René's URL Explorer Experiment

Title: bpo-9566: Fix some Windows x64 compiler warnings by segevfiner · Pull Request #2492 · python/cpython · GitHub

Open Graph Title: bpo-9566: Fix some Windows x64 compiler warnings by segevfiner · Pull Request #2492 · python/cpython

X Title: bpo-9566: Fix some Windows x64 compiler warnings by segevfiner · Pull Request #2492 · python/cpython

Description: I tried to fix the x64 compiler warnings on Windows. Here is a list of the ones I encountered: python-x64-warnings.txt The following remain: LINK : warning LNK4013: image size 0x29B000 exceeds specified maximum 0x200000 Caused by externals\tcl-core-8.6.6.0\win\coffbase.txt not giving enough room for an x64 debug build. – "Fixed" by pre-building tcl in bpo-30916 (#2688) ..\Modules\gcmodule.c(1085): warning C4244: 'function': conversion from 'Py_ssize_t' to 'int', possible loss of data Caused by DTrace stubs being defined as taking int. It might be correct to change them to take a larger type since it's possible the real ones take something bigger, but I don't know what the real ones take... EDIT: The real one takes a long (Include/pydtrace.d:12), which is probably 64-bit on architectures where DTrace is available but it is 32-bit on Windows x64. So just changing the stub to the correct type won't help (Though I think we can simply change the stub to a larger integer type without it causing issues...) It might be possible to redefine the real probe to take a larger integral type (It really needs Py_ssize_t but I'm not sure we can use it there. But I'm not familiar with enough... – #2852 ..\Modules\_tracemalloc.c(88): warning C4359: '': Alignment specifier is less than actual alignment (8), and will be ignored. This optimization doesn't really work as is. Need to use #pragma pack instead. – bpo-31018 (#2848) ..\Python\getargs.c(2058): warning C4244: '=': conversion from 'Py_ssize_t' to 'int', possible loss of data I'm not sure about this one. – #2890 peephole.c warnings They seem to be related to mixing opcode args which are unsigned int with size_t/Py_ssize_t. ..\Modules\_ssl.c(2947): warning C4244: '=': conversion from 'Py_ssize_t' to 'int', possible loss of data The Py_buffer can theoretically be larger... But I'm not sure that the protocol even supports that... Should this be casted away or fixed some other way?? – Split to #2495 ..\Modules\_ssl.c(4158): warning C4244: 'function': conversion from 'Py_ssize_t' to 'int', possible loss of data Someone can theoretically pass a buffer big enough... But I doubt such a buffer is going to be written at once... Should this be casted away...? – Split to #2495 cc @zooba https://bugs.python.org/issue9566

Open Graph Description: I tried to fix the x64 compiler warnings on Windows. Here is a list of the ones I encountered: python-x64-warnings.txt The following remain: LINK : warning LNK4013: image size 0x29B000 exceeds sp...

X Description: I tried to fix the x64 compiler warnings on Windows. Here is a list of the ones I encountered: python-x64-warnings.txt The following remain: LINK : warning LNK4013: image size 0x29B000 exceeds sp...

Opengraph URL: https://github.com/python/cpython/pull/2492

X: @github

direct link

Domain: github.com

Links:

Viewport: width=device-width