exclude @PaxHeader from archive when exctracting #7172
Conversation
| @@ -55,7 +55,7 @@ def decompress_file(archive, dir_): | |||
| if len(suffixes) > 1 and suffixes[-2] == '.tar': | |||
| # we have a compressed tar file that needs to be fed through the | |||
| # decompressor first | |||
| cmd = '7z x {} -so | 7z x -si -ttar'.format(quote_cmdlinearg(archive)) | |||
| cmd = '7z x {} -so | 7z x "-x.\.\@PaxHeader" -si -ttar'.format(quote_cmdlinearg(archive)) | |||
There was a problem hiding this comment.
I'm under the impression that -x.\.\@PaxHeader refers to filenames on windows. I think the \ should be os.path.separator instead.
There was a problem hiding this comment.
@bpinsard Are you planning to finish this or should I take it? I think above mentioned change is all it takes.
There was a problem hiding this comment.
Weirdly, I managed to make it work once, but now I can't get it to work. Not sure this is the right solution for now. I think I will be migrating to something else that tar.gz, to allow "random" access similarly to what is done with the ORA 7z archives.
There was a problem hiding this comment.
In fact maybe the second 7z should be replaced with tar itself to be able to handle these metadatas. Unless there is something I am missing that requires 7z or is it to reduce number of dependencies.
There was a problem hiding this comment.
- PR might be worth to be against
maint - comment better be placed to note why that file is to be excluded etc.
There was a problem hiding this comment.
or is it to reduce number of dependencies.
IIRC that was the point, yes. However, it seems to me it's fine to depend on tar for a dataset that uses tarballs.
PR checklist
Fixes #7171 by excluding @PaxHeader from extracted files
Thanks for contributing!