A prototype for solve dividing two integers #66
Conversation
for rounding down to zero for integer division
r888800009
changed the title
pyclibrary/c_parser.py
Outdated
| @@ -35,6 +35,91 @@ | |||
| __all__ = ['win_defs', 'CParser'] | |||
|
|
|||
|
|
|||
| def wrap_int(t): | |||
| logger.debug('wrap_int: {} {}'.format(t.dump(), type(t))) | |||
| t[0] = "CInt({})".format(eval(t[0])) | |||
There was a problem hiding this comment.
hello @kalvdans,
Thanks for your reply, avoiding using eval is a very good simple suggestion to avoid security issues and everyone should know.
One solution is to write a C integer parser, but I'm not sure it's worth writing. Because my purpose is to bind existing open source c libraries. If some people who want to use untrusted libraries may be worried about this.
Also, I think must write in clean code for better maintenance, but I don't know of any python built-in function that can convert C integer string to a Python integer. If you know a function, I will appreciate if you tell me :)
See also this code https://github.com/r888800009/pyclibrary/blob/f5f8bf80113b9deb247e92705bb4b73529b95912/pyclibrary/c_parser.py#L1736-L1739
The token passed in by wrapper must match the regular expression. If you know how to bypass this regex to achieve python execute code, please tell me and I will be able to confirm that this is indeed a bad approach
Of course, to avoid misuse, we'd better avoiding using eval, but we also need have a good enough solution to make it easy to maintain.
For your concerns, maybe you'll want to look at the existing code in c_parser.py
pyclibrary/pyclibrary/c_parser.py
Line 1514 in 1d4dbfc
If we care about security issues, maybe we can consider rewriting the eval function of the entire project
Best regards
There was a problem hiding this comment.
I did not check in depth but could you use ast.literal_eval ?
There was a problem hiding this comment.
hmm, int(t[0], 0) should handle 0x prefixes fine, but doesn't work with octal.
pyclibrary/c_parser.py
Outdated
| @@ -35,6 +35,91 @@ | |||
| __all__ = ['win_defs', 'CParser'] | |||
|
|
|||
|
|
|||
| def wrap_int(t): | |||
| logger.debug('wrap_int: {} {}'.format(t.dump(), type(t))) | |||
| t[0] = "CInt({})".format(int(t[0], 0)) | |||
There was a problem hiding this comment.
Why do we return a string from here? I would expect this function to return a CInt instance.
There was a problem hiding this comment.
I realise this should be part of a bigger rewrite, so no need to change in this PR.
pyclibrary/c_parser.py
Outdated
| # The floating regex is ugly but it is because we do not want to match | ||
| # integer to it. | ||
| floating = Regex(r'[+-]?\s*((((\d(\.\d*)?)|(\.\d+))[eE][+-]?\d+)|((\d\.\d*)|(\.\d+)))') | ||
| number = (floating | integer) | ||
| number = (floating |integer) |
There was a problem hiding this comment.
coding style: use spaces around binary operators
| @@ -290,6 +290,17 @@ def test_values(self): | |||
| macros['MACRO_H3'] == '0X000002UL' and | |||
| values['MACRO_H3'] == 2) | |||
|
|
|||
| # Octal integer | |||
| @@ -566,6 +577,24 @@ def test_variables(self): | |||
| assert ('x2' in variables and | |||
| variables['x2'] == (88342528, Type('int'))) | |||
|
|
|||
| # Test int div 9 / 2 should be 4 | |||
| assert ('x3' in variables and | |||
There was a problem hiding this comment.
| assert ('x3' in variables and | |
| assert ( |
the indexing will anyway if the key does not exist in the dict. No need to test it separately.
hello I created a prototype to solve issue #65, but need to do more test cases and evaluate how to improve this code
I created a CInt wrapper, and wrapped the integer in the parsing stage, which can be divided as a c integer in eval.
However, in performance, I have not yet evaluated it, but it has passed the original and new test cases.
Need more discussion, thanks