This issue tracker has been migrated to GitHub, and is currently read-only.
For more information, see the GitHub FAQs in the Python's Developer Guide.

Author larry
Recipients benjamin.peterson, eric.snow, larry, ncoghlan, serhiy.storchaka, vstinner
Date 2018-01-04.01:31:25
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <1515029486.46.0.467229070634.issue32455@psf.upfronthosting.co.za>
In-reply-to
Content
> 1. This will actually simplify the code for calculating the stack size.

Simplify whose code, the caller or the callee?  It seems like it's simplifying the life of the callee (PyCompile_OpcodeStackEffectEx) at the expense of pushing complexity into the caller.


> 2. This will simplify third-party compilers (like
> https://github.com/vstinner/bytecode). They wouldn't need to
> duplicate the complicated code.

What complicated code?  Compilers simply call PyCompile_OpcodeStackEffect(), passing in an opcode + oparg pair, and it simply returns the maximum stack effect.  No crazy "is it a jump" logic needed.  Again: this may mean overallocating the stack.  But CPython has behaved like this for twenty years and I'm not convinced it's a problem that needs solving.


Victor, do you want to use Serhiy's proposed API in your library?
History
Date User Action Args
2018-01-04 01:31:26larrysetrecipients: + larry, ncoghlan, vstinner, benjamin.peterson, eric.snow, serhiy.storchaka
2018-01-04 01:31:26larrysetmessageid: <1515029486.46.0.467229070634.issue32455@psf.upfronthosting.co.za>
2018-01-04 01:31:26larrylinkissue32455 messages
2018-01-04 01:31:25larrycreate