09a9ed9734
Fix #! at top of shell scripts for py3 and venvs
2020-08-05 17:01:47 -04:00
e7f52a4013
Update versioneer
2020-08-05 16:42:46 -04:00
549a27e66c
Add flake8 tests and clean up warnings
2020-08-03 23:46:34 -04:00
cd68389e9a
Update for Python 3 and psutil >= 2.
2020-08-03 23:08:52 -04:00
6faf563bda
Remove nilmrun.testfilter
...
This module was only used to assist testing, but wasn't actually
available in a test environment where no copies of nilmrun have been
installed. Just code the contents directly into the filters that the
tests are running.
2020-08-03 23:05:32 -04:00
58fd9d1559
Run cherrypy always in "embedded" mode
...
Non-embedded mode is not used in the test suite or wsgi server;
it was an option in the standalong nilmdb-server script, but it's
really not necessary, and removing it gets rid of some untested
code.
2020-08-03 16:57:17 -04:00
2bc939d42d
CherryPy bug 1200 is no longer an issue
2020-08-03 16:55:23 -04:00
fe36722684
Use built-in json module rather than external simplejson
2020-08-03 16:54:47 -04:00
24740a838e
Run 2to3 over all code
2020-08-03 16:54:33 -04:00
cbc7a7dd55
Start Python 3 conversion
2020-08-03 16:51:01 -04:00
477c27a4e6
Clean up temp dirs and processes at exit
2013-07-30 20:15:01 -04:00
bed26e099b
Put each code in its own dir; save args too
2013-07-30 20:14:48 -04:00
9224566f9b
More robust process killing
2013-07-30 20:13:30 -04:00
a8ecad9329
Use NilmDB serializer for ProcessManager
2013-07-24 14:55:59 -04:00
5b878378f3
Translate UTF-8 in command output more robustly
2013-07-22 13:03:09 -04:00
d7551bde0b
Make 'args' optional to /run/code
2013-07-21 19:49:30 -04:00
40fd377a38
Remove 'name' from spawned processes
2013-07-21 19:49:15 -04:00
7c605a469a
Cleanup dependencies
2013-07-18 11:00:53 -04:00
f5225f88f9
Add max CPU percentage
2013-07-17 18:48:55 -04:00
32e59310ef
Fix for dead processes
2013-07-17 18:19:52 -04:00
5a33ef48cc
Add /process/info request
2013-07-17 18:12:44 -04:00
18a5cd6334
Improve boolean parameter parsing
2013-07-15 14:39:28 -04:00
e0709f0d17
Remove multiprocessing due to mod_wsgi incompatibility; use subprocess
...
Multiprocessing and Apache's mod_wsgi don't play nicely. Switch to
manually managing processes via subprocess.Popen etc instead. When
running arbitrary code, we write it to an external file, and running
functions directly is no longer supported.
2013-07-11 11:39:22 -04:00
a7b9656916
Remove parameters from status output
2013-07-10 11:35:17 -04:00
2e9ec63675
Don't catch SystemExit from a subprocess
2013-07-09 13:15:27 -04:00
6d295b840a
Delete trainola; it will live in nilmtools repo
2013-07-08 11:57:45 -04:00
74a05d05d6
Clear out traceback object to avoid reference cycles
2013-07-08 11:44:19 -04:00
35b20c90a5
Rename and reorganize stuff
2013-07-08 11:33:27 -04:00
6d9ee7b405
Ability to run user provided code now
2013-07-07 20:18:52 -04:00
721d6c4936
Add run_code_string function, untested
2013-07-07 15:36:34 -04:00
973d328e1e
Rename some URLs (/process/command -> /run/command, etc)
2013-07-07 14:12:48 -04:00
0338d40226
One fewer process when spawning commands
2013-07-07 14:09:47 -04:00
734e1d9b52
Add /process/command; fix killing of forked processes
...
Now an entire process tree is killed in /process/remove, as long as
each child hasn't changed its process group.
2013-07-07 12:13:13 -04:00
30a3559253
Add "force" option for /process/remove
2013-07-06 16:03:18 -04:00
e2c9575937
More tests
2013-07-05 22:14:43 -04:00
258fe2358d
Filled out test coverage, fixed lots of bugs
2013-07-05 21:58:30 -04:00
f73de35ee6
Work in progress on the process manager. Very finicky.
...
The multiprocessing environment is really finicky.
I'm seeing deadlocks in the process at exit, that are probably related
to tracebacks being printed and still getting redirected to the
logfile or something.
2013-07-05 19:35:09 -04:00
65e48caf5f
Start test framework; work on server and manager
2013-07-05 16:47:50 -04:00