The VTY prompt may contain regular expression metacharacters in it,
so the self.re_prompt will not match as expected. One example is
the modem app in osmocom-bb.git using promot 'OsmocomBB(modem)'.
Change-Id: I47b1f1c1765acb1b935dcbcc54bbb5eb539db4f0
Tag a new release and to update debian/changelog, with the intention to bring
this package to the latest feed (and not only nightly).
Related: OS#5604
Change-Id: I4cb57c82f4ef28e59495bbe406a042209d210add
Remove the paragraph about writing to the Free Software Foundation's
mailing address. The FSF has changed addresses in the past, and may do
so again. In 2021 this is not useful, let's rather have a bit less
boilerplate at the start of source files.
Change-Id: Ifd7b7e548c9c6db140fa8cbc8a3c8dfaf4603d18
The UserManual documentation specifies errors to use the "ERROR"
keyword, not the "ERR" one. Furthermore, libosmocore (reference
implementation) sends "ERROR" too.
Change-Id: I82cbe8ad0ab51fa073ac3b2d0b959a1610f683b1
Python2 support was dropped recently, remove references informing how to
install python2 modules (operation would fail).
Change-Id: Iffe403377afe9534dbd61d65bfd52e479e125e68
The existing code assumed that tmpdir only contained regular files,
not an entire hierarchy of files with sub-directories or the like.
In case sub-directories exist, the current code fails as follows:
line 176, in test_all_apps
remove_tmpdir(tmpdir)
File "/usr/local/lib/python3.5/dist-packages/osmopython-0.2.0-py3.5.egg/EGG-INFO/scripts/osmotestconfig.py", line 142, in remove_tmpdir
os.unlink(os.path.join(tmpdir, f))
IsADirectoryError: [Errno 21] Is a directory: 'writtenconfig/ericsson'
Change-Id: I896b99032d94ba0cdd340a8eed7c7b625661ad69
Re-apply reverted commit Iabda95073faa2191fd117e9637e0858c589e9d9e
("Drop python2 support"), but with additional changes to make the
scripts actually work with python3 and to make it build without python2.
I have verified, that the contrib/jenkins.sh scripts of all Osmocom
repositories (with their python3 patches on top) are working with this
patch and that all Osmocom repositories with the python3 patches build
in OBS (tested in own namespace).
All related patches for changing from python2 to 3 in other repositories
must be merged shortly after this one, as soon as the build slaves were
(automatically) updated to have the new osmo-python-tests installed:
https://gerrit.osmocom.org/q/topic:drop-py2
New fixes:
* osmopy/obscvty.py: verify: fix compare
Comparing maps in python3 does not work the same as in python2. Convert
them to lists first, so the compare works as intended again.
Fix error:
File "/home/user/code/osmo-dev/src/osmo-python-tests/scripts/osmotestvty.py", line 57, in test_history
assert(self.vty.w_verify(test_str, [t1]))
AssertionError
* osmopy/obscvty.py: use enc/dec with send/recv
Fix error:
self.socket.send("%s\r" % request)
TypeError: a bytes-like object is required, not 'str'
* scripts/osmotestconfig.py: use encode() before writing to file
Fix error:
File "/home/user/code/osmo-dev/src/osmo-python-tests/scripts/osmotestconfig.py", line 91, in copy_config
tmpfile.write(open(config).read())
File "/usr/lib/python3.5/tempfile.py", line 622, in func_wrapper
return func(*args, **kwargs)
TypeError: a bytes-like object is required, not 'str'
* debian/control: add --buildsystem=pybuild. Otherwise "--with python3"
is ignored and the build fails if python2 is not installed, with:
Can't exec "pyversions": No such file or directory at /usr/[...]/python_distutils.pm line 120.
Related: OS#2819
Change-Id: I3ffc3519bf6c22536a49dad7a966188ddad351a7
This reverts commit fb1dc7c405.
I was under the impression, that all previous scripts were already
working with python 3. But as it turns out, this isn't true. Reverting,
so I can properly post follow-up patches, that fix the issues before we
apply this "drop python2" patch again.
Related: OS#2819
Change-Id: Ic1559d1a9f7839fa86a841d62a04b22e1caed466
Remove all compatibility code for python2.
All scripts are already python 3 compatible since
I80e5850a8978d78cda793e2192ef4bd3fd54a121 and
I1b4a629f12863c498a8681b555f57b4e255cebfb.
dpkg-buildpackage shows that it is still invoking setup.py with python
in addition to python3, on debian stretch. But after spending quite some
time on trying to convince it to not care about python2 without success
(trying different variables, overrides, --without python2 flags etc.),
I'm leaving it as is. The resulting package is the python3 package, which
is what we want.
Related: OS#2819
Change-Id: Iabda95073faa2191fd117e9637e0858c589e9d9e
That code in common.py is hit when invoking via osmo_interact_vty.py. It has a
unicode string already, its attempt to decode hits a python exception (no
'decode' method).
Must be a long standing bug that no-one saw because we're only ever using
the osmo_verify_transcript_vty.py variant.
Change-Id: I1b4a629f12863c498a8681b555f57b4e255cebfb
As a result, osmo_interact_vty.py was printing incorrect output:
"""
- <param name='shutdown' doc='Remove the APN from administrative shut-down' />
- </params>
+ <param name='shutdown' doc='Remove the APN from administrative shut-down' /> </params
"""
Change-Id: Ib1dbf39db1b27331ea4c39051e550a87780d9f76
Use a loop instead. Without it the script will eventually crash with a
RecursionError.
File "/usr/bin/osmo_trap2cgi.py", line 211, in conn_client
await ctrl_client(proxy, reader, writer)
File "/usr/bin/osmo_trap2cgi.py", line 202, in ctrl_client
proxy.dispatch(wr, data)
[...]
File "/usr/bin/osmo_trap2cgi.py", line 202, in ctrl_client
proxy.dispatch(wr, data)
File "/usr/bin/osmo_trap2cgi.py", line 201, in ctrl_client
[...]
RecursionError: maximum recursion depth exceeded in comparison
Change-Id: Ic909e371771f3056cb87e18793fd4225ffb90a2c
Related: SYS#4399
* Install systemd file with debian
* Depend on python3-aiohttp in python3-osmopy-utils as it is required
for osmo-trap2cgi
Related: SYS#4399
Sponsored-by: On-Waves ehf
Change-Id: I6add5f37b63dd5fcfdefe0459b539e0050d03897
* split hash generation into simpler functions
* move them into helper file
* split log init into separate function
This facilitates code sharing with new version of ctrl2cgi proxy in
folow-up patches.
Change-Id: Ic5c349349d2c246f934d6701435b1a2d9de77de1
Related: SYS#4399
It might be removed in one of the future releases as we move away from
Twisted. Let's mark it as deprecated to make sure it won't be a
surprise.
Change-Id: Iefd69e7712d456e3bc42c3277bbfa103a2448bb6
Previously we've tried to determine the type of TRAP and handle as many
as possible via generic method dispatcher similar to parent class. Let's
make code simpler by explicitly ignoring all but 'location-state'
TRAPs. This should also reduce debug log output because we'll only log
the TRAP variable for ignored messages instead of entire content.
Change-Id: I42b715cb82eb309950ff387649726504b6c7cf61
Related: SYS#4399
Save and propagate separate timestamp for HTTP request duration to
distinguish between time spent waiting on connection limit availability
from actual http request time.
Change-Id: I14d1db5206864ac9ab7c64f96f67139113ab3fcf
Related: SYS#4399
osmo-mgw's configure.ac won't find osmotestvty.py, unless it is
executable. This allows passing its ./configure with placing the scripts
dir of osmo-python-tests in PATH, instead of installing it.
Change-Id: If6cefd32d0dc544093db08efc114db86b9bb18b0
The same messages are better handled in derived classes using
message-specific functions instead of generic dispatchers.
Change-Id: I4a054eec9ef7ab39e15ac0cd64834a251d4fb2d7
Log repr() of error because it's more conscience which makes it easier to
read the logs: in case of "e" there's additional
"Failure instance: Traceback (failure with no frames)" prefix which is
pretty useless because we won't have proper tracebacks for deferred anyway.
Change-Id: I39870b83cacee9a589636974d04e356ef0d32345
Log http request timestamp and duration on debug loglevel. This is
especially helpful while troubleshooting issues with multiple concurrent
requests under significant load while network issues are present.
Change-Id: I11c8ac67a2730a9c6912694e5b83bbdf08fe357d
Related: SYS#4399
Log the basic location-state trap triggering request and command reply
as info level - that way we can receive useful information without
turning on full debug output.
Change-Id: If8eae9f36292930571954becf6613fada43d301c
When receiving commands for particular BSC, log BSC id known at the time
when request was made, not the one which is part of the reply.
Change-Id: I6acdfddb9a1132f978f2b55c769559b0c29eb3e8
Manual acquire()/release() of semaphore does not limit number of
concurrent requests when combined with explicit yield. Fix this by using
semaphore.run() and removing inilineCallbacks decorator.
Change-Id: I47b8b9f5b726ca0905bb7c023d63b325c7f7d85f
Related: SYS#4399
Previously command-line arguments without defaults took precedence over
config file variables while values from config file which had
command-line counterparts with default value were silently ignored.
Let's fix this by making config file option mandatory values and
removing overlap between command-line and config file parameters.
Change-Id: I471b5a6497eadce6456e835233fdaba88a593324
Related: SYS#4399
Since the scripts are intended to be used as systemd services, there's
no need in separate logging via syslog: systemd will take care of
properly collecting and storing script output. Hence we can drop extra
options and function parameters.
Change-Id: Ifcad1877d45d43b3a2e617775a1c9b256e190591
Related: SYS#4399
* reorder imports to make pylint3 happy
* drop unused imports
* use proper spacing for list constants
* don't use reserved names for internal variables
The check was run as follows:
pylint3 -d C0103,C0301,C0410,C0326,R0913,R0901 ...
to disable useless warnings.
Change-Id: I5b90ee790f73dc509081401776911f25e43f1801
Previously handle_reply() was marked as deferred callback unlike soap.py
where it's synchronous function. This seems to be causing issues where
some of the callbacks are not yield properly. Let's move to the
known-to-work semantics of soap.py where async functions are limited to
Trap() class.
Change-Id: Ib2c28dd7f79cbd28d475de93750703659ddd18f1
Related: SYS#4399
Sometimes when we set variable we might receive arbitrary number of TRAP
messages before we get SET_REPLY. Those could be either separate
messages or combined together with SET_REPLY depending on tcp buffering
at server side. Let's handle this gracefully by skipping over all TRAP
messages.
An example command which often triggers this behavior:
./osmo_ctrl.py -s -d 127.0.0.1 -p 4249 bts.0.location (date +%s)",fix2d,1,2,3"
Change-Id: Ia6de02c2f13a56f0381c97a9ab02c6c7a31cc32f
Related: SYS#4399
Oliver Smith
Pau Espin
Vadim Yanitskiy
Vadim Yanitskiy
Harald Welte
Neels Hofmeyr
Daniel Willmann