gh-77273: Better bytecodes for f-strings#6132
Conversation
Sorry, something went wrong.
markshannon
force-pushed
the
better-bytecode-fstring
branch
from
554c1b6 to
d79893a
Compare
March 17, 2018 13:08
markshannon
force-pushed
the
better-bytecode-fstring
branch
from
35dea3c to
6a88b26
Compare
May 6, 2019 16:10
|
Any update on this @taleinat suggested sharing benchmarks here https://bugs.python.org/issue33092. Otherwise due to lack of activity I think we can close this. |
Sorry, something went wrong.
markshannon
force-pushed
the
better-bytecode-fstring
branch
from
6a88b26 to
cc0f91e
Compare
December 1, 2020 16:25
markshannon
force-pushed
the
better-bytecode-fstring
branch
from
cc0f91e to
7c1f2a6
Compare
August 25, 2021 13:30
|
This PR is stale because it has been open for 30 days with no activity. |
Sorry, something went wrong.
github-actions
Bot
removed
the
stale
Stale PR or inactive for long period of time.
label
May 14, 2023
markshannon
requested review from
brettcannon,
ericsnowcurrently,
iritkatriel,
ncoghlan and
warsaw
as code owners
June 8, 2023 11:32
markshannon
force-pushed
the
better-bytecode-fstring
branch
from
7c1f2a6 to
3af4084
Compare
June 8, 2023 11:34
|
Updated to use the new bytecode definition format. Removes the conditional stack effect, simplifying tooling around these instructions. |
Sorry, something went wrong.
|
Looks like you missed out the doc update. |
Sorry, something went wrong.
Sorry, something went wrong.
* main: (57 commits) pythongh-105831: Fix NEWS blurb from pythongh-105828 (python#105833) pythongh-105820: Fix tok_mode expression buffer in file & readline tokenizer (python#105828) pythongh-105751, test_ctypes: Remove disabled tests (python#105826) pythongh-105821: Use a raw f-string in test_httpservers.py (python#105822) pythongh-105751: Remove platform usage in test_ctypes (python#105819) pythongh-105751: Reenable disable test_ctypes tests (python#105818) pythongh-105751: Remove dead code in test_ctypes (python#105817) More reorganisation of the typing docs (python#105787) Improve docs for `typing.dataclass_transform` (python#105792) pythonGH-89812: Churn `pathlib.Path` test methods (python#105807) pythongh-105800: Issue SyntaxWarning in f-strings for invalid escape sequences (python#105801) pythongh-105751: Cleanup test_ctypes imports (python#105803) pythongh-105481: add HAS_JUMP flag to opcode metadata (python#105791) pythongh-105751: test_ctypes avoids the operator module (pythonGH-105797) pythongh-105751: test_ctypes: Remove @need_symbol decorator (pythonGH-105798) pythongh-104909: Implement conditional stack effects for macros (python#105748) pythongh-75905: Remove test_xmlrpc_net: skipped since 2017 (python#105796) pythongh-105481: Fix types and a bug for pseudos (python#105788) Update DSL docs for cases generator (python#105753) pythonGH-77273: Better bytecodes for f-strings (pythonGH-6132) ...
Reviewers
7 participants
Simplify and speed up interpreter for f-strings.
tldr; Do work in the compiler instead of the interpreter.
Splits the FORMAT_VALUE opcode into CONVERT_VALUE, FORMAT_SIMPLE and FORMAT_WITH_SPEC.
The compiler can then emit the optimal sequence for each format expression.
The three new opcodes are much simpler than the one they replace, actually removing a few lines of code from the interpreter.
The CONVERT_VALUE is emitted only if a conversion is present.
In the standard library only (approx) 20% of f-strings include a conversion.
FORMAT_WITH_SPEC is emitted if a format specifier is present.
Otherwise FORMAT_SIMPLE is emitted.
In the standard library about 70% of format expressions can be formatted with just the simple (and fast)
FORMAT_SIMPLE opcode.
https://bugs.python.org/issue33092