Move strconv and strings packages' taint-tracking to stdlib, and expand them#318
Move strconv and strings packages' taint-tracking to stdlib, and expand them#318gagliardetto wants to merge 5 commits intogithub:mainfrom
strconv and strings packages' taint-tracking to stdlib, and expand them#318Conversation
|
Should I use a regex for function name matching for similar functions, or it does not make any difference from a performance standpoint? |
|
Personally I like that the particular functions are enumerated rather than using a glob that expresses intent but may hit things it doesn't intend to. Looks like the second commit, while described as just a move, actually adds support for at least NewReader, causing the test failures seen; suggest changing its commit message to fully describe what it's doing. Finally I'm surprised to see this doesn't change StdlibTaintFlow.expected -- is that expected? |
123034e to
afbbe64
Compare
…rconv and move into stdlib
…xpected: calls to strings.NewReader are a step now.
afbbe64 to
35f6f3d
Compare
FFR: see #317 (comment) |
strconv and strings packages' taint-tracking to stdlibstrconv and strings packages' taint-tracking to stdlib, and expand them
max-schaefer
left a comment
There was a problem hiding this comment.
As for the other PR, LGTM in general, but I'd prefer to drop the taint steps that track individual bytes or runes.
|
@max-schaefer I think the rune comment is applied; I'm going to make a derived PR that introduces the sanitiser for |
|
Rebased and derived here: #330 |
|
Superseded by #330. |
Part of #167
codeboxcommands: