Aldo Cortesi
9bea616441
Be stricter about handling connetcts in the pathoc test suite
...
Attempt to clear a niggling Appveyor exception buggering up our CI.
2016-06-12 10:39:19 +12:00
Aldo Cortesi
7b86560ded
debug: On SIGUSR2, we dump tracebacks for all threads to screen
2016-06-11 23:07:42 +12:00
Aldo Cortesi
09edbd9492
Improve debugging of thread and other leaks
...
- Add basethread.BaseThread that all threads outside of test suites should use
- Add a signal handler to mitmproxy, mitmdump and mitmweb that dumps resource
information to screen when SIGUSR1 is received.
- Improve thread naming throughout to make thread dumps understandable
2016-06-11 19:52:24 +12:00
Aldo Cortesi
5b9f07c81c
debug.sysinfo: tests and coverage
2016-06-11 17:56:17 +12:00
Aldo Cortesi
2cf79b7912
Remove timestamps from pathoc output
...
Pathoc is an interactive tool, no need for a long leading timestamp. More
generally, make timestamps optional in the logging mechanism so we can
configure this with command-line flags or something down the track.
2016-06-10 10:47:28 +12:00
Aldo Cortesi
c421c41307
Remove odict
...
- Adds default implementations for _kconv and _reduce_values to MultiDict.
Without these, operations fail in really, really non-obvious ways.
- Replace the remaining few instances of ODict
Fixes #1159
2016-06-09 13:28:43 +12:00
Shadab Zafar
b3b4156c2f
Py3: Fix test_language_http2 tests by using byte literals
2016-06-08 16:15:54 +05:30
Shadab Zafar
6b03df2633
Py3: Use global next() instead of iterator method
2016-06-08 16:15:54 +05:30
Shadab Zafar
0280af9522
Py3: Use BytesIO instead of StringIO
2016-06-08 16:15:54 +05:30
Aldo Cortesi
a5cb241c7c
If a message has been acked, all other processors are skipped
...
This applies the constraint, but does to clumsily. When we've unified modules
and processors it will be much nicer. We also make some exceptions for the
master processors that we may want to re-evaluate down the track.
2016-06-08 12:58:58 +12:00
Aldo Cortesi
b3bf754e53
Simplify script concurrency helpers
...
We now have take() to prevent double-replies.
2016-06-08 11:21:38 +12:00
Aldo Cortesi
a388ddfd78
A new interface for reply
...
Reply is now explicit - it's no longer a callable itself. Instead, we have:
reply.kill() - kill the flow
reply.ack() - ack, but don't send anything
reply.send(message) - send a response
This is part of an incremental move to detach reply from our flow objects,
and unify the script and handler interfaces.
2016-06-08 10:44:20 +12:00
Aldo Cortesi
982077ec31
Add reply.ack and reply.kill
2016-06-08 10:14:34 +12:00
Aldo Cortesi
db11fe0087
Merge branch 'master' into beautiful-json
2016-06-08 09:19:25 +12:00
Maximilian Hils
c98c83190b
tests++
2016-06-07 01:03:08 -07:00
Shadab Zafar
d254fe791a
Py3: Fix a test_language_writer test by using a byte literal
2016-06-07 00:01:41 -07:00
Maximilian Hils
ba466dbb35
simplify pretty_json, improve test coverage
2016-06-06 23:08:58 -07:00
Aldo Cortesi
1ffc273c94
Utils cleanups
...
- Move more stuff that belongs in netlib.human
- Move some stuff to near the only use
- Zap mitmproxy.utils.timestamp(). I see the rationale, but we used it
interchangeably with time.time() throughout the project. Since time.time()
dominates in the codebase and timestamp() is such low utility, away it goes.
2016-06-07 17:12:52 +12:00
Maximilian Hils
80292c0ac3
fix all pathod tests on Windows 🎉
2016-06-06 16:03:45 -07:00
Maximilian Hils
2ee5e8ef0e
fix NestedMessage encoding
2016-06-06 13:39:38 -07:00
Maximilian Hils
6447c8ae22
fix .freeze(), improve tests
2016-06-06 13:31:44 -07:00
Shadab Zafar
f48073af56
Py3: Fix status_code and other tests by using byte literals
2016-06-06 11:52:45 -07:00
Shadab Zafar
5a2932adc1
Py3: Fix corresponding base.Integer test in test_language_base
2016-06-06 11:52:45 -07:00
Shadab Zafar
40056e1cd7
Py3: Fix useragent tests
2016-06-06 11:52:45 -07:00
Shadab Zafar
f5a6b295f7
Py3: Fix websocket tests
2016-06-06 11:52:45 -07:00
Shadab Zafar
b86ef9fd04
Py3: Use BytesIO in tutils.render
2016-06-06 11:52:45 -07:00
Shadab Zafar
bafce6cfe9
Py3: Use byte string literals
2016-06-06 11:51:36 -07:00
Shadab Zafar
9a19540f8b
Py3: Use BytesIO instead of StringIO for tests
2016-06-06 11:51:36 -07:00
Shadab Zafar
12dd6f6707
Py3: Use global next() instead of iterator method
2016-06-06 11:51:36 -07:00
Shadab Zafar
ced6fb4cf9
Py3: Fix tests by using byte literals
2016-06-06 11:48:58 -07:00
Shadab Zafar
efb0f7133c
Py3: Use global next() instead of iterator method
2016-06-06 11:48:54 -07:00
Aldo Cortesi
210c376901
Fix sequencing in Pathoc websocket termination test
2016-06-06 18:17:22 +12:00
Aldo Cortesi
9458eaece7
Make more of the pathod test suite use synchronisation
2016-06-06 16:05:44 +12:00
Aldo Cortesi
08344ee38b
Merge pull request #1215 from redfast00/master
...
Now keeps header capitalization
2016-06-06 08:45:35 +12:00
Aldo Cortesi
c31b9c461d
Merge pull request #1211 from cortesi/pathod
...
WIP: Radical webectomy of pathod
2016-06-06 08:40:20 +12:00
redfast00
25505733bc
fix multidict unittests
2016-06-05 21:20:37 +02:00
Shadab Zafar
435bfeca0b
Use BytesIO instead of StringIO
2016-06-06 00:36:15 +05:30
Shadab Zafar
7b6da98719
Py3: Use byte literals for comparison
2016-06-06 00:07:44 +05:30
Shadab Zafar
be306c8439
Py3: Use global next() function instead of iterator method
2016-06-06 00:07:44 +05:30
Aldo Cortesi
375d7c9741
Remove last vestiges of noapi and noweb from pathod
...
Also hide HTTP2 freeze bug by making explain configurable in the tests
2016-06-05 13:24:46 +12:00
Aldo Cortesi
48da24ae7e
First-order removal of pathod API and app
2016-06-05 13:04:13 +12:00
Aldo Cortesi
c0c45c051a
Fix test that may fail due to binary header data
...
This is just inherently not a determinisitc test. We don't use the log HTTP
interface any more, so it can just go. A more radical "solution" is inbound
shortly.
Fixes #1207
2016-06-05 11:47:52 +12:00
Shadab Zafar
6ffe8b376e
Py3: Write bytes to file
2016-06-04 18:28:50 +05:30
Shadab Zafar
0d4437a404
Fix function name typo
2016-06-04 18:28:03 +05:30
Shadab Zafar
bd7b275d44
Py3: Convert all data to bytes in pathod.language.writer test
2016-06-04 15:23:09 +05:30
Shadab Zafar
614a3d7a54
Py3: Use global next() instead of iterator method
2016-06-04 15:23:09 +05:30
Aldo Cortesi
d5532319c1
Basic ConsoleMaster tests, based on mastertester
2016-06-04 19:07:27 +12:00
Aldo Cortesi
af63a3e064
console/__init__.py -> console/master.py
2016-06-04 16:06:44 +12:00
Aldo Cortesi
12a4f9ecf9
Very basic test truss for web.app
2016-06-04 15:30:28 +12:00
Aldo Cortesi
f7f9cab5dc
Rebase on master
2016-06-04 14:37:07 +12:00
Aldo Cortesi
e60860e65d
Make tcp.Client.connect return a context manager that closes the connection
2016-06-03 12:06:36 +12:00
Aldo Cortesi
c5076f5e01
Implement a service connection handler counter, use it in Pathod test suite
...
Lots of failures, but that's a good thing.
2016-06-03 11:47:07 +12:00
Aldo Cortesi
c6d15904f3
Merge remote-tracking branch 'upstream/master' into testsuite
2016-06-03 10:19:22 +12:00
Maximilian Hils
29c52d98dc
Merge remote-tracking branch 'duffer/pathod-lang-generators'
2016-06-02 09:53:22 -07:00
Aldo Cortesi
5fd52970fd
Satisfy linting gods
2016-06-02 19:55:52 +12:00
Aldo Cortesi
c9dd574f48
Silence requests warnings
...
Once we convert the rest of the suite to pathoc, this can go.
2016-06-02 19:40:39 +12:00
Aldo Cortesi
d8c52964c7
Reimplement test retrievals on pathoc and ditch requests
...
Requests uses urllib3, which has a connection pool that's not easy to disable
(https://github.com/shazow/urllib3/issues/383 ).
2016-06-02 19:36:02 +12:00
Shadab Zafar
b510616c69
Py3: Return bytes from FileGenerator
...
and use __getitem__ instead of __getslice__
2016-06-02 13:00:44 +05:30
Shadab Zafar
a04d7fd166
Py3: Return bytes from RandomGenerator
...
and use __getitem__ rather than __getslice__
2016-06-02 13:00:02 +05:30
Aldo Cortesi
a7522d9308
pathod.test shouldn't use the API at all
2016-06-02 18:27:51 +12:00
Aldo Cortesi
254614e9f7
Since we have locks over the logs, use direct access rather than API requests to get to them
2016-06-02 18:10:36 +12:00
Aldo Cortesi
ffca395e48
Honesty in testing
...
- Don't skip tests that fail
- Don't omit console from coverage
I'm working on the pathod races, and can't seem to reproduce the failures
locally.
Also, let's not kid ourselves about the console coverage - it should be
tracked. To improve it, we should measure it. Regressions in coverage of the
module should be flagged by CI.
2016-06-02 14:44:06 +12:00
Aldo Cortesi
09da1febbd
Shift a bunch more string-related functions to strutils
2016-06-02 13:04:19 +12:00
Aldo Cortesi
31012d782f
Consolidate some functions to strutils.bytes_to_escaped_str
2016-06-02 12:53:12 +12:00
Aldo Cortesi
cccdc98426
Utils reorganisation: add netlib.strutils
...
Extract a number of string and format-related functions to netlib.strutils.
2016-06-02 12:31:41 +12:00
Aldo Cortesi
eaa3b308f7
Fix non-deterministic test failures in export
...
We had various places in the code where we relied on incidental order of dict
keys. Add a helper to multidict, and fix.
2016-06-02 11:37:18 +12:00
Maximilian Hils
898602ad21
Merge pull request #1180 from dufferzafar/pathod-base
...
Python 3 - pathod.language.base
2016-05-31 14:46:28 -07:00
Maximilian Hils
30fff1fb32
be more strict about module namespaces
2016-05-31 14:05:57 -07:00
Aldo Cortesi
42e91fcfe1
http2.frame -> http2.framereader
2016-05-31 20:32:28 +12:00
Aldo Cortesi
5a75ea3fc6
Fix test failures
2016-05-31 20:19:54 +12:00
Aldo Cortesi
40a030f215
Satisfy flake8
2016-05-31 19:58:28 +12:00
Aldo Cortesi
f62efed304
Unify and make symmetric pretty_size and parse_size
2016-05-31 19:45:48 +12:00
Aldo Cortesi
b2f63458fc
Move human-friendly format functions to netlib.human, remove redundant implementations
2016-05-31 19:32:08 +12:00
Aldo Cortesi
ec34cae618
utils.multipartdecode -> http.multipart.decode
...
also
utils.parse_content_type -> http.headers.parse_content_type
2016-05-31 19:07:55 +12:00
Aldo Cortesi
15b2374ef9
netlib.utils.get_header_tokens -> netlib.http1.read.get_header_tokens
...
Placing this next to its only use.
2016-05-31 18:54:42 +12:00
Aldo Cortesi
6dda2b2ee5
Module is part of the name - url.decode, not url.urldecode
...
A pattern we need to use far more often in the codebase
2016-05-31 18:46:19 +12:00
Aldo Cortesi
4e6c9c4e93
Extract url functions from netlib.utils and move to netlib.http.url
2016-05-31 18:42:56 +12:00
Aldo Cortesi
08fbe6f111
Start cleaning up netlib.utils
...
- Remove http2 functions, move to http2.frame
- Remove Serializable, move to netlib.basetypes
2016-05-31 17:16:31 +12:00
Aldo Cortesi
9ea68ebd28
Improve handling of pseudo-headers
...
- The canonical source for :method, :scheme and :path are the .method, .scheme
and .path attributes on the request object.
- These pseudo-headers are stripped after reading the request, and re-inserted
just before sending.
- The :authority header remains, and should be handled analagously to the Host
header in HTTP1 with respect to display and user interaction.
2016-05-31 14:36:17 +12:00
Aldo Cortesi
bc6509ab6d
Fix test import error
2016-05-31 10:25:02 +12:00
Shadab Zafar
25e30f42be
Py3: Use byte literals for comparison
2016-05-31 01:13:50 +05:30
Maximilian Hils
89f07603ca
split mitmproxy.flow into mitmproxy.flow.*
2016-05-30 01:40:09 -07:00
Thomas Kriechbaumer
be9438b01c
fix
2016-05-29 22:58:56 +02:00
Thomas Kriechbaumer
8cd0c9c330
fix windows paths
2016-05-29 22:47:51 +02:00
Thomas Kriechbaumer
fb0b591242
fix har test
2016-05-29 22:30:15 +02:00
Thomas Kriechbaumer
03167d5e71
fix headers order
2016-05-29 22:07:22 +02:00
Thomas Kriechbaumer
01fcaa64c3
fix #1162
2016-05-29 13:58:14 +02:00
Thomas Kriechbaumer
e2447406cc
more style cleanup
2016-05-29 13:43:20 +02:00
Thomas Kriechbaumer
8eeab37a07
more style cleanup
...
Use this to check:
flake8 --count mitmproxy netlib pathod examples test
2016-05-29 11:55:12 +02:00
Aldo Cortesi
f3bee6f245
Merge pull request #1172 from cortesi/solidcore
...
First steps to solidifying the core
2016-05-29 20:03:40 +12:00
Maximilian Hils
85aa5da6e3
fix tests
2016-05-28 19:31:43 -07:00
Aldo Cortesi
0a875421c5
handle_* -> *
...
Now that we have the controller.handler decorator, the _handler prefix
stutters.
2016-05-29 12:49:01 +12:00
Aldo Cortesi
8bb9285678
Explicitly list all events
2016-05-29 12:15:52 +12:00
Aldo Cortesi
e4f043f69c
Flatten ServerMaster into Master
2016-05-29 12:01:46 +12:00
Aldo Cortesi
ed415877d4
Merge branch 'master' into solidcore
2016-05-29 11:14:46 +12:00
Thomas Kriechbaumer
caee1adeb6
cleanup imports with flake8
2016-05-28 14:43:32 +02:00
Maximilian Hils
fa7964988f
fix Python 2 tests
2016-05-28 00:17:57 -07:00
Shadab Zafar
b71ef054b5
Py3: Use cStringIO from six.moves
2016-05-28 00:17:57 -07:00
Aldo Cortesi
f4813d5649
Make @controller.handler inheritance-friendly
...
Use this to adapt mitmweb and mitproxy console
2016-05-28 12:38:07 +12:00
Aldo Cortesi
f29f6210bb
Mandate that all handlers must be wrapped, make tests pass
...
mitmproxy, mitmdump and mitmweb masters still to be done
2016-05-28 11:00:00 +12:00
Aldo Cortesi
7c4843857b
Sketch out a more solid core
...
- Decorator for handler methods
- Stricter checking for double-acks and non-acks
2016-05-28 11:00:00 +12:00
Shadab Zafar
bc6cd13356
Go Python 3 by default with the bytes conversion
2016-05-26 23:07:37 +05:30
Shadab Zafar
1bbb178b6a
Remove inner_repr, fixup escape_unprintables
2016-05-26 23:07:37 +05:30
Shadab Zafar
22e4bc1938
Py3: Handle bytes case in inner_repr
2016-05-26 23:07:37 +05:30
Maximilian Hils
d3477e27fa
bytes_to_escaped_str: always escape single quotes
2016-05-25 20:11:34 -07:00
Maximilian Hils
7a8da48a30
escaped_str_to_bytes: support unicode on python 2
2016-05-25 19:16:02 -07:00
Thomas Kriechbaumer
ebaad91484
Merge pull request #1127 from mitmproxy/tcp-flows
...
mitmdump: Add Basic Support for TCP Flows
2016-05-23 20:54:14 +02:00
Aldo Cortesi
97f3077082
Merge branch 'mhils-multidict'
2016-05-21 15:01:19 +12:00
Aldo Cortesi
43d7965532
Clean un-needed imports
2016-05-21 15:00:52 +12:00
Aldo Cortesi
123e3b130c
We don't need 3 slightly different implementations of Data
2016-05-21 11:53:14 +12:00
Aldo Cortesi
a5c4cd0340
A clearer implementation of MultiDictView
...
This makes MultiDictView work with a simple getter/setter pair, rather than
using attributes with implicit leading underscores. Also move MultiDictView
into multidict.py and adds some simple unit tests.
2016-05-21 11:37:36 +12:00
Aldo Cortesi
96d8ec1ee3
Merge branch 'multidict' of https://github.com/mhils/mitmproxy into mhils-multidict
2016-05-21 10:15:37 +12:00
Maximilian Hils
84144ca0c6
tests: silence flaky pathod test
2016-05-20 14:05:04 -07:00
Maximilian Hils
a3946d2a2d
tests++
2016-05-20 13:27:26 -07:00
Maximilian Hils
b538138ead
tests++
2016-05-20 11:04:27 -07:00
Maximilian Hils
f96697646c
add FlowMaster.active_flows
2016-05-19 22:09:00 -07:00
Maximilian Hils
d38989fe7e
tests: finish connections properly
2016-05-19 22:07:56 -07:00
Maximilian Hils
d3c30d9005
fix tests, don't double-add error'd flows
2016-05-19 22:07:56 -07:00
Maximilian Hils
acd51befbb
minor fixes, adjust tests
2016-05-19 22:07:56 -07:00
Thomas Kriechbaumer
79cb17ef03
Merge pull request #1152 from mhils/cleanup-odict
...
clean up odict
2016-05-19 16:47:51 +02:00
Maximilian Hils
a7b9e3033d
tests: properly skip pathod tests
2016-05-18 22:51:27 -07:00
Maximilian Hils
6f8db2d7eb
improve MultiDict, add ImmutableMultiDict, adjust response.cookies
2016-05-18 22:50:19 -07:00
Maximilian Hils
8e39b7bf38
test flow export with duplicate query string
2016-05-18 19:28:23 -07:00
Maximilian Hils
d1fc694952
clean up odict
2016-05-18 19:01:49 -07:00
Maximilian Hils
44ac64aa72
add MultiDict
...
This commit introduces MultiDict, a multi-dictionary similar to
ODict, but with improved semantics (as in the Headers class).
MultiDict fixes a few issues that were present in the Request/Response
API. In particular, `request.cookies["foo"] = "bar"` has previously been a
no-op, as the cookies property returned a mutable _copy_ of the cookies.
2016-05-18 18:46:42 -07:00
Maximilian Hils
4c3fb8f509
tests: xfail -> skip
2016-05-18 18:42:35 -07:00
Thomas Kriechbaumer
09b90fce9e
remove debug print
2016-05-17 22:45:09 -07:00
Thomas Kriechbaumer
e61014d203
http2: add connection-lost test
2016-05-17 21:25:10 -07:00
Aldo Cortesi
d27fd55657
Fix broken test_flow_export tests
...
- Make the tests safe for concurrency
- Fix a test that only passed due to prior modification of global state
This also fixes the project generally for parallel test execution.
2016-05-18 12:34:45 +12:00
Aldo Cortesi
00cc836470
Clean up flow export tests
...
- Externalise test data
- Define a less finicky comparator for Python code
2016-05-18 12:16:50 +12:00
Thomas Kriechbaumer
2887480bcb
fix naming which caused overwriting of functions
2016-05-15 10:39:39 -07:00
Thomas Kriechbaumer
f4a13d6f9f
http2: improve header tests ( #1143 ) ( #1144 )
2016-05-14 10:14:42 -07:00
Maximilian Hils
f1c922c652
Sanitize Print ( #1135 )
...
* sanitize strings with shell control characters
* netlib: add utilities to safe-print bytes
* escaped str: add TODO for multi-byte chars
2016-05-12 11:03:57 -06:00
Thomas Kriechbaumer
7a813936c6
disable http2 header encoding, use bytes everywhere
2016-05-11 16:47:58 -05:00
Thomas Kriechbaumer
bef72c1b02
Merge pull request #1121 from Kriechi/fix-cookies
...
improve cookie parsing
2016-05-11 12:30:44 -05:00
Thomas Kriechbaumer
e2ee48a0ee
replace SimpleCookie with our own parser lib
2016-05-11 12:20:21 -05:00
Maximilian Hils
67a37e6d1f
improve script loading
2016-05-11 11:15:36 -06:00
Maximilian Hils
4c8e334642
actually fix tests
2016-05-10 12:19:26 -06:00
Maximilian Hils
bb4f9611f5
Revert "fix tests"
...
This reverts commit f315dc1eb9
.
2016-05-10 12:18:05 -06:00
Maximilian Hils
f315dc1eb9
fix tests
2016-05-10 12:17:25 -06:00
Thomas Kriechbaumer
08002282e8
improve cookie parsing
...
allows '/' to be within a cookie name
removes deprecated cookie getter/setter
fixes #1118
2016-05-10 11:13:12 -05:00
Maximilian Hils
9d47d3b1ec
fix #1120
2016-05-09 12:15:20 -06:00
Maximilian Hils
74cfd7a4e2
stateobject: support lists
2016-04-29 20:59:26 -07:00
Maximilian Hils
317b4becca
mitmdump: don't fail for other flowtypes
2016-04-29 20:18:32 -07:00
Maximilian Hils
54c4859a95
remove netlib version check
2016-04-29 12:00:38 -07:00
Maximilian Hils
909d5ec87e
FlowReadError -> FlowReadException
2016-04-29 11:17:49 -07:00
Steven Van Acker
66267ad276
support for setting/sending multiple cookies ( #1091 )
...
* support for setting/sending multiple cookies
* py.test for multiple cookie support
2016-04-22 11:16:05 -07:00
Maximilian Hils
5b12720962
silence flaky pathod test
2016-04-21 18:29:16 -07:00
Maximilian Hils
028f5c4929
downgrade pyparsing to fix #1087 and #1090
2016-04-21 17:11:28 -07:00
Maximilian Hils
423c076c61
cleanup mitmproxy.controller, raise Kill in Channel ( #1085 )
2016-04-14 12:03:29 -07:00