Author pablogsal
Recipients Mark.Shannon, methane, nascheme, pablogsal, vstinner, yselivanov
Date 2020-10-22.03:24:42
SpamBayes Score -1.0
Marked as misclassified Yes
Message-id <1603337082.78.0.53398842019.issue42115@roundup.psfhosted.org>
In-reply-to
Content
To clarify what I mean with:

> - We could also do the same for operations like "some_container[]" if the container is some builtin. We can substitute/specialize the opcode for someone that directly uses built-in operations instead of the generic BINARY_SUBSCR.

If a given function has a BINARY_SUBSCR opcode and when executing it a given number of times we see that the object to get the subscript is always a list, we change the BINARY_SUBSCR to BINARY_SUBSCR_FOR_LISTS and it that opcode we do a quick check for PyList_CheckExact and if is correct we call "Objects/listobject.c:list_subscript" directly and if is false we put back a BINARY_SUBSCR.
History
Date User Action Args
2020-10-22 03:24:42pablogsalsetrecipients: + pablogsal, nascheme, vstinner, methane, Mark.Shannon, yselivanov
2020-10-22 03:24:42pablogsalsetmessageid: <1603337082.78.0.53398842019.issue42115@roundup.psfhosted.org>
2020-10-22 03:24:42pablogsallinkissue42115 messages
2020-10-22 03:24:42pablogsalcreate