Make sure that routerlists contain only routers; give a better error message if they start with a nonrouter.


svn:r2393

svn:r2392

fix a memory leak


svn:r2391

svn:r2390

If we are an authoritative dirserver, check out the fingerprint list when setting connection nicknames


svn:r2389

gave us a sigpipe, and we logged that we were ignoring it,
causing us to fail to log that, and delete the log entry. Then
when the signal handler exited, we proceeded to delete the log
entry that had already been deleted.

Now we make sure to only log inside our signal handler if we'll
be exit()ing right after.


svn:r2388

svn:r2387

svn:r2386

svn:r2385

rendezvous publishes and fetches, since they're still running 0.0.8. so
make them speak the old way until we've upgraded the authdirservers.


svn:r2384

still not working


svn:r2383

- don't complain that you won't cache it, then cache it anyway.
- change confusing terminology from 'deflate' to 'compress', so
  we'll stop getting our logic backwards. i assume the rfc author
  has shot himself by now.


svn:r2382

svn:r2381

files have really old descriptors for the authdirservers, so we're
asking them in the new format because they're too old.

now we actually compare the version to a cutoff version, and act
appropriately.

also take this chance to use only >=0.0.8 servers for dns resolves,
because of the recent bugs. we'll bump to >=0.0.9pre1 once there are
some servers running that.


svn:r2380

transition easier


svn:r2379

dirservers rewrite url /foo/ to /tor/foo/, which will help us run things
behind apache once 0.0.8 is obsolete


svn:r2378

svn:r2377

svn:r2376

Make read_all distinguish between error and EOF; read_file_to_string should only check bytes read against st_size when in binary mode.


svn:r2375

thanks to duck for pointing out this omission


svn:r2374

svn:r2373

svn:r2372

keys in /tmp


svn:r2371

wait 10 seconds after directory gets dirty, before regenerating


svn:r2370

than once per minute


svn:r2369

rather than seg fault


svn:r2368

actually unlink the log entry.

ok, that wasn't cleverly hidden enough. let's try again.


svn:r2367

so if there were any problems opening logs, we would complain to
stderr (which wouldn't work) and then mysteriously exit.


svn:r2366

don't crash. just stop logging to that log and continue.

nick, can you find the bug i've cleverly hidden in this commit?


svn:r2365

svn:r2364

raise it to 100 for now. we need a better solution long-term.
but alas, there are no better solutions that are also portable.
hm.


svn:r2363

svn:r2362

svn:r2361

Attach dummy resolve connections to a circuit *before* calling dns_resolve().  This fixes a bug where cached answers would never be sent in RESOLVED cells.


svn:r2360

svn:r2359

the end cell back. also, give a better reason for failure.


svn:r2358

svn:r2357

svn:r2356

Bugfix: When a dns request failed, we would detach any "resolving" streams from the circuit before we would send a the resolve_failed cells.


svn:r2355

svn:r2354