Feature #11664
closed[PATCH] introduce rb_autoload_value to replace rb_autoload
Description
rb_autoload_value
may be safer by preventing premature GC. It
can also be more efficient by passing a pre-frozen string that
can be deduped using rb_fstring
. Common autoload callers (e.g.
rubygems, rdoc) already use string literals as the file
argument.
There seems to be no reason to expose rb_autoload_value
to the
public C API since autoload is not performance-critical.
Applications may declare autoloads in Ruby code or via
rb_funcall
; so merely deprecate rb_autoload
without exposing
rb_autoload_value
to new users.
Running: valgrind -v ruby -rrdoc -rubygems -e exit
shows a minor memory reduction (32-bit userspace)
before:
in use at exit: 1,600,621 bytes in 28,819 blocks
total heap usage: 55,786 allocs, 26,967 frees, 6,693,790 bytes allocated
after:
in use at exit: 1,599,778 bytes in 28,789 blocks
total heap usage: 55,739 allocs, 26,950 frees, 6,692,973 bytes allocated
Files
Updated by normalperson (Eric Wong) about 9 years ago
normalperson@yhbt.net wrote:
Comments? It's low-risk, but does deprecate a (probably unused)
public C-API.
Thanks.
Updated by nobu (Nobuyoshi Nakada) about 9 years ago
- Description updated (diff)
The new function looks fine to me, but why deprecating rb_autoload
?
We used to append _str
in many cases, IIRC.
Updated by normalperson (Eric Wong) about 9 years ago
nobu@ruby-lang.org wrote:
The new function looks fine to me, but why deprecating
rb_autoload
?
In general, I prefer to make the C API smaller so it is easier to
support and improve Ruby internals. The rb_autoload API is also
tricky/dangerous with RSTRING_PTR use because of GC.
Since declaring autoload is not performance-critical, users who
need autoload from C-ext can use rb_funcall instead.
We used to append
_str
in many cases, IIRC.
So rename rb_autoload_value => rb_autoload_str?
Updated by normalperson (Eric Wong) about 9 years ago
Eric Wong normalperson@yhbt.net wrote:
nobu@ruby-lang.org wrote:
We used to append
_str
in many cases, IIRC.So rename rb_autoload_value => rb_autoload_str?
Did that, I'd like to commit the following, soon:
http://80x24.org/spew/20151204004808.3288-1-e%4080x24.org/raw
Updated by Anonymous about 9 years ago
- Status changed from Open to Closed
Applied in changeset r52909.
introduce rb_autoload_str to replace rb_autoload
rb_autoload_str may be safer by preventing premature GC. It
can also be more efficient by passing a pre-frozen string that
can be deduped using rb_fstring. Common autoload callers (e.g.
rubygems, rdoc) already use string literals as the file
argument.
There seems to be no reason to expose rb_autoload_str to the
public C API since autoload is not performance-critical.
Applications may declare autoloads in Ruby code or via
rb_funcall; so merely deprecate rb_autoload without exposing
rb_autoload_str to new users.
Running: valgrind -v ruby -rrdoc -rubygems -e exit
shows a minor memory reduction (32-bit userspace)
before:
in use at exit: 1,600,621 bytes in 28,819 blocks
total heap usage: 55,786 allocs, 26,967 frees, 6,693,790 bytes allocated
after:
in use at exit: 1,599,778 bytes in 28,789 blocks
total heap usage: 55,739 allocs, 26,950 frees, 6,692,973 bytes allocated
- include/ruby/intern.h (rb_autoload): deprecate
- internal.h (rb_autoload_str): declare
- load.c (rb_mod_autoload): use rb_autoload_str
- variable.c (rb_autoload): become compatibility wrapper
(rb_autoload_str): hoisted out from old rb_autoload
[ruby-core:71369] [Feature #11664]
Updated by nobu (Nobuyoshi Nakada) about 9 years ago
I don't think there is good enough reason to deprecate rb_autoload
.
Updated by normalperson (Eric Wong) about 9 years ago
nobu@ruby-lang.org wrote:
I don't think there is good enough reason to deprecate
rb_autoload
.
Who uses rb_autoload from C extensions?
It is not performance-critical.
I'd like to make the C API smaller so we can have less to support
in the future. Deprecating unused, non-performance-critical APIs
won't hurt anybody; and hopefully we'll be able to remove the
old function in a few years.
If you want to un-deprecate, go ahead. I just don't see the point
of maintaining the old function forever.
(and maybe autoload will be removed entirely by matz)