dbs [Tue, 22 Feb 2011 14:28:19 +0000 (14:28 +0000)]
Resolve encoding problem found by attempting to register patrons with Unicode names
Symptoms were:
1. a gateway.log entry showing http_translator receiving an XMPP error message
2. a corresponding osrfsys.log entry "server: died with error Wide character in syswrite at /usr/local/share/perl5/OpenSRF/Server.pm line 239"
Per normal Perl Unicode handling practices, decoding input and encoding output
using the Encode module and decode_utf8() / encode_utf8() appears to resolve
the problem.
dbs [Tue, 11 Jan 2011 22:59:18 +0000 (22:59 +0000)]
Clean up Python server implementation, guided by pylint
Add some docstrings, keep line lengths at 80 or less (MORE
WHITESPACE!), remove some dead code, and add a hint to check
max_children config value if we're bumping up against the
limit.
dbs [Mon, 20 Dec 2010 02:55:24 +0000 (02:55 +0000)]
Straighten out -f / -d flag confusion in libmemcached prerequisite
There are two different styles of checking for previously downloaded
tarballs in this Makefile; the libmemcached entry combined them both
with unfortunate results.
erickson [Mon, 20 Dec 2010 01:20:48 +0000 (01:20 +0000)]
Python srfsh enhancements
* Srfsh plugins can now insert new commands and add words to the tab
completion word bank.
* Addded support reading script files
* Added support for service open/close (connect/disconnect) for stateful
connections
* Moved to class-based srfsh module for easier state maintenance
* More doc strings
dbs [Tue, 7 Dec 2010 14:30:45 +0000 (14:30 +0000)]
Apply libtools versioning to OpenSRF c-apps as well
We apply a -version-info argument of 2:0:2, meaning:
* interface = 2: interface '2' of this library
* revision = 0: first implementation of interface '2'
* age = 2: this library can be linked with executables going back
two interface versions
See http://sourceware.org/autobook/autobook/autobook_91.html for more
information on the -version-info argument.
dbs [Tue, 30 Nov 2010 04:16:30 +0000 (04:16 +0000)]
Prepare for version 2.0.0 with versioning for 2.0.0
Use the libtool -version-info option to provide an explicit version
for the libopensrf.so library. Introducing versioned libraries will
give us the ability to force recompiles of linked applications when
binary compatibility has been broken.
erickson [Mon, 29 Nov 2010 19:15:07 +0000 (19:15 +0000)]
add fault tolerance (e.g. recover from interrupted select/sysread); always perform idle maintenance when there are any active child processes; exit check status early when no active processes exist
erickson [Mon, 22 Nov 2010 22:21:35 +0000 (22:21 +0000)]
in the rare case the select() is interrupted while waiting on child statuses, exit early to prevent read()'s on invalid file handles and allow the calling code to loop back around and try again
erickson [Sat, 20 Nov 2010 15:24:55 +0000 (15:24 +0000)]
added support for paging through 'less' (like C); added 'router' query docs; added separate env vars for pretty printing json and for print network object keys; explicit opensrf disconnect on exit
erickson [Tue, 16 Nov 2010 14:08:29 +0000 (14:08 +0000)]
repaired logic in the as-yet-unused import_array_data; use the more efficient enumerate func; capture array length outside of loop for efficiency; avoid calls to set/get_field and access _data directly inside of the NetworkObject class for faster accessor/mutator's
erickson [Tue, 16 Nov 2010 03:36:49 +0000 (03:36 +0000)]
rely on logic, not failure through try/catch, to determine of an object is a registered network object; if an object is unregistered, do not treat it as an __unknown network object (which obfuscates and provides no value).
erickson [Tue, 16 Nov 2010 03:29:32 +0000 (03:29 +0000)]
killed the terminal colors in srfsh.py output. They are cute and all, but are likely the cause of readline oddities/artifacts (and kind of distracting). Default to raw (json) output for consistency
dbs [Thu, 11 Nov 2010 14:47:12 +0000 (14:47 +0000)]
If an alternate config file is pointed to in osrf_ctl.sh, respect it
In a previous edit, I accidentally removed the config file option
from the start_python command, which would break if anything other
than the configured default config file was specified as an option to
osrf_ctl.sh. This makes opensrf.py once again respect the authority
of osrf_ctl.sh.
dbs [Thu, 11 Nov 2010 14:25:02 +0000 (14:25 +0000)]
Use the same PID directory as the opensrf-perl.pl and opensrf.py scripts
The default PID directories for the opensrf-perl.pl and opensrf.py scripts
was /openils/var/run/opensrf (assuming a prefix of '/openils'), while the
default PID directory for osrf_ctl.sh was /openils/var/run. This complicated
matters when trying to restart individual Python or Perl services, as without
passing an explicit PID directory argument the restart command from the
language-specific script would check for a PID in a different PID directory,
find none, and end up running a second copy of the service instead of
restarting the existing service.
The change groups all PIDs in the 'run/opensrf' subdirectory to avoid any
PID file conflicts (as unlikely as that might be) if OpenSRF is installed
in the default location - so rather than /var/local/run/*.pid, one will
be able to look at /var/local/run/opensrf/*.pid
dbs [Tue, 9 Nov 2010 18:17:50 +0000 (18:17 +0000)]
Enhanced do_start() implementation in Python management script
Unclean shutdowns and fork() misery can create PID files that have
no actual process behind them. Instead of just trusting the PID file, check
for the running PID; if the process is not running, then remove the PID file
and actually start the service.
dbs [Tue, 9 Nov 2010 16:21:15 +0000 (16:21 +0000)]
Give opensrf.py reasonable defaults for options
Rather than:
opensrf.py -l -d -f /openils/conf/opensrf_core.xml -p /openils/var/run/ -a start_all
you can now use:
opensrf.py -l -d -a start_all
Isn't that better? Note that we put the PIDs into PID_DIR/run/opensrf/
so that if/when OpenSRF is installed outside of the /openils/ prefix,
the names of the processes won't conflict with any other application PIDs.
Unlikely, but you never know.
dbs [Tue, 9 Nov 2010 14:22:25 +0000 (14:22 +0000)]
Fix daemonize problem that surfaced in start_all
Thanks to Michael Giarlo for reporting the problem, Bill Erickson
for pointing the way to the solution, and http://bugs.python.org/issue5313
for providing me with more context for the problem & solution.
dbs [Mon, 8 Nov 2010 21:03:55 +0000 (21:03 +0000)]
Make opensrf-perl.pl handle missing configuration for services more gracefully
As we might want to ship configuration files with commented out sections,
this will alert the user that a service was listed in the hosts section
but the configuration for that service could not be found.
erickson [Tue, 26 Oct 2010 14:12:15 +0000 (14:12 +0000)]
consistent w/ the C libs, only log server message processing duration at INFO. log server response processing duration at debug and w/ slighly different wording. let the processing duration log message act as the indication of a successfully handled message to reduce logging
scottmk [Thu, 14 Oct 2010 16:30:17 +0000 (16:30 +0000)]
Plug a minor memory leak that could occur when we opened a TCP connection
to a server. In practice this meant whenever we opened a Jabber session.
We call getaddrinfo() to get a dynamically allocated linked list of
addresses for a given server name. Then we traverse the list, looking
for one that accepts streaming connections over IPV4, in order to get
an IP address.
At the end, we call freeaddrinfo() to free the linked list.
Previously we would pass to freeaddrinfo() a pointer, not to the
head of the list, but to the node that we used for getting an IP
address. Prior nodes, if any, would leak.
Also: added calls to freeaddrinfo() in the case of early returns, to
avoid leaking the list in the event of an error.
erickson [Tue, 12 Oct 2010 14:53:41 +0000 (14:53 +0000)]
implemented an optional per-service stderr log for capturing miscellaneous stderr output from services, similar to the old-style _unix.log files, since there are often useful warnings and error messages that never bubble up to syslog. for clarity, the files now use _stderr as a suffix instead of _unix. stderr logs are enabled by default. included opensrf.xml example of how to disable it for a given service
erickson [Tue, 12 Oct 2010 02:58:16 +0000 (02:58 +0000)]
when we add xhr responses onto the end of the queue, stash the original session as well, otherwise we're attempting to ask a session questions about requests it's not tracking
Work around a Net::Domain bug that can result in fqdn's like foo.example.com,bar.com
The bug manifested when trying to run autogen.sh, with the output:
Updating fieldmapper
No Response from settings server...going to sleep
This was because SettingsParser was generating invalid XPath and consquently
errors in the opensrf.settings service; with no response, the attempt to run
autogen.sh would die.
This workaround splits the fqdn on commas and tries each possible domain in
the server setting XPath. Long term we either need to wait for Net::Domain
bug #60729 (https://rt.cpan.org/Public/Bug/Display.html?id=60729) to be
resolved, or consider alternatives.
removing vestigial references to Inbound.pm, which no longer exists in the repository. this also removes get/set_listener, which were never taken advantage of and are no longer of use since the listener is thoroughly hard-coded into the opensrf stack now
Bump trunk version number for OpenSRF up to 1.8.0; better than 1.4.0.
perlmodstyle says "A correct CPAN version number is a floating point
number with at least 2 digits after the decimal." so we're not in
compliance in any case. Oh well.
Update the manifest for Perl modules to avoid build warnings
r2016 removed two Perl modules, src/perl/lib/OpenSRF/Transport/SlimJabber/Inbound.pm
and src/perl/lib/OpenSRF/UnixServer.pm, and the build gets upset because the
manifest suggests they're missing. But no, we know where they are, and we
don't expect them to come back.
gmc [Tue, 31 Aug 2010 20:49:44 +0000 (20:49 +0000)]
remove ntp and ntpdate dependencies
Not used by OpenSRF; these are legacies of
Evergreen dependencies at the time OpenSRF was
split off. Quite likely a bunch of deps in
Makefile.install aren't needed by OpenSRF
proper.
erickson [Wed, 18 Aug 2010 20:11:18 +0000 (20:11 +0000)]
Due to the async nature of xmllhttprequest, processing http responses (which often lead to further async-request laden callbacks) at receive time can result in out-of-order message handling. To bring order to this chaos, push all inbound message onto a queue, then go back to processing the older messages in order of oldest to newest.
scottmk [Mon, 16 Aug 2010 15:18:08 +0000 (15:18 +0000)]
Several minor and inconsequential changes:
1. Tidied up the white space in a few places.
2. Commented out several unused functions, used only for development,
in order to eliminate some annoying compiler warnings.
3. Added the const qualifier to several variables.
4. Added three branches to a switch/case structure, in order to
eliminate some compiler warnings about untested enum values. The
new branches merely issue warning messages to the log (presumably
we don't expect to see RESULT or STATUS messages in this
context).
scottmk [Sat, 14 Aug 2010 11:09:47 +0000 (11:09 +0000)]
Eliminate the old JSON parser, implemented in osrf_json_parser.c and
the associated header osrf_json_utils.h.
This parser has been completely replaced by a newer one implemented
in osrf_parse_json.c, plus an incremental JSON parser (so far
unused) in jsonpush.c..
The even older parser of JSON-with-comments is not affected.
D include/opensrf/osrf_json_utils.h
M include/opensrf/osrf_json.h
D src/libopensrf/osrf_json_parser.c
M src/libopensrf/osrf_json_object.c
M src/libopensrf/Makefile.json
M src/libopensrf/osrf_json_tools.c
M src/libopensrf/Makefile.am
M src/libopensrf/osrf_parse_json.c
M Makefile.am