This is the mail archive of the binutils@sourceware.org mailing list for the binutils project.
| Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
|---|---|---|
| Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |
| Other format: | [Raw text] | |
On 2019-10-29, Nick Clifton wrote:
There is a proposal https://reviews.llvm.org/D69418 that adds --output to llvm-ar. --output specifies the output directory for the x (extract) operation.I have started working on an implementation of this feature for the binutils version of ar. I have run across one issue however, which I would like to run by you: https://sourceware.org/bugzilla/show_bug.cgi?id=17533#c7 Currently the binutils version of ar checks that files are extracted into the current directory, or into sub-directories of the current directory, but not elsewhere in the file system. Do you think that the code handling the --output option should just waive these checks, on the assumption that the if the user wants to extract the files into another place then they should know what they are doing, or issue a warning if the resolved path would be outside of the current directory, or refuse to create such files ?
I think that the --output option should keep the is_valid_archive_path check. 1) absolute paths should still be rejected with --output 2) a path component .. can still cause directory traversal problems with --output, so it should be rejected as well. I sent a patch a few days ago: https://sourceware.org/ml/binutils/2019-10/msg00193.html (both absolute paths and paths containing .. are rejected) but it may be better for a maintainer to implement it because I know little about the ar.c code and may miss something...
| Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
|---|---|---|
| Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |