666c59cbfb
actually fix merge 🙈
2016-10-16 21:33:24 -07:00
bdaa13d498
fix merge
2016-10-16 21:23:04 -07:00
9389601025
Merge branch 'such-types'
2016-10-16 21:17:39 -07:00
ae3ff8ee1e
Merge pull request #1615 from cortesi/python3a
...
exterminate six
2016-10-17 17:03:02 +13:00
5a07892bfc
py2--: inline type info
2016-10-16 20:56:46 -07:00
ce98a9219e
test & examples: zap six
2016-10-17 16:45:45 +13:00
3fbce7e981
update pydivert dependency ( #1614 )
2016-10-16 20:30:28 -07:00
839813a84c
mitmproxy: zap six
2016-10-17 16:22:44 +13:00
d60ef617e3
pathod: remove six
2016-10-17 15:43:38 +13:00
421679a770
Merge pull request #1613 from cortesi/python3
...
First-order conversion to Python3-only
2016-10-17 15:35:09 +13:00
8360f70024
First-order conversion to Python3-only
...
- Zap various occurrences of Python2 in docs and scripts
- Remove six from netlib, and some other places where obvious project-wide
search and replace works.
2016-10-17 15:18:47 +13:00
4918feb725
Add stub converter for 0.19 io format
2016-10-17 08:47:03 +13:00
dcbb968b1b
Bump version post release
2016-10-17 00:29:49 +13:00
5db860a334
Bump to 0.18.1 to fix pypi obstreperousness
...
Also include contriburors update.
2016-10-16 23:52:47 +13:00
e4a32d6844
Merge pull request #1602 from cortesi/scriptdocs
...
docs: stub out new script documentation
2016-10-16 20:50:09 +13:00
00603021d9
docs: concurrency, developing scripts
2016-10-16 20:39:34 +13:00
9a0195bf64
scripts: keep scripts just after the ScriptLoader in addon chain
...
We need scripts to run _before_ filestreamer, so we can't just add them to the
end of the chain. This patch also fixes an issue that could cause scripts to be
initialised un-necessarily if only the order of scripts in options changed.
2016-10-16 20:26:06 +13:00
57b8ed21a9
docs: scripts on saved flows
2016-10-16 20:26:06 +13:00
97b594b848
mitmdump: fix addon order - dumper must be last
...
This is so we can see the effects of script rewriting using -dd.
2016-10-16 20:26:06 +13:00
072fff90f1
docs: link to correct tag
2016-10-16 20:26:06 +13:00
55cb2a8547
docs: logging and the context
2016-10-16 20:26:06 +13:00
a6c7a1ff91
scripts: handle SystemExit from scripts explicitly
2016-10-16 20:26:06 +13:00
dc19ff7a9d
mitmdump: show script errors on startup
...
Add the terminal logger before any other addons, so we see script errors on
startup.
2016-10-16 20:26:06 +13:00
fb69c9c345
docs: overview, classes, arguments
2016-10-16 20:26:06 +13:00
61040a7bcd
docs: improve external source links, tweak code docs
2016-10-16 20:26:06 +13:00
c8f2f1019d
Clean up models.http a bit
...
- We don't need a deprecation warning here
- Bring imports inline with policy
2016-10-16 20:26:06 +13:00
fdb6a44245
docs: cleanups improvements and fighting sphinx
...
- Hide links to internal code listings, and link to github instead
- Improve formatting of code/example captions
- Fix outdated documentation of command-line options
- Complete documentation of all events + improved formatting
- tcp_open -> tcp_start, tcp_close -> tcp_end to reduce confusion
2016-10-16 20:26:06 +13:00
26af9b29fc
fix #1611
2016-10-15 18:25:18 -07:00
8c3eed16d4
update CHANGELOG
2016-10-15 18:15:28 -07:00
d5873302c1
Merge pull request #1610 from mhils/issue-1605
...
Fix #1605
2016-10-14 02:00:36 -07:00
e9284ab7ee
add regression test for #1605
2016-10-14 01:22:30 -07:00
5fab94580e
fix #1605
2016-10-13 20:31:14 -07:00
61a1b96ca4
Issue 1461 ( #1609 )
...
* doc changes
* doc changes
* doc changes
* doc changes
* Documentation updated
* Documentation for custom certificate updated
2016-10-13 19:46:12 -07:00
8be0d78eaa
Merge pull request #1603 from mhils/eof
...
Simplify EOF detection
2016-10-05 19:37:22 -07:00
b422ba5328
simplify eof detection
2016-10-05 17:45:48 -07:00
8e7ec6117a
Merge pull request #1601 from cortesi/certcap
...
certutils: cap the cert store size at 100 by default
2016-10-05 11:32:05 +11:00
89d36713e3
certutils: cap the cert store size at 100 by default
...
This should be enough to give us reuse without growing infinitely. This is part
of fixing the memory situation in mitmdump.
TODO: There's an opportunity here for a better algorithm, that expires certs
based on least-recently-accessed time, rather than oldest generated time.
2016-10-05 10:44:31 +11:00
6d343c7ca3
debug: use correct fp for memory object output
2016-10-05 09:54:23 +11:00
ff388f2e62
Merge pull request #1598 from cortesi/dummystate
...
mitmdump: use a dummy state object
2016-10-04 15:36:31 +02:00
90898f9084
use flt instead of filt for flowfilter expressions
2016-10-04 14:51:04 +02:00
76ae9fdbaa
Merge pull request #1597 from cortesi/debug
...
debug: add a dump of objects in memory
2016-10-04 17:00:44 +11:00
e189652680
mitmdump: use a dummy state object
...
Most of the mitmdump memory leak turns out to be due to our leaky handling of
the state object. Since mitmdump doesn't actually use the state object, we can
replace it with a shell.
2016-10-04 16:52:58 +11:00
882c363ec5
debug: Hard exit after debug signal
2016-10-04 16:48:13 +11:00
2130f248de
Merge pull request #1596 from cortesi/addons
...
Consolidate logging
2016-10-04 16:30:25 +11:00
3318e531a6
debug: add a dump of objects in memory
2016-10-04 16:22:01 +11:00
7fe09f4d26
Consolidate logging
...
- Make log event handler work
- Move terminal event log into an addon
- Clean various log related errors and duplications
2016-10-04 13:00:15 +11:00
3d5b811994
Merge pull request #1592 from cortesi/ws
...
docs and API-related cleanups
2016-10-04 10:54:15 +11:00
2bd8686629
http2: support the requestheaders event
...
We do this by splitting read_request into read_request_headers and
read_request_body.
2016-10-04 10:33:57 +11:00
2e48b60ff1
Clarify channel interaction pattern
...
Before this patch, it wasn't clear when the arguments to an event could be
modified, and when a new object should be returned. After this patch, the right
thing to do is to modify the arguments in 99% of cases. The only exception to
this is root_layer, for which we can find a better structure down the track.
2016-10-04 08:13:56 +11:00
8cec187f9a
http2: frames don't have a human_readable method
2016-10-04 08:13:56 +11:00
Maximilian Hils