ARGF supports compaction, but it pins all of its references, which means
that none of it can move. This commit changes it to actually support
compaction.
[Feature #21109]
By always freezing when setting the global rb_rs variable, we can ensure
it is not modified and can be accessed from a ractor.
We're also making sure it's an instance of String and does not have any
instance variables.
Of course, if $/ is changed at runtime, it may cause surprising behavior
but doing so is deprecated already anyway.
Co-authored-by: Jean Boussier <jean.boussier@gmail.com>
This does not change any actual behaviour, but provides a choke point for blocking IO operations.
* Update `IO::Buffer` to use `rb_io_blocking_region`.
* Update `File` to use `rb_io_blocking_region`.
* Update `IO` to use `rb_io_blocking_region`.
Since `IO.new` accepts one or two positional arguments except for the
optional hash argument, exclude the optional hash argument from the
check for delegation to `IO.new`.
This `st_table` is used to both mark and pin classes
defined from the C API. But `vm->mark_object_ary` already
does both much more efficiently.
Currently a Ruby process starts with 252 rooted classes,
which uses `7224B` in an `st_table` or `2016B` in an `RArray`.
So a baseline of 5kB saved, but since `mark_object_ary` is
preallocated with `1024` slots but only use `405` of them,
it's a net `7kB` save.
`vm->mark_object_ary` is also being refactored.
Prior to this changes, `mark_object_ary` was a regular `RArray`, but
since this allows for references to be moved, it was marked a second
time from `rb_vm_mark()` to pin these objects.
This has the detrimental effect of marking these references on every
minors even though it's a mostly append only list.
But using a custom TypedData we can save from having to mark
all the references on minor GC runs.
Addtionally, immediate values are now ignored and not appended
to `vm->mark_object_ary` as it's just wasted space.
This frees FL_USER0 on both T_MODULE and T_CLASS.
Note: prior to this, FL_SINGLETON was never set on T_MODULE,
so checking for `FL_SINGLETON` without first checking that
`FL_TYPE` was `T_CLASS` was valid. That's no longer the case.
[Bug #20169]
Embedded strings are not safe for system calls without the GVL because
compaction can cause pages to be locked causing the operation to fail
with EFAULT. This commit changes io_fwrite to use rb_str_tmp_frozen_no_embed_acquire,
which guarantees that the return string is not embedded.
The macro SafeStringValue() became just StringValue() in c5c05460ac,
and it is deprecated nowadays.
This patch replaces remaining macro usage. Some occurrences are left in
ext/stringio and ext/win32ole, they should be fixed upstream.
The macro itself is not deleted, because it may be used in extensions.
When opening a file with `File.open`, and then setting the encoding with
`IO#set_encoding`, it still correctly performs CRLF -> LF conversion on
Windows when reading files with a CRLF line ending in them (in text
mode).
However, the file is opened instead with either the `rb_io_fdopen` or
`rb_file_open` APIs from C, the CRLF conversion is _NOT_ set up
correctly; it works if the encoding is not specified, but if
`IO#set_encoding` is called, the conversion stops happening. This seems
to be because the encflags never get ECONV_DEFAULT_NEWLINE_DECORATOR
set in these codepaths.
Concretely, this means that the conversion doesn't happen in the
following circumstances:
* When loading ruby files with require (that calls rb_io_fdopen)
* When parsing ruuby files with RubyVM::AbstractSyntaxTree (that calls
rb_file_open).
This then causes the ErrorHighlight tests to fail on windows if git has
checked them out with CRLF line endings - the error messages it's
testing wind up with literal \r\n sequences in them because the iseq
text from the parser contains un-newline-converted strings.
This commit fixes the problem by copy-pasting the relevant snippet which
sets this up in `rb_io_extract_modeenc` (for the File.open path) into
the relevant codepaths for `rb_io_fdopen` and `rb_file_open`.
[Bug #20101]
Before this patch, the MN scheduler waits for the IO with the
following steps:
1. `poll(fd, timeout=0)` to check fd is ready or not.
2. if fd is not ready, waits with MN thread scheduler
3. call `func` to issue the blocking I/O call
The advantage of advanced `poll()` is we can wait for the
IO ready for any fds. However `poll()` becomes overhead
for already ready fds.
This patch changes the steps like:
1. call `func` to issue the blocking I/O call
2. if the `func` returns `EWOULDBLOCK` the fd is `O_NONBLOCK`
and we need to wait for fd is ready so that waits with MN
thread scheduler.
In this case, we can wait only for `O_NONBLOCK` fds. Otherwise
it waits with blocking operations such as `read()` system call.
However we don't need to call `poll()` to check fd is ready
in advance.
With this patch we can observe performance improvement
on microbenchmark which repeats blocking I/O (not
`O_NONBLOCK` fd) with and without MN thread scheduler.
```ruby
require 'benchmark'
f = open('/dev/null', 'w')
f.sync = true
TN = 1
N = 1_000_000 / TN
Benchmark.bm{|x|
x.report{
TN.times.map{
Thread.new{
N.times{f.print '.'}
}
}.each(&:join)
}
}
__END__
TN = 1
user system total real
ruby32 0.393966 0.101122 0.495088 ( 0.495235)
ruby33 0.493963 0.089521 0.583484 ( 0.584091)
ruby33+MN 0.639333 0.200843 0.840176 ( 0.840291) <- Slow
this+MN 0.512231 0.099091 0.611322 ( 0.611074) <- Good
```
