cbb5d790e5common: Add mandatory length field to UDP receive calls
Tom Tsou
2016-05-02 17:56:15 -0700
29ef2c82dbcommon: Restrict UDP binding to localhost only
Tom Tsou
2016-05-02 17:55:42 -0700
11d8baa826also ignore LockTest
Michael Iedema
2014-12-04 21:16:43 +0100
cc62ca4bb7sync from commercial 270263e2437d5105b36fec2015176cfddec48ffb
Michael Iedema
2014-12-04 19:50:34 +0100
3ad343b97bignore new test progam binaries
Michael Iedema
2014-07-16 06:13:19 +0200
88311c0670merge 5.0 preview from commercial
Michael Iedema
2014-07-16 04:32:22 +0200
00d91e90bbNo real code changes only to comments and logging for GPRS (upstream r8287)
Scott Van Gundy
2014-04-18 11:02:18 +0200
7e87828c3bAllow an empty log level to mean use the default level. This allows the CLI command "unconfig" on Log.Level.* or Log.Group.* to work as expected. (upstream r8227)
Pat Thompson
2014-04-18 10:59:46 +0200
d45cfa5ab5Bug 1707 related changes - convert timestamps to use ISO date/time format. (upstream r8170)
dgotwisner
2014-04-18 10:55:44 +0200
df638854dbAllow IP:PORT in address lists
Shaddi Hasan
2014-03-03 11:53:51 -0800
7c1ebc3899Fixed zero length sms message in both text and 3gpp mode. Fixed many crahses along the way. Changes were made in openbts and smqueue. Fixed issue with loading smqueue from back up file. All cases of plain text and 3gpp are working now. Improved logging in smqueue. Fixed error where data was be sent in 3gpp mode when text was selected.
svangundy
2014-03-31 12:34:00 +0200
778aa50799Add some trivial new routines to class SimpleKeyValue.
pat.thompson
2014-03-31 12:29:21 +0200
2f92ebc717Change LOGHEX to use long instead of unsigned type so it will compile on 64 bit machines.
pat.thompson
2014-03-31 12:28:39 +0200
6ee4814b7eThe socket code throws errors, but since those are usually uncaught, it causes the containing program to end. So added a devassert in these cases so we can catch how this happens if the Log.Level is DEBUG. I did this to track down a bug, but of course, now the bug no longer occurs.
pat.thompson
2014-03-31 12:27:55 +0200
3560e43d9dFixes: Calling LOG from sqlite3util when the database cannot be accessed causes infinite recursion when the LOG message attempts to access the database. Fixes ticket #1733
pat.thompson
2014-03-31 12:27:00 +0200
1b7ea66cf9typo in text output
olga.bobrova
2014-03-31 12:26:00 +0200
0dc6d369e8Back out r6976, r6977, r6980, r6981, which were namespace removals.
pat.thompson
2014-03-31 12:17:26 +0200
c17e42bd8eResult of merging features/dev -r7759:7824 to trunk.
pat.thompson
2014-03-31 12:15:14 +0200
b21473440fBug 242 (yes, that old).
dgotwisner
2014-03-31 12:12:35 +0200
7b0e351a23Remove smarts to cache thread id, as I believe there was a locking issue and significantly added overhead dealing with locking each time than in just doing the system call. I expect that this will allow 1655 to work for Neel.
dgotwisner
2014-03-31 12:11:05 +0200
9cf9130509Use LOG(CRIT) when existing so a diagnostic message is printed to the console before exiting. Also clean up code slightly.
pat.thompson
2014-03-31 12:08:20 +0200
6129dbfbc2Changed logging system to "cache" gettid() values, to avoid system calls at runtime (ie, when logging). The implementation is cludgy. This is because we create threads before main, as a result of global thread objects (as opposed to pointers to thread objects that get new'ed and started in main). A view of the system through strace indicates one or two calls to gettid() per thread (usually 1, occasionally 2). Changing the read lock in Log() to a write lock deadlocks.
dgotwisner
2014-03-31 12:04:56 +0200
73ef35474fEliminate getpid() calls when logging.
dgotwisner
2014-03-31 12:03:44 +0200
06343d3812Changed logger to use getpid() and gettid() instead of pthread_self(), as pthread_self() just returns a pointer to an opaque data structure used by pthreads, and the other two give information that is useful to relate to top, ps, etc. Accidently changed indent in SelfDetect::Exit(). Sorry Michael. Found as a result of working through bug 1608.
dgotwisner
2014-03-31 12:02:24 +0200
7f33081522- remove need for 3rdParty include reference, libcoredump now can be built-from-scratch and installed as lib and -dev headers
Michael Iedema
2014-03-31 11:30:56 +0200
74ab41ae30Add code to save various /proc files to a file when we crash (with core generation). Tied to 2 configuration variables.
dgotwisner
2014-03-31 11:16:25 +0200
9c17f09fe6Add support to CommonLib? to support google core dumper.
dgotwisner
2014-03-31 11:15:28 +0200
3a2602c75dRemove "hack" and put in correct signal number define.
dgotwisner
2014-03-31 11:01:19 +0200
414179aa82- merging in Dave's selfDetect branch so we can limit our components to only a single instance - small changes to log messages, passing signal information through to the log and also echoing failed starts to to cout - a HACK was needed to avoid signal 17, Dave can smooth this out in a more intelligent way than my workaround - ticket #1549
Michael Iedema
2014-03-31 10:59:57 +0200
4575b4d264Fix more code with respect to bypassing "using namespace". Currently, Utils.h still has a using namespace, to prevent non-CommonLibs? code from breaking. All CommonLibs? code has been compiled with tue Utils.h "using namespace" commented out, though.
dgotwisner
2014-03-31 10:50:02 +0200
2b6d0c6fa7Remove more "using namespace" from source files.
dgotwisner
2014-03-31 10:32:27 +0200
66a47d9bffEliminate more "using namespace" statements.
dgotwisner
2014-03-31 10:31:29 +0200
b8ca812fdeRemove "using namespace std" from these source files, and fix the usage of string, cout, etc. accordingly, to match both our coding convention (from the wiki) and the typical C++ ruleset of never using "using namespace".
dgotwisner
2014-03-30 19:26:54 +0200
00294a9fbdChange Vector constructor code from call to vConcat() to call to this->vConcat(). This allows us to compile under gcc 4.8.1 (Ubuntu 13.10's compiler). This is more consistent with the C++ language definition of calling a base construtor from inside a derived constructor.
dgotwisner
2014-03-30 19:25:31 +0200
abec8dabb6merged in commercial commonlibs
Michael Iedema
2014-03-24 22:57:31 +0100
bd94d0a988ignore some things
Michael Iedema
2014-03-22 16:18:17 +0100
94d53215c1allow EGSM
Kurtis Heimerl
2013-12-31 19:55:16 +0000
b08981537f- ignoring some stuff
Michael Iedema
2013-09-18 08:58:36 +0000
7a7862c9d5syncing CommonLibs, and including BSD-licensed A5/1 implementation.
Kurtis Heimerl
2013-08-13 23:17:46 +0000
59515c11f0Complete Configuration sync. While testing, I discovered what I think is a race condition: getStr("SIP.SMSC") returned another config key ("GSM.MS.TA.Damping") which I assume is a race condition in the config cache. However, I've been unable to replicate the issue and I think it may resolve itself quickly (as these things are successfully cached)
Kurtis Heimerl
2013-05-31 21:47:42 +0000
e6a48a3690syncing commonlibs with Many thanks to Michael Iedema for these patches, makes config a lot better.
Kurtis Heimerl
2013-05-31 21:47:25 +0000
9b8ad3802eHarvind found a bug in BitVector.cpp:
Kurtis Heimerl
2012-03-08 07:13:15 +0000