summaryrefslogtreecommitdiff
path: root/src
AgeCommit message (Collapse)Author
2004-04-27some compilers don't like an array of length zeroRoger Dingledine
nick: should we remove support for IVs for now, since we don't use them and don't plan to use them? svn:r1732
2004-04-27set Content-Type on the directory and hidserv descriptorRoger Dingledine
svn:r1731
2004-04-27make my assumption explicitRoger Dingledine
svn:r1730
2004-04-27it was the second bug that was nailing us.Roger Dingledine
periodically we expire some circuits if we already have enough, and we were counting rend circs in that number. (the first one wasn't a bug after all) svn:r1729
2004-04-27don't expire joined-rend circs that have an exit connection attachedRoger Dingledine
and don't expire non-general excess circuits svn:r1728
2004-04-26Some versions of openssl have an SSL_pending function that erroneouslyNick Mathewson
returns bytes when there is a non-application record pending. I have no idea when/why this would even happen, but let's catch it and make sure tor_tls_get_pending_bytes stays correct. svn:r1727
2004-04-26Log number of bytes pending after read.Nick Mathewson
svn:r1726
2004-04-26rendmid says it drops but actually sends nackRoger Dingledine
fix log messages svn:r1725
2004-04-26Remove spurious semiNick Mathewson
svn:r1724
2004-04-26Add a macro to catch unhandled openssl errors.Nick Mathewson
svn:r1723
2004-04-26and other debugsRoger Dingledine
svn:r1722
2004-04-26log debug so nick can see it tooRoger Dingledine
svn:r1721
2004-04-26The dots are there so you see test doing something. Therefore we should ↵Peter Palfrader
flush them after printing svn:r1720
2004-04-26add a debugging entry, to find the wants-to-read-but-can't bugRoger Dingledine
svn:r1719
2004-04-26Include strerror(errno) with tls syscall errorsNick Mathewson
svn:r1718
2004-04-26Refactor crypto error handling to be more like TLS error handling:Nick Mathewson
crypto_perror is a no-no, since an operation can set more than one error. Also, fix a bug in the unix crypto_seed_rng: mixing stdio with /dev/urandom is a bad idea, since fopen can make all kinds of weird extraneous syscalls (mmap, fcntl, stat64, etc.) and since fread tends to buffer data in big chunks, thus depleting the entropy pool. svn:r1717
2004-04-26Log pending TLS errors in a couple more places, in case they are possible.Nick Mathewson
svn:r1716
2004-04-26tell us when we're rotating thingsRoger Dingledine
svn:r1712
2004-04-26since we don't support truncateds much, don't bother sending them;Roger Dingledine
just close the circ. (this wasn't relevant before, because we were mis-handling destroys.) svn:r1711
2004-04-26fix a bug that's been lurking since 27 may 03 (!)Roger Dingledine
when passing back a destroy cell, we would use the wrong circ id. how the heck did this work? svn:r1710
2004-04-26Call tls_log_errors at a more appropriate location; we can remove the other ↵Nick Mathewson
calls in tor_tls_verify once we are sure they never happen. svn:r1709
2004-04-26Better error msg on unknown circuit id.Nick Mathewson
svn:r1708
2004-04-26Very blunt debugging code: log pending errors at start and end of tor_tls_verifyNick Mathewson
svn:r1707
2004-04-26tell us the nickname of the OR that hung up on usRoger Dingledine
svn:r1706
2004-04-25Check for machine/limits.h in addition to sys/limits.hNick Mathewson
svn:r1705
2004-04-25don't crash, if a conn that sent a begin has suddenly lost its circuitRoger Dingledine
svn:r1704
2004-04-25resolve a double-mark-for-close when things die inside connection_handle_writeRoger Dingledine
svn:r1702
2004-04-25give us a new dirservers fileRoger Dingledine
svn:r1700
2004-04-25now assume all routers support rendezvous cellsRoger Dingledine
svn:r1699
2004-04-25use tor_assertRoger Dingledine
remove obsolete BUF_OK macro svn:r1697
2004-04-25use tor_assert and PUBLIC_KEY_OKRoger Dingledine
but don't use tor_assert inside log.c, to avoid loops svn:r1696
2004-04-25Add a tor_assert macro that logs failed assertions.Nick Mathewson
svn:r1695
2004-04-25Add assert on crypto_pk_write_private_key,and macros to make sure we have ↵Nick Mathewson
real keys svn:r1694
2004-04-25Remove onion_pkey from connection, since onion keys can change more often ↵Nick Mathewson
than connections. Also add more log messages svn:r1693
2004-04-25remove spurious router_rebuild_descriptorNick Mathewson
svn:r1692
2004-04-25don't build and upload a new desc twice in a rowRoger Dingledine
svn:r1691
2004-04-25log correctly if decoding onion failedRoger Dingledine
svn:r1690
2004-04-25don't warn when private_decrypt_hybrid failsRoger Dingledine
svn:r1689
2004-04-25bugfix: when rotating onion key, do it (more) correctlyRoger Dingledine
svn:r1688
2004-04-24more touch-upsRoger Dingledine
svn:r1687
2004-04-24Spurious link-key should not be an errorNick Mathewson
svn:r1686
2004-04-24Note router addr:port in log messagesNick Mathewson
svn:r1685
2004-04-24Compare nicknames casei.Nick Mathewson
svn:r1684
2004-04-24Merge flagday into main branch.Nick Mathewson
svn:r1683
2004-04-20only expire wedged dir conns that haven't already been expiredRoger Dingledine
svn:r1676
2004-04-19use the real socketpair if we've got itRoger Dingledine
now it's just windows that has to use the kludged one svn:r1675
2004-04-19the socketpair bug is no longer a bugRoger Dingledine
now it is a documented behavior svn:r1674
2004-04-19expire wedged dir conns after 5min without write, not simply after 5minRoger Dingledine
svn:r1673
2004-04-19give wedged dir conns a last chance to flushRoger Dingledine
svn:r1671
2004-04-18tmp patch to also expire old clean rend_ready circs for nowRoger Dingledine
will have to be fixed once we build-and-open rend circs preemptively svn:r1668
7apu2bq3rhoylwmucxizk54afw5jyda7pho4j5zdcqpwkufke7zdzwad.onion