tree d4b8f9ea9da632806260ffd92135c8f314028fe2
parent 012050e3fc92ef5068b7415f15751e0083263dee
author Ani Betts <anais@anaisbetts.org> 1645219488 -0500
committer GitHub <noreply@github.com> 1645219488 +0100
gpgsig -----BEGIN PGP SIGNATURE-----
 
 wsBcBAABCAAQBQJiEA6gCRBK7hj4Ov3rIwAA6TcIAH3Wz4Ny0lt25vFBsd/4IVls
 MzL3FgEMuuu6GdOB4wWJ343og1F9kAnCdgLIgMFuHObgYfq1y7nBy2ReKtvOA9Hr
 JLa4j93XRvYc9lx2RrOGVnLvdy13ZU4sVCPYXa7+8Lp7WQE6vZ/mPR/PgQlHRBuZ
 BdIykOs+uM9x1SGxrlAnSEAkPOX25UqErAQmODBF9Q5e/AhS4cWOMpjy9YAl4vxl
 z9dL78KCrrInTBFgeEqAxkzMGo9yyZmlMn+6JEqcGqfrzJ9NIyIYIAdjzNEeGoBB
 uFa9db4Fs4CZHN8rKery1TP39oJL1oMXLY3U5eVBpiCYRy7AwQZtJ17hAUAUq34=
 =Qseq
 -----END PGP SIGNATURE-----
 

Decode Tar archives with malformed UTF-8 data in headers (#19)

* Decode Tar archives with malformed UTF-8 data in headers

Some Tar archives suck and leave malformed data in their headers, such as
the [Arch Linux bootstrap
image](http://mirror.rackspace.com/archlinux/iso/2022.02.01/archlinux-bootstrap-2022.02.01-x86_64.tar.gz).

To make these decodable via tar, we need to not throw when handling these
entries

* Only decode the value if necessary

If we know that we're going to discard the key, we can safely ignore
decoding the value altogether

* Still check that keys are valid

Reject tar files with malformed keys.

Co-authored-by: Simon Binder <oss@simonbinder.eu>