Hi Nick,
as the subject says, two quick questions about dir-spec.txt:
1. The "s" lines of network status entries are specified as:
"s" SP Flags NL
Now, when a relay has no flags at all, which can happen in a vote or could happen in older consensus methods, that relay would have an "s" line just consisting of "s", not "s ".
I think it makes sense to omit the SP, but should this be specified? If so, how?
There are other lines where it's not specified whether a trailing SP is permitted or not. But I think I only encountered "s" lines without any list elements so far.
(I ran into problems here a while ago, when I looked for lines starting with "s " to overwrite an sLine string variable. But there was no "s " line for some relays, so the program used the "s " line from the previous relay. Went unnoticed. Ugh.)
2. The "m" lines in votes are not in dir-spec.txt. I haven't looked at proposal 158 yet. Is that proposal up-to-date?
Thanks, Karsten
On Wed, Dec 14, 2011 at 1:37 PM, Karsten Loesing karsten.loesing@gmx.net wrote:
Hi Nick,
as the subject says, two quick questions about dir-spec.txt:
- The "s" lines of network status entries are specified as:
"s" SP Flags NL
Now, when a relay has no flags at all, which can happen in a vote or could happen in older consensus methods, that relay would have an "s" line just consisting of "s", not "s ".
I think it makes sense to omit the SP, but should this be specified?
Yes.
If so, how?
One of the easiest ways would be "s" NL | "s" SP Flags NL
There are other lines where it's not specified whether a trailing SP is permitted or not. But I think I only encountered "s" lines without any list elements so far.
IIRC trailing space is permitted while parsing but should never be required or emitted.
(I ran into problems here a while ago, when I looked for lines starting with "s " to overwrite an sLine string variable. But there was no "s " line for some relays, so the program used the "s " line from the previous relay. Went unnoticed. Ugh.)
- The "m" lines in votes are not in dir-spec.txt. I haven't looked at
proposal 158 yet. Is that proposal up-to-date?
It should be, IIRC, but see also 162. I have a ticket somewhere to merge it back into dir-spec