 Nick Mathewson
|
2ff6cfccd6
Forward-port fix to make dual-purpose NT executable work.
|
20 years ago |
 Roger Dingledine
|
55c3f15ed1
Forward-port the trick to bump up ConnLimit as far as it can
|
20 years ago |
|
Nick Mathewson
|
6e6d95b3db
Change from inet_ntoa to a threadproof tor_inet_ntoa.
|
21 years ago |
|
Nick Mathewson
|
e8da6b26b6
Note a non-bug in read_file_to_str as used. If we start to rely on it behaving differently, we need to change how it behaves. [Noticed by Chris Palmer]
|
21 years ago |
|
Nick Mathewson
|
d37f4dd8a9
But on windows, localtime and gmtime _are_ threadsafe.
|
21 years ago |
|
Nick Mathewson
|
70c3580f81
Patch to localtime/gmtime handling: use the _r variants where available. Use mutexes to fake _r where necessary. Make mutexes no-ops where no threading is enabled.
|
21 years ago |
|
Nick Mathewson
|
d01718841e
Remove assumptions that maximum path length is anything like 1024. From comments by Chris Palmer
|
21 years ago |
|
Nick Mathewson
|
7d9820e80d
Check for pthread dependency on pthread_detach as well as pthread_create
|
21 years ago |
|
Nick Mathewson
|
bb53a0bddd
remove spurious semicolons
|
21 years ago |
|
Nick Mathewson
|
d060f845f2
Snarf some logic from python, adapted to our own needs, to handle gethostbyname_r correctly across platforms.
|
21 years ago |
|
Nick Mathewson
|
a68fcbf807
Initial IRIX compatibility fixes: check for inttypes.h; check for platform cell_t.
|
21 years ago |
|
Nick Mathewson
|
c2d5b67dd2
After we successfully connect, try handling pending stuff immediately, rather than waiting for the next call to connection_handle_write. Roger: is this right?
|
21 years ago |
|
Nick Mathewson
|
e673536e3f
Print uname on start of unittests.
|
21 years ago |
|
Nick Mathewson
|
1a55422006
Fix windows build: juggle }s and #endifs around in main.c
|
21 years ago |
|
Nick Mathewson
|
6bd81c9ee2
Fix windows build: do not spuriously include openssl where not needed; handle lack of gethostbyname_r properly.
|
21 years ago |
|
Roger Dingledine
|
8991af604c
make the addressmap internals less noisy
|
21 years ago |
|
Roger Dingledine
|
6d85627bc9
man page entries for addressmap and trackhostexits
|
21 years ago |
|
Roger Dingledine
|
8727acf2f1
Add a new AddressMap directive to rewrite incoming socks addresses.
|
21 years ago |
|
Roger Dingledine
|
b5bddd8cfd
bugfix: actually use the argument to strmap_free()
|
21 years ago |
|
Roger Dingledine
|
38f53cff41
forward-port the onion queue thing
|
21 years ago |
|
Roger Dingledine
|
596d65ec3b
avoid case (not yet triggered) where smartlists could grow out
|
21 years ago |
|
Nick Mathewson
|
f309168301
Tweak control-spec: add a GETINFO/INFOVALUE type pair to handle non-configuration stuff and to avoid proliferating more GETFOO/FOOVALUE pairs. Remove spurious 0x from signal codes. Simplify address-mapping format. Clarify what an identity looks like.
|
21 years ago |
|
Roger Dingledine
|
b3d39cb689
make it clearer that our screenshots are of old versions.
|
21 years ago |
|
Roger Dingledine
|
cbb4803ec8
forward-port some fixes
|
21 years ago |
 Geoff Goodell
|
d418cd5f70
integrating changes related to building circuits, assigning streams, and exchanging descriptors (discussed on return trip from airport)
|
21 years ago |
|
Roger Dingledine
|
797419a62c
remove unused code
|
21 years ago |
|
Nick Mathewson
|
bfa9a72d5e
More fixes for IRIX compilation: make very sure we have NULL in configure.in. This may work now.
|
21 years ago |
|
Nick Mathewson
|
e6dcb17ba4
Fix check for NULL representation to include string.h; this makes fixes building on Irix, I hope.
|
21 years ago |
|
Roger Dingledine
|
ecf7bb5ff4
make the wikipedia statement less strict
|
21 years ago |
|
Nick Mathewson
|
0289144229
Fix another crash bug found by Jaroslaw--it is possible for dns_resolve to mark a connection to be freed, if the circuit it was attached to fails for some reason.
|
21 years ago |
