On Sun, Dec 02, 2007 at 08:59:56PM -0800, Russ Allbery wrote: > It was closer, but there were stray x89 and x94 strings (as literal > strings, not as escapes or hex encodings of characters). Ah, well, that makes some sense. That's what gpg's dumping to me: $ gpg --with-colons --list-key 65B4B162 | hd ... 00000040 2d 33 31 3a 3a 3a 2d 3a 41 75 72 c3 a9 6c 69 65 |-31:::-:Aur..lie| 00000050 6e 20 47 c3 5c 78 38 39 52 c3 5c 78 39 34 4d 45 |n G.\x89R.\x94ME| 00000060 20 3c 61 67 40 72 6f 78 6f 72 2e 63 78 3e 3a 3a | <ag@roxor.cx>::| ... except that I'm sending '\x89' to psql instead of '\\x89'. Doing the latter makes my output include the literal text "\x89", and makes psql include the same thing. That at least stops it repeatedly reporting the name. And making the script de-escape '\\x89' into the actual chacter '\x89' makes it work properly. Yay! Cheers, aj
Attachment:
signature.asc
Description: Digital signature