Commit Graph

3336 Commits

Author SHA1 Message Date
Nick Mathewson
a9e1a4c33c Ah. We need a new zlib_state for each networkstatus we spool out.
svn:r6667
2006-06-20 06:27:13 +00:00
Nick Mathewson
4283b81a48 Start spooling v2 networkstatus docs as well.
svn:r6664
2006-06-20 00:48:23 +00:00
Nick Mathewson
3f7f069a05 Fix a tricky crash: making the_directory heap-allocated (so we could refcount it and have multiple instances as neeeded) means that calls to dirserv_regenerate_directory could invalidate the auth_dir value passed to dirserv_pick_cached_dir_obj. Big fun.
svn:r6663
2006-06-18 21:30:03 +00:00
Nick Mathewson
2552d5e60f Fix an assert that still isnt the assert we are hunting.
svn:r6662
2006-06-18 21:15:01 +00:00
Roger Dingledine
37c441d96f correct a function comment in compute_preferred_testing_list()
svn:r6660
2006-06-18 21:07:45 +00:00
Nick Mathewson
f1993b8b30 Become capable of noticing that we are done sending a directory.
svn:r6659
2006-06-18 20:58:27 +00:00
Nick Mathewson
5c0d7fb016 Resolve control flow warning.
svn:r6658
2006-06-18 20:39:46 +00:00
Nick Mathewson
0932f463ff Ah. That seems to work.
svn:r6657
2006-06-18 16:39:26 +00:00
Nick Mathewson
5ba72d0ae5 another write_to_buf_zlib fix.
svn:r6656
2006-06-18 16:20:38 +00:00
Nick Mathewson
b5c599412f Backport candidate: implement the "is this uptime change cosmetic" test properly.
svn:r6654
2006-06-18 16:05:54 +00:00
Nick Mathewson
a9cffd2ff0 Re-enable buffer RAM guard values.
svn:r6653
2006-06-18 15:53:54 +00:00
Nick Mathewson
11c2e97d51 Another _zlib fix.
svn:r6652
2006-06-18 09:03:48 +00:00
Nick Mathewson
337ad6420b Make connection_write_to_buf_zlib set outbuf_flushlen right.
svn:r6651
2006-06-18 08:53:09 +00:00
Nick Mathewson
537104c908 write_to_buf != connection_write_to_buf. Also, add a connection_write_to_buf_zlib wrapper that sucks.
svn:r6650
2006-06-18 08:46:55 +00:00
Nick Mathewson
10b7b57616 Temporarily disable sentinels on buffers
svn:r6649
2006-06-18 08:44:34 +00:00
Nick Mathewson
9376d901ed make zlib buffer function set buf_highwater properly
svn:r6648
2006-06-18 08:21:27 +00:00
Nick Mathewson
4d02ccfed8 Stop trying to refresh when we are out of data.
svn:r6647
2006-06-18 08:19:35 +00:00
Nick Mathewson
b9013cbe98 And actually check the url when it exists. that might work better.
svn:r6646
2006-06-18 08:16:05 +00:00
Nick Mathewson
c374340c70 That dir_refresh_src fix will only work if I enable it.
svn:r6645
2006-06-18 08:13:45 +00:00
Nick Mathewson
347bcec84a Fix a couple of bugs in last patch.
svn:r6644
2006-06-18 08:07:16 +00:00
Nick Mathewson
c58cc8c16c add coverage for a default case
svn:r6643
2006-06-18 07:57:47 +00:00
Nick Mathewson
b1a8b7869a Oops. conn->requested_resource is client only.
svn:r6642
2006-06-18 07:55:04 +00:00
Nick Mathewson
1d9923da7c Instead of adding servers and v1 directories to buffers en masse, directory servers add them on the fly as their outbufs are depleted. This will save ram on busy dirservers.
svn:r6641
2006-06-18 07:38:55 +00:00
Nick Mathewson
8fa8616ec7 Add tests for several of the more recently committed functions.
svn:r6639
2006-06-18 07:35:10 +00:00
Nick Mathewson
cd38511a1e Part of incremental encryption logic for buffers: there is a subtle yucky point documented in a comment.
svn:r6637
2006-06-18 07:27:47 +00:00
Peter Palfrader
2ff24dbc26 And a minor bugfix to the weaselhack
svn:r6629
2006-06-15 23:14:01 +00:00
Roger Dingledine
af0e47c921 fix spacing
svn:r6628
2006-06-15 22:59:07 +00:00
Peter Palfrader
bc64becd12 Add a /tor/dir-all-weaselhack directory resource so I do not have to update my scripts
svn:r6627
2006-06-15 22:52:56 +00:00
Roger Dingledine
4e4cc9aa27 lower the number of seconds before we yell about clock jump.
and make the yelling only happen if you're a server.


svn:r6625
2006-06-15 09:03:15 +00:00
Roger Dingledine
de7a9c83d8 upgrade the severity of the 'clock jump' warn, and ask people
to report if it occurs.


svn:r6623
2006-06-14 22:28:16 +00:00
Roger Dingledine
09dfe1b265 bugfix: discourage picking directory authorities as our TestVia
hops, even if they're running the right versions, since we probably
already have a connection established to them.


svn:r6622
2006-06-14 11:06:43 +00:00
Nick Mathewson
66db3afb74 eventdns: Apply a couple of patches from AGL; start working on windows compat; note some TODOs.
svn:r6621
2006-06-13 21:49:56 +00:00
Roger Dingledine
977574e48d back off and add the $ at the beginning of the preferrednodes list
we generate.


svn:r6620
2006-06-13 12:57:19 +00:00
Roger Dingledine
2cd349903c be more lax about recognizing valid hexdigests.
svn:r6619
2006-06-13 12:05:59 +00:00
Roger Dingledine
b4e415ba3c now we can tell dirserv_dump_directory_to_string() whether we want
it to include down/invalid descriptors or not.


svn:r6618
2006-06-13 11:11:19 +00:00
Roger Dingledine
88dadc917c bugfix in exit_policy_is_general_exit() that weasel found.
this time for sure!


svn:r6617
2006-06-13 10:48:26 +00:00
Roger Dingledine
23e2c6a7dd first cut at a workaround for the reachability bug: explicitly find a
server running the right version, if we can, and ask for that one.


svn:r6616
2006-06-13 10:25:22 +00:00
Roger Dingledine
31b51849b5 export the default exit policy via the control port, so controllers
don't need to guess what it is / will be later.


svn:r6615
2006-06-13 09:16:09 +00:00
Roger Dingledine
279e86f3c8 Defense in depth: fix the reachability bug a second way too.
Now if we establish a connection with the right digest, regardless
of what the addr/port is, and we have pending create cells, use it.


svn:r6614
2006-06-13 05:51:28 +00:00
Roger Dingledine
a20a384b09 harmless typo
svn:r6613
2006-06-13 05:50:24 +00:00
Roger Dingledine
ab61c40661 Fix the bug that was causing servers to not find themselves
reachable if they changed IP addresses. This happened because
middle servers knew the old descriptor, and kept swapping the
addr/port we asked for with the one they thought was right.

So the create cell never got sent, because it was asking for a
different addr/port than we believed we had connected to.


svn:r6610
2006-06-13 05:36:35 +00:00
Roger Dingledine
19408cf8d7 reintroduce the logic to exit_policy_is_general_exit() to count
how many ports are allowd. require two ports open, not just one.


svn:r6609
2006-06-13 02:48:06 +00:00
Peter Palfrader
cce8ae1190 Make the Exit tag in status documents actually work in head too
svn:r6608
2006-06-12 22:03:25 +00:00
Roger Dingledine
87e3800dfc Finally solve the "closing wedged cpuworkers" bug. Woo.
This happened when we got two create cells in a row from the same TLS
connection. It would hand one to the cpuworker, and then immediately
handle the second one -- after it had registered that the first one was
busy, but before it had updated the timestamp that we use to decide how
*long* it's been busy.


svn:r6606
2006-06-12 11:59:19 +00:00
Roger Dingledine
cbc4cd93ac allow people to start their tor with runasdaemon set but
with no logs set at all.


svn:r6604
2006-06-12 06:03:15 +00:00
Roger Dingledine
14687fc443 take out the reachability bandaid in 0.1.2.x as well.
maybe we will actually be able to fix it, instead.


svn:r6585
2006-06-10 00:57:12 +00:00
Roger Dingledine
889b8d5bac Add a new config option TestVia, that lets you specify preferred middle
hops to use for testing circuits. Perhaps this will let me debug the
reachability problem better.


svn:r6581
2006-06-10 00:26:39 +00:00
Roger Dingledine
6531a31aad actually, don't fix it that far. we should still do some
error checking.


svn:r6576
2006-06-09 09:07:59 +00:00
Roger Dingledine
0071d67efd when only one router is labelled as a guard, and we've
already picked him, we would cycle endlessly picking him
again, being unhappy about it, and so forth.

now we specifically exclude guards when picking a new guard.


svn:r6575
2006-06-09 09:02:32 +00:00
Roger Dingledine
30f6c57fd7 and forward-port too.
svn:r6573
2006-06-09 06:52:49 +00:00