 Roger Dingledine
|
419a672222
holding until flush was borked
|
21 years ago |
|
Roger Dingledine
|
6022bfea11
better debugging for lonely flushes
|
21 years ago |
|
Roger Dingledine
|
d8e5f3a3a2
close immediately when write() fails too, else things will turn bad
|
21 years ago |
|
Roger Dingledine
|
dbdd2eaa2a
don't mark_for_close again when the destroy arrives after the end cell
|
21 years ago |
|
Roger Dingledine
|
3c71d09402
use conn->hold_open_until_flushed for streams
|
21 years ago |
 Nick Mathewson
|
f4e4dac801
Implement hold_open_until_flushed. I may have missed something important.
|
21 years ago |
|
Roger Dingledine
|
8275e2302c
make socks5 not give a spurious warning
|
21 years ago |
|
Roger Dingledine
|
77bb3e054e
reenable part of assert_circuit_ok, leave some still disabled
|
21 years ago |
|
Roger Dingledine
|
de146de0c5
make circuit-building less noisy when a node in your exclude/exit/entrynodes is down
|
21 years ago |
|
Nick Mathewson
|
20ecb66c7d
Fix (maybe) assert_circuit_ok. leave it disabled till arma confirms
|
21 years ago |
|
Nick Mathewson
|
5a053a6a81
remove redundant check
|
21 years ago |
|
Nick Mathewson
|
6379dbbe2a
warn more when losing data on closed/marked connections.
|
21 years ago |
|
Roger Dingledine
|
49281424ae
bugfix: close_immediate called buf_clear but listeners don't have bufs
|
21 years ago |
|
Nick Mathewson
|
5985aef18a
Nth time is the charm on list manip.
|
21 years ago |
|
Roger Dingledine
|
d16f142fa9
fix the rpm spec description again
|
21 years ago |
|
Nick Mathewson
|
a1503f667e
Integrate jbash's RPM spec into build process. (Requires "rpmbuild" to
|
21 years ago |
|
Roger Dingledine
|
4f6cf727b9
fix a pair of lurking segfaults
|
21 years ago |
|
Nick Mathewson
|
6f3c6d1289
Split out mark_for_close with circuits. Seems to work for me.
|
21 years ago |
|
Nick Mathewson
|
a1e4c6dc85
Fix assertion failure spotted by arma.
|
21 years ago |
|
Roger Dingledine
|
b961732246
add --version, note that --help is probably obsolete
|
21 years ago |
|
Roger Dingledine
|
1147289dc0
make config_assign fail rather than warn for unrecognized config options
|
21 years ago |
|
Roger Dingledine
|
08346f13f0
make parse_line_from_file fail rather than warn for malformed lines
|
21 years ago |
|
Roger Dingledine
|
f81ad1550e
note an assert trigger bug
|
21 years ago |
|
Roger Dingledine
|
c0b5002d2c
update the todo
|
21 years ago |
|
Nick Mathewson
|
5072488a95
make 'make test' exit(1) if a test fails.
|
21 years ago |
|
Nick Mathewson
|
3138941e68
revise spec. It still needs more explanation to be intelligible by the uninitiated, but it is at least far less incorrect.
|
21 years ago |
|
Roger Dingledine
|
2c341cd032
put switch_id and start_daemon earlier
|
21 years ago |
|
Roger Dingledine
|
e2881631cf
fix duplicate call to connection_mark_for_close
|
21 years ago |
|
Roger Dingledine
|
2ac37c64f2
bugfix: we were checking the wrong thing to see if 'reason'
|
21 years ago |
 Peter Palfrader
|
247fe31af2
New upstream release.
|
21 years ago |
