bpo-41463: Generate information about jumps from 'opcode.py' rather than duplicating it in 'compile.c'#21714
Uh oh!
There was an error while loading. Please reload this page.
Conversation
…te it in 'compile.c'
Python/compile.c Outdated
| inti_lineno; | ||
| }; | ||
| staticint |
There was a problem hiding this comment.
| staticint | |
| staticinlineint |
Python/compile.c Outdated
| return (_PyOpcode_RelativeJump[opcode>>5]>>(opcode&31))&1; | ||
| } | ||
| staticint |
There was a problem hiding this comment.
| staticint | |
| staticinlineint |
Python/compile.c Outdated
| is_relative_jump(structinstr*i) | ||
| { | ||
| intopcode=i->i_opcode; | ||
| return (_PyOpcode_RelativeJump[opcode>>5]>>(opcode&31))&1; |
There was a problem hiding this comment.
Can you define 5 and 31 as globals with a name (using #define or with static variables). That way this will read a bit better
pablogsal left a comment
pablogsal left a comment There was a problem hiding this comment.
Look great, I left some minor comments
| intoparg=inst->i_oparg; | ||
| intnextop=i+1<bb->b_iused ? bb->b_instr[i+1].i_opcode : 0; | ||
| if (inst->i_jabs||inst->i_jrel){ | ||
| if (is_jump(inst)){ |
Python/compile.c Outdated
| staticinlineint | ||
| is_relative_jump(structinstr*i) | ||
| { | ||
| intopcode=i->i_opcode; |
There was a problem hiding this comment.
Apologies for insisting a bit but it still took me a minute to parse the expression to follow how the function is determining if is a relative jump. I would suggest to maybe add a comment here briefly explaining the process as someone with less context will certainly struggle
pablogsal left a comment
pablogsal left a comment There was a problem hiding this comment.
LGTM
I left a comment, feel free to address it before landing
…han duplicating it in 'compile.c' (pythonGH-21714) Generate information about jumps from 'opcode.py' rather than duplicate it in 'compile.c'
…han duplicating it in 'compile.c' (pythonGH-21714) Generate information about jumps from 'opcode.py' rather than duplicate it in 'compile.c'
4 participants
The motivation for this PR is to avoid errors when modifying compile.c.
Setting the
i_jabsandi_jrelfields of every instruction is error prone and unnecessary.https://bugs.python.org/issue41463