From: jean.boussier@gmail.com
To: ruby-core@ruby-lang.org
Subject: [ruby-core:101102] [Ruby master Feature#13381] [PATCH] Expose rb_fstring and its family to C extensions
Date: Thu, 26 Nov 2020 17:44:42 +0000 (UTC) [thread overview]
Message-ID: <redmine.journal-88777.20201126174442.7945@ruby-lang.org> (raw)
In-Reply-To: redmine.issue-13381.20170329052912.7945@ruby-lang.org
Issue #13381 has been updated by byroot (Jean Boussier).
> It sounds like that fstring doesn't match that purpose.
I'm not sure why it wouldn't. Ultimately the prupose is the same than `String#-@`, but from the C API and by passing a `char *`.
> Is it really better to divert fstring than separated string pools?
It would be way less efficient, consider the following case:
```ruby
objects = JSON.load_file('path/to.json') # [{"field": 1}, {"field": 2}, ...]
objects.map { |o| o['field'] }
```
Here `some_field` since it is a literal is part of the fstring table. As of Ruby 2.7 the `json` extension has to `rb_str_new()` many times, before calling `hash_aset` which will then deduplicate these strings.
In some use cases like ours, this generates a huge amount of GC pressure that could be avoided if the `json` extension (and some others) could directly lookup interned strings from a `char` pointer.
----------------------------------------
Feature #13381: [PATCH] Expose rb_fstring and its family to C extensions
https://bugs.ruby-lang.org/issues/13381#change-88777
* Author: eagletmt (Kohei Suzuki)
* Status: Closed
* Priority: Normal
----------------------------------------
https://github.com/ruby/ruby/pull/1559
Currently, C extensions cannot use fstrings. I'd like to use
`rb_fstring_cstr` instead of `rb_str_new_cstr` for static strings in C
extensions to avoid excess allocation.
I think there's several use cases.
- https://github.com/k0kubun/hamlit/blob/v2.8.0/ext/hamlit/hamlit.c#L508-L512
- https://bitbucket.org/ged/ruby-pg/src/e5eb92cca97abc0c6fc168acfad993c2ad314589/ext/pg_connection.c?at=v0.20.0&fileviewer=file-view-default#pg_connection.c-3679
- https://bitbucket.org/ged/ruby-pg/src/e5eb92cca97abc0c6fc168acfad993c2ad314589/ext/pg_copy_coder.c?at=v0.20.0&fileviewer=file-view-default#pg_copy_coder.c-38
--
https://bugs.ruby-lang.org/
next prev parent reply other threads:[~2020-11-26 17:44 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <redmine.issue-13381.20170329052912.7945@ruby-lang.org>
2020-07-17 4:19 ` [ruby-core:99199] [Ruby master Feature#13381] [PATCH] Expose rb_fstring and its family to C extensions ko1
2020-07-17 9:22 ` [ruby-core:99200] " eregontp
2020-07-20 7:25 ` [ruby-core:99236] " nobu
2020-07-20 7:51 ` [ruby-core:99237] " jean.boussier
2020-08-12 6:36 ` [ruby-core:99564] " ko1
2020-09-23 6:07 ` [ruby-core:100083] " ko1
2020-09-23 7:23 ` [ruby-core:100084] " jean.boussier
2020-09-25 1:01 ` [ruby-core:100114] " daniel
2020-11-18 15:34 ` [ruby-core:100935] " jean.boussier
2020-11-26 7:41 ` [ruby-core:101082] " nobu
2020-11-26 17:44 ` jean.boussier [this message]
2020-11-26 18:42 ` [ruby-core:101103] " jean.boussier
2020-11-26 19:02 ` [ruby-core:101104] " jean.boussier
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-list from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
List information: https://www.ruby-lang.org/en/community/mailing-lists/
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=redmine.journal-88777.20201126174442.7945@ruby-lang.org \
--to=ruby-core@ruby-lang.org \
/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 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).