(https://github.com/ruby/irb/pull/648)
IRB already has several features that rely on rdoc, such as:
- Autocompletion's document dialog
- Autocompletion's `PerfectMatchedProc`
- The `show_doc` command
- Easter egg
And we could use its pager more in the future too. So it makes sense to
declare rdoc as a dependency instead of relying on the one bundled with
Ruby.
https://github.com/ruby/irb/commit/4dffbb1dd3
Remove rb_control_frame_t::__bp__ and optimize bmethod calls
This commit removes the __bp__ field from rb_control_frame_t. It was
introduced to help MJIT, but since MJIT was replaced by RJIT, we can use
vm_base_ptr() to compute it from the SP of the previous control frame
instead. Removing the field avoids needing to set it up when pushing new
frames.
Simply removing __bp__ would cause crashes since RJIT and YJIT used a
slightly different stack layout for bmethod calls than the interpreter.
At the moment of the call, the two layouts looked as follows:
┌────────────┐ ┌────────────┐
│ frame_base │ │ frame_base │
├────────────┤ ├────────────┤
│ ... │ │ ... │
├────────────┤ ├────────────┤
│ args │ │ args │
├────────────┤ └────────────┘<─prev_frame_sp
│ receiver │
prev_frame_sp─>└────────────┘
RJIT & YJIT interpreter
Essentially, vm_base_ptr() needs to compute the address to frame_base
given prev_frame_sp in the diagrams. The presence of the receiver
created an off-by-one situation.
Make the interpreter use the layout the JITs use for iseq-to-iseq
bmethod calls. Doing so removes unnecessary argument shifting and
vm_exec_core() re-entry from the interpreter, yielding a speed
improvement visible through `benchmark/vm_defined_method.yml`:
patched: 7578743.1 i/s
master: 4796596.3 i/s - 1.58x slower
C-to-iseq bmethod calls now store one more VALUE than before, but that
should have negligible impact on overall performance.
Note that re-entering vm_exec_core() used to be necessary for firing
TracePoint events, but that's no longer the case since
9121e57a5f.
Closesruby/ruby#6428
[Feature #18885]
For now, the optimizations performed are:
- Run a major GC
- Compact the heap
- Promote all surviving objects to oldgen
Other optimizations may follow.
In `rb_ruby_ripper_parser_allocate`, `r->p` is NULL between creating
`self` and `parser_params` assignment. As GC can happen there, the
typed-data functions for it need to consider the case.
As long as "floating point numbers" cannot accurately represent an
irrational number, the result of the natural logarithm cannot be
accurate. Logarithms with an integer base may have the possibility to
represent more accurately.
There’s no reason to prevent RUBYOPT from controlling the backtrace
limit. In fact, Matz said [0] he was expecting this to be possible.
[0] https://bugs.ruby-lang.org/issues/8661#note-27
(https://github.com/ruby/irb/pull/642)
* Define RelineInputMethod::HISTORY
The HistorySavingAbility module doesn't do anything if the input method
class doesn't define HISTORY.
- 3ac96be660/lib/irb/history.rb (L10)
- 3ac96be660/lib/irb/history.rb (L34)
This patch defines RelineInputMethod::HISTORY to avoid this.
* Improve history-saving's ability check
Instead of checking the existence of `input_method_class::HISTORY`, we should
make every input method class declare if it supports history saving or not.
Since the default value is `false`, it shouldn't break any custom input method
that inherits from `IRB::InputMethod`.
https://github.com/ruby/irb/commit/aec7a5b3f5
Nobuyoshi Nakada
Stan Lo
Ivanov-Anton
Hiroshi SHIBATA
Dimitar Haralanov
Alan Wu
Maxime Chevalier-Boisvert
Jean Boussier
yui-knk
tomoya ishida
git
ima1zumi
Tom Stuart
alexandre184
Marcelo Pereira
Jemma Issroff
Peter Zhu
jinroq
Kazuhiro NISHIYAMA
Jeremy Evans
Takashi Kokubun
Matt Valentine-House