-24 23:06 UTC|newest]
Thread overview: 36+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-07-24 19:04 [PATCH v4 0/10] port tag.c to use ref-filter APIs Karthik Nayak
2015-07-24 19:04 ` [PATCH v4 01/10] ref-filter: add option to align atoms to the left Karthik Nayak
2015-07-24 21:54 ` Junio C Hamano
2015-07-24 23:00 ` Junio C Hamano [this message]
2015-07-25 4:14 ` Karthik Nayak
2015-07-26 4:08 ` Eric Sunshine
2015-07-26 4:36 ` Karthik Nayak
2015-07-26 5:58 ` Eric Sunshine
2015-07-26 6:03 ` Karthik Nayak
2015-07-27 12:01 ` Matthieu Moy
2015-07-27 0:39 ` Duy Nguyen
2015-07-27 7:39 ` Jacob Keller
2015-07-27 10:18 ` Duy Nguyen
2015-07-28 10:35 ` Duy Nguyen
2015-07-24 19:04 ` [PATCH v4 02/10] ref-filter: make the 'color' use ref_formatting_state Karthik Nayak
2015-07-24 21:46 ` Junio C Hamano
2015-07-25 4:15 ` Karthik Nayak
2015-07-26 4:28 ` Eric Sunshine
2015-07-26 4:12 ` Eric Sunshine
2015-07-26 5:40 ` Karthik Nayak
2015-07-24 19:04 ` [PATCH v4 03/10] ref-filter: add option to filter only tags Karthik Nayak
2015-07-24 19:04 ` [PATCH v4 04/10] ref-filter: support printing N lines from tag annotation Karthik Nayak
2015-07-26 4:46 ` Eric Sunshine
2015-07-26 5:15 ` Karthik Nayak
2015-07-24 19:04 ` [PATCH v4 05/10] ref-filter: add support to sort by version Karthik Nayak
2015-07-25 22:40 ` Junio C Hamano
2015-07-26 5:07 ` Karthik Nayak
2015-07-27 15:24 ` Junio C Hamano
2015-07-27 17:03 ` Karthik Nayak
2015-07-24 19:04 ` [PATCH v4 06/10] ref-filter: add option to match literal pattern Karthik Nayak
2015-07-26 5:15 ` Eric Sunshine
2015-07-26 18:21 ` Karthik Nayak
2015-07-24 19:04 ` [PATCH v4 07/10] tag.c: use 'ref-filter' data structures Karthik Nayak
2015-07-24 19:19 ` [PATCH v4 08/10] tag.c: use 'ref-filter' APIs Karthik Nayak
2015-07-24 19:19 ` [PATCH v4 09/10] tag.c: implement '--format' option Karthik Nayak
2015-07-24 19:19 ` [PATCH v4 10/10] tag.c: implement '--merged' and '--no-merged' options Karthik Nayak
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
List information: http://vger.kernel.org/majordomo-info.html
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=xmqqpp3hjh3r.fsf@gitster.dls.corp.google.com \
--to=gitster@pobox.com \
--cc=Matthieu.Moy@grenoble-inp.fr \
--cc=christian.couder@gmail.com \
--cc=git@vger.kernel.org \
--cc=karthik.188@gmail.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.
Code repositories for project(s) associated with this public inbox
https://80x24.org/mirrors/git.git
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).