Message340513
I decided to come back to this after a python meetup last night. By messing with this a bit, building in VS2019 with /W4, I see that fully 2/3rds of the total warnings are from two specific warnings:
C4100 (unreferenced formal parameter)
C4127 (conditional expression is constant)
...This seems to be a stylistic thing across the codebase. If it were a new codebase, I'd simply recommend not giving unreferenced formal parameters a variable name - the typical way of doing it - but there's no way anybody is gonna care enough to comment them out across the whole codebase. C4127 is not A Big Deal, since dispatching based on data type sizes in conditionals is just the easiest way to do The Right Thing in C.
The rest of the warnings are mostly datatype coercions ('=': conversion from 'int' to 'char', possible loss of data), old style declarators, and a bunch of type indirection mismatches ('function': 'volatile int *' differs in indirection to slightly different base types from 'volatile long *'), type cast truncation ('type cast': truncation from 'volatile __int64' to 'PyThreadState *'), named type declarations in parenthesis ('timeval': named type definition in parentheses), and assignments in conditionals (which I don't like, but are not a huge deal).
There really are only a few things that actually look sketchy. For example, in run_child (launcher.c), SetInformationJobObject is passing sizeof(info) as cbJobObjectInformationLength, where it should instead be the local variable rc. |
|
Date |
User |
Action |
Args |
2019-04-18 21:01:59 | Alexander Riccio | set | recipients:
+ Alexander Riccio, loewis, paul.moore, tim.golden, skrah, Jim.Jewett, zach.ware, serhiy.storchaka, steve.dower |
2019-04-18 21:01:59 | Alexander Riccio | set | messageid: <1555621319.71.0.560462809846.issue25878@roundup.psfhosted.org> |
2019-04-18 21:01:59 | Alexander Riccio | link | issue25878 messages |
2019-04-18 21:01:59 | Alexander Riccio | create | |
|