Changes between Initial Version and Version 1 of Ticket #21693, comment 2


Ignore:
Timestamp:
Mar 10, 2017, 12:56:09 PM (3 years ago)
Author:
asn
Comment:

Legend:

Unmodified
Added
Removed
Modified
  • Ticket #21693, comment 2

    initial v1  
    55Hmm, I think it's because we use the routerparse API (`tokenize_string()` etc.)  to parse each layer, and the routerparse API mainly works on NUL-terminated strings.
    66
    7 Maybe we could play with the `start` and `end` arguments of `tokenize_string()` to tell it where it should stop parsing due to binary data. In that case, I think we would need some sort of separator between the ASCII part of the descriptor layer and the binary part of the descriptor layer.
     7Maybe we could play with the `start` and `end` arguments of `tokenize_string()` to tell it where it should stop parsing due to binary data. On this end, we could use the `encrypted` field of the descriptor as a separator between ASCII and binary data.
     8
     9Of course, this way we lose the ASCIIness of the super-encrypted part of the descriptor; i.e. you can't nicely dump it on your terminal.