Victor Seva via sr-dev writes:
> @juha-h please use ``cherry-pick -x `` when back-porting
I used `git cherry-pick . This gets too complicated for
me so better that I leave the git stuff to those who know.
___
Kamailio (SER) - Development Mailing List
Module: kamailio
Branch: 5.8
Commit: 3a34b468ad93904c507bd0fc7ef592b25321807f
URL:
https://github.com/kamailio/kamailio/commit/3a34b468ad93904c507bd0fc7ef592b25321807f
Author: Juha Heinanen
Committer: Juha Heinanen
Date: 2024-05-20T19:54:45+03:00
tm: add mandatory Max-Forwards header to local
Module: kamailio
Branch: master
Commit: 0054b875045b4badfedd3e8424bf4772d6a9bc6b
URL:
https://github.com/kamailio/kamailio/commit/0054b875045b4badfedd3e8424bf4772d6a9bc6b
Author: Juha Heinanen
Committer: Juha Heinanen
Date: 2024-03-12T22:18:20+02:00
Avoid 'qm_strnstr' defined but not used
Module: kamailio
Branch: 5.8
Commit: 2122afae2a6604c0bc44a43ab37ee274a15a700a
URL:
https://github.com/kamailio/kamailio/commit/2122afae2a6604c0bc44a43ab37ee274a15a700a
Author: Juha Heinanen
Committer: Juha Heinanen
Date: 2024-03-12T22:16:23+02:00
Avoid 'qm_strnstr' defined but not used
Module: kamailio
Branch: 5.8
Commit: 1928af8f90ebf02f610c0904bc3d24a00aaedb81
URL:
https://github.com/kamailio/kamailio/commit/1928af8f90ebf02f610c0904bc3d24a00aaedb81
Author: Juha Heinanen
Committer: Juha Heinanen
Date: 2024-03-12T07:54:50+02:00
core: set null-terminated char for modparam
How about the diff below?
Also, is there plan to backport ksr_tcp_msg_data_timeout,
ksr_tcp_msg_read_timeout, and ksr_tcp_check_timer to 5.7, since they can
help in protecting from DoS attacks that we have seen in the wild.
-- Juha
diff --git a/src/main.c b/src/main.c
index
Daniel-Constantin Mierla writes:
> > In order to make configuration simpler, how about having a dynamic default
> > as I suggested?
>
> Default is half (hardcoded), but its own parameter gives more
> flexibility for granularity -- one extra parameter is not adding much
> complexity, imo, and
Daniel-Constantin Mierla writes:
> Smaller is recommended for better accuracy.
In order to make configuration simpler, how about having a dynamic default
as I suggested?
___
Kamailio (SER) - Development Mailing List
To unsubscribe send an email to
If so, how about defaulting to half of tcp_msg_read_timeout or
tcp_msg_data_timeout whichever is smallest?
___
Kamailio (SER) - Development Mailing List
To unsubscribe send an email to sr-dev-le...@lists.kamailio.org
Daniel-Constantin Mierla via sr-dev writes:
> core: added tcp_check_timer parameter
>
> - set the check interval (in seconds) for tcp connections
> - default 10
Does this need to be smaller than tcp_msg_read_timeout and
tcp_msg_data_timeout in order for them to have effect?
Victor Seva writes:
> @juha-h can you please test https://github.com/kamailio/kamailio/pull/3482?
> At least the OPENSSL_fork_* warnings should be solved.
I gave the PR a try and now got the warnings below (no fork warnings
anymore).
-- Juha
tls_mod.c: In
Kamailio 5.7 build on Debian Bookworm gave me the following build
warnings. The only curl related build dependency that I have is
libcurl4-openssl-dev and the only openssl related build dependency is
libssl-dev.
-- Juha
--
CC (gcc) [M http_client.so]
I guess I missed the clang-format discussion or did it happen?
Fr example, I would have voted for
SpaceBeforeParens: ControlStatements
-- Juha
___
Kamailio (SER) - Development Mailing List
To unsubscribe send an email to sr-dev-le...@lists.kamailio.org
Daniel-Constantin Mierla writes:
> I am for it, maybe the next major release will be bumped to 6.0, so
> going with clang formatting would be good because makes the readability
> better.
Which style? https://clang.llvm.org/docs/ClangFormatStyleOptions.html
supports many:
LLVM A style complying
Daniel-Constantin Mierla writes:
> I will check it, but in general the kemi function is executed by the
> native kamailio.cfg function, after it evaluates its parameters with
> kamailio pseudo-variables. It is nothing specific for the kemi
> function than just needing the string/int values of the
Daniel-Constantin Mierla writes:
> Hopefully the dispatcher and jsonrpcs warnings are fixed. I guess that
> local.so is a custom module, you have to change the logging/printf()
> specifiers from %d to%ld because some internal fields in kamailio were
> changed from int to long int for a better
Latest K master crashes on Debian 11 when handling SUBSCRIBE message.
(Earlier messages didn't go through when there was word subscribe in
subject.)
-- Juha
--
(gdb) bt
#0 0x7f58c362cad0 in trim_leading (_s=0x7f58bceeb558) at
../../core/trim.h:59
#1
I'm getting a few compiler warnings when building latest master on
Debian 11.
-- Juha
---
CC (gcc) [M dispatcher.so] ds_ht.o
In file included from ../../core/mem/shm.h:43,
from ../../core/mem/shm_mem.h:32,
from
Michael Furmur writes:
> https://github.com/kamailio/kamailio/pull/2812
Thanks, Juha
___
Kamailio (SER) - Development Mailing List
sr-dev@lists.kamailio.org
https://lists.kamailio.org/cgi-bin/mailman/listinfo/sr-dev
C (gcc) [M outbound.so] outbound_mod.o
outbound_mod.c: In function 'mod_init':
outbound_mod.c:112:25: warning: pointer targets in passing argument 1 of 'SHA1'
differ in signedness [-Wpointer-sign]
SHA1(flow_token_secret.s, flow_token_secret.len, ob_key.s);
~^~
Daniel-Constantin Mierla writes:
> Give it a try again with latest master, I did also some basic tests and
> seems ok.
Thanks, works now.
-- Juha
___
Kamailio (SER) - Development Mailing List
sr-dev@lists.kamailio.org
This used to work:
$var(count) = $(var(servers){param.count,%});
but latest master gives:
Jun 1 16:02:04 lab sip-proxy[11965]: 0(12020) INFO: pv [pv_shv.c:60]:
shvar_init_locks(): locks array size 16
Jun 1 16:02:04 lab sip-proxy[11965]: 0(12020) ERROR: pv [pv_trans.c:3254]:
Daniel-Constantin Mierla writes:
> Regarding topos_redis, I just pushed a commit to it in master branch,
> can you give it another try?
Still when A receives 183 Session Progress from P1 and send PRACK, P1
rejects it with 404 Not Found (pcap enclosed). Didn't have debug on,
but can do if
Daniel-Constantin Mierla writes:
> The idea was to try to optimize and tailor it for redis specific
> operation, otherwise you can try with topos+db_redis, which should give
> a more SQL-like approach and the columns in mysql tables should appear
> as attributes inside redis db. The db_redis was
Daniel,
I studied a bit topos_redis source and noticed that it lacks x_context
td and tt keys.
Is there something else missing in order to make spirals work also when
redis backend is in use?
-- Juha
___
Kamailio (SER) - Development Mailing List
Daniel-Constantin Mierla writes:
> Author: Daniel-Constantin Mierla
> Committer: Daniel-Constantin Mierla
> Date: 2021-04-26T11:22:20+02:00
>
> core/mem: q_malloc - use #else with DBG_QM_MALLOC for debug on freed fragments
The large number of log messages are still there, but perhaps this was
Daniel-Constantin Mierla writes:
> Pushed a commit to fix the error message.
Now kamailio starts, but when baresip sip client tries to connect over
TLS, I get these errors:
Mar 23 14:02:23 lohi /usr/bin/sip-proxy[1732]: ERROR: tls [tls_server.c:1283]:
tls_h_read_f(): protocol level error
Mar
Daniel-Constantin Mierla writes:
> I pushed a commit to make that error message print the file name. Try
> again with latest master and paste here again the new error message to
> see what's there.
Mar 23 13:41:37 lohi /usr/bin/sip-proxy[7261]: ERROR: tls [tls_domain.c:604]:
load_ca_list():
I got latest master properly installed and gave ca_path a try.
I placed four ca certs in directory /etc/sip-proxy/certs/ca_list:
# ls /etc/sip-proxy/certs/ca_list
class3_X0E.crt dst_root_ca_x3.pem lets-encrypt-x3-cross-signed.pem
root_X0F.crt
Then I created file ca_list.pem that contained
Daniel-Constantin Mierla writes:
> I was just about to write you on the other thread to test and see if
> works, as I did the development but without having the time to test.
I tested with tcp and tls listeners
Mar 12 20:16:43 sip-proxy[21780]: Listening on
Mar 12 20:16:43 sip-proxy[21780]:
Daniel-Constantin Mierla writes:
> core: execute event_route[core:worker-one-init] for sctp or tcp
>
> - if no udp worker is started, the event route is executed by next
> available transport, in the order: sctp, tcp (or tls)
Thanks, Juha
___
Module: kamailio
Branch: master
Commit: 857e9deb4c2455ee2ea0c05446da54f97e1f0a63
URL:
https://github.com/kamailio/kamailio/commit/857e9deb4c2455ee2ea0c05446da54f97e1f0a63
Author: Juha Heinanen
Committer: Juha Heinanen
Date: 2021-02-11T17:44:19+02:00
Improved README.
---
Modified: src
Daniel-Constantin Mierla writes:
> rr: use rr params with record_route_preset()
>
> - param buffer was checked, but not used
Thanks. Tested and now add_rr_param() has effect on R-R headers added by
record_route_preset(). add_rr_param() can be called both before and
after
Module: kamailio
Branch: master
Commit: 2664a51c35a1e8b5adc4c7e6cedde60bc01f0b28
URL:
https://github.com/kamailio/kamailio/commit/2664a51c35a1e8b5adc4c7e6cedde60bc01f0b28
Author: Juha Heinanen
Committer: Juha Heinanen
Date: 2021-02-11T17:27:44+02:00
Fixed typo.
---
Modified: src/modules/rr
Daniel-Constantin Mierla writes:
> @juha-h - can you try with latest master again? I found an issue in
> ctl module that was introduced in master branch a while ago, but after
> 5.4. It may be that one. If it is not fixed, then get the `kamctl
> trap` after it hangs, to see where the processes
Daniel-Constantin Mierla writes:
> I pushed a fix for two cases when allocation was not initialized. Can
> you try with latest master?
Thanks, now keepalive.list works when there is nothing in the list.
But the second issue (that I didn't create yet) still exists, i.e.,
trying to add something
Juha Heinanen writes:
> keepalive.list rpc command causes kamailio master to crash (at least
> when there is nothing in the list).
I tried to add something to the list with rpc command:
keepalive.add sip:example.com test
and the command hangs.
Perhaps keepalive module is not supported a
keepalive.list rpc command causes kamailio master to crash (at least
when there is nothing in the list).
-- Juha
---
gdb) bt full
#0 rpc_struct_add (s=0x560b4df75830, fmt=0x7f2da23322c3 "SS") at
binrpc_run.c:1092
ap = {{gp_offset = 32, fp_offset = 48,
sergey-safarov writes:
> In this case, the TCP session may be expired on the NAT device, and
> future calls to the user phone not possible.
Usually clients generate TCP keepalives and there is no need for
Kamailio proxy to do anything.
-- Juha
___
I gave --cfg-print a try. I have this in config:
loadmodule "tls"
...
# -- tls params
#!ifdef MOD_tls
modparam("tls", "config", "/etc/sip-proxy/tls.cfg")
modparam("tls", "cipher_list", "HIGH:!ADH:!AECDH")
#!endif
# -- tm params
and --cfg-print gave this kind of output:
loadmodule "tls"
...
#
Daniel-Constantin Mierla writes:
> So for now, if someone doesn't want the #!KAMAILIO (or the other
> variants) in the first line, can add it temporarily only when it uses
> the --cfg-print, then remove it. Of course, if someone wants to look
> more into it, is more than welcome to make pull
Daniel-Constantin Mierla writes:
> Very important - for now the `--cfg-print` works if the main
> configuration file has the first line:
> #!KAMAILIO
Can there be other stuff on the line after #!KAMAILIO?
For example emacs can be told about the major mode used in the file by
checking if first
Daniel-Constantin Mierla writes:
> This is due to some ongoing work started in the morning -- I just
> updated the wiki core cookbook and the upgrade guidelines:
Just after sending the email, I saw the emails of the related commits
and figured out the change.
-- Juha
Daniel-Constantin Mierla writes:
> I looked at the code and indeed there was an issue with the flag-based
> protocol detection. I pushed a commit to fix it, let me know if works ok
> now.
Works now, thanks,
-- Juha
___
Kamailio (SER) - Development
Daniel-Constantin Mierla writes:
> Actually that error message log is after non-sip msg callback. That part
> should not be affected by my changes and execution should not get there
> if xhttp module is properly loaded and configured. Can you check if
> other error messages are printed? Maybe the
Daniel-Constantin Mierla writes:
> For the moment set core parameter:
>
> sip_parser_mode=0
Didn't help. Still
Oct 14 20:53:47 lohi /usr/bin/sip-proxy[23198]: ERROR:
[core/receive.c:356]: receive_msg(): no via found in request
-- Juha
___
With current master, event_route [xhttp:request] is not run. Instead
error message shows up:
Oct 14 20:00:48 lohi /usr/bin/sip-proxy[8820]: ERROR:
[core/receive.c:356]: receive_msg(): no via found in request
After spending one hour in trying to figure out, what is wrong in my
config, I build
Daniel-Constantin Mierla writes:
> I am fine to add it. So far I preferred to discard unwanted traffic at
> the brorder of the core platform, so I didn't need such feature.
I agree, better to do blocking at the edge.
-- Juha
___
Kamailio (SER) -
How about adding source addr param to pike_check_req()?
Motivation is that main proxy may be front-ended by another proxy that
passes the real source addr to the main proxy in a header.
-- Juha
___
Kamailio (SER) - Development Mailing List
Daniel-Constantin Mierla writes:
> I pushed a commit to fix it, can you try with latest master?
Tried with latest master and crash does not happen anymore at shutdown.
Thanks, Juha
___
Kamailio (SER) - Development Mailing List
Daniel-Constantin Mierla writes:
> The kamailio wiki has a turorial about troubleshooting with core files
> where it gives the details about how to enable core file per pid (I am
> on mobile link right now, not easy to give the url).
OK, will try to look.
> To understand that it crashes only on
Juha Heinanen writes:
> Just built K from current master and it crashes every time with similar
> gdb trace when I stop it using 'service sip-proxy stop'.
As a test, djJust build K from latest 5.4 branch and it does not crash
at stop with exactly same config setup and environment. So look
Daniel-Constantin Mierla writes:
> The backtrace is for the main process during shut down, which can be a
> side effect of the real issue. You should enable core dump per pid to
> see if you get core files from more than one process, then grab
> backtraces from all those core files to figure out
K master has recently crashed several times on Debian 10. Could this have
something to do with openssl libs?
-- Juha
Oct 5 08:32:21 lohi systemd[1]: Stopping LSB: Start/stop OpenSIPg SIP Proxy...
Oct 5 08:32:21 lohi sip-proxy[19922]: Stopping sip-proxy
Oct 5 08:32:21 lohi
Module: kamailio
Branch: master
Commit: d93de03c10eb7b88024d7eb9ba85200601cf6c7a
URL:
https://github.com/kamailio/kamailio/commit/d93de03c10eb7b88024d7eb9ba85200601cf6c7a
Author: Juha Heinanen
Committer: Juha Heinanen
Date: 2020-08-25T20:11:59+03:00
sqlops module: added one more sql query
Juha Heinanen writes:
> Victor Seva writes:
>
> > I cannot be sure 100% but from what I understand if you set etag... it has
> > the same behavior than previously.
> > If is the first PUBLISH, there's no record, so no changes in logic.
> >
> > Change
Victor Seva writes:
> I cannot be sure 100% but from what I understand if you set etag... it has
> the same behavior than previously.
> If is the first PUBLISH, there's no record, so no changes in logic.
>
> Change is merged already do you detect any regression?
I can test when the merge is
Henning Westerholt writes:
> * 5 new modules have been added (pv_headers, kafka, secsipid,
> * systemdops, dlgs)
Could not find many of these on page:
https://kamailio.org/docs/modules/devel/
-- Juha
___
Kamailio (SER) - Development Mailing
Daniel-Constantin Mierla writes:
> I pushed a fix for it (not tested) -- it was the new siptrace mode
> trying to save to database using core callbacks and not checking it is
> sip traffic or not. The errors were for the http traffic.
Thanks, the error messages are now gone.
-- Juha
This is the config command that my SIP proxy executes on the incoming
JSON over HTTP request:
jsonrpc_dispatch();
-- Juha
___
Kamailio (SER) - Development Mailing List
sr-dev@lists.kamailio.org
Juha Heinanen writes:
> I noticed that json rpc commands have started to fail, for example,
Actually, the commands still work even when error messages are produced.
-- Juha
___
Kamailio (SER) - Development Mailing List
sr-dev@lists.kamailio.org
ht
I noticed that json rpc commands have started to fail, for example,
May 2 15:41:15 char /usr/bin/sip-proxy[10440]: INFO: Executing JSON request
<{"jsonrpc":"2.0","method":"siptrace.status","params":["on"],"id":1}> from
<127.0.0.1> with host <127.0.0.1:6060>
May 2 15:41:15 char
Daniel-Constantin Mierla writes:
> It was introduced by my previous commit. I pushed the fix for it.
OK, thanks, running now.
-- Juha
___
Kamailio (SER) - Development Mailing List
sr-dev@lists.kamailio.org
I build my sip proxy from today's master and it didn't start:
May 2 15:12:51 char sip-proxy[22422]: 0(22474) ERROR:
[core/sr_module.c:512]: load_module(): could not open module
:
/usr/lib/x86_64-linux-gnu/sip-proxy/modules/usrloc.so: undefined symbol:
ul_cseq_delay
-- Juha
So if tracing is controlled only by rpc commands, then trace_init_mode
can have value 1?
-- Juha
> + 1 - module initialized to do tracing only with
> core callback
> + functions (see also 'trace_mode' parameter).
Daniel-Constantin Mierla writes:
> On the other hand, as enhancement, would it make sense to have
> trace_mode also changeable via rpc? Not sure if someone would want to
> change from mirroring to hep or uri to storing to database, or the other
> way around ...
It might be useful to be able to
Daniel-Constantin Mierla writes:
> Can you try the master branch with trace_on=1 and trace_mode=2 for
> siptrace module? You need to set also the db connection parameters,
> otherwise no siptrace flag set or function use in routing blocks. See
> also the docs of trace_mode param.
Thanks for the
Daniel-Constantin Mierla writes:
> That might not be that complex to do, right now all traffic can be
> mirrored via HEP by setting the trace_mode parameter. If someone wants
> to test it, I can extend it when I get a bit of time during the next few
> days with an option to write the traffic to
Federico Cabiddu writes:
> I would like to have the developers' feedback before opening a PR, there
> could be other scenarios/use cases I'm not considering here.
My feeedback as a K user is this: I would prefer capability to turn
on/off tracing of all messages (requests/replies) that
Daniel-Constantin Mierla writes:
> OK, indeed, the previous behaviour should be preserved in this case. Is
> sip_trace() without params now doing transaction mode capturing?
If I remember correctly, there was discussion about this a few months
ago.
Behavior of
modparam("siptrace", "db_url",
Daniel-Constantin Mierla writes:
> Apparently clang-libc on bsd/macos has different types for tmval fields
> than gcc on linux. I pushed a commit to do casts and avoid warning in
> any of those platforms.
Thanks, warnings are now gone on Debian.
-- Juha
version.xml:14: namespace error : Namespace prefix db on para is not defined
^
___
Kamailio (SER) - Development Mailing List
sr-dev@lists.kamailio.org
https://lists.kamailio.org/cgi-bin/mailman/listinfo/sr-dev
CC (gcc) [M usrloc.so] ul_keepalive.o
ul_keepalive.c: In function 'ul_ka_urecord':
ul_keepalive.c:61:18: warning: format '%x' expects argument of type 'unsigned
int', but argument 26 has type '__suseconds_t' {aka 'long int'} [-Wformat=]
#define ULKA_MSG "%.*s %.*s SIP/2.0\r\n" \
Daniel-Constantin Mierla writes:
> They were from the previous commits indeed -- clang on mac was not
> reporting them. I just pushed patches to fix them.
Thanks, works now, Juha
___
Kamailio (SER) - Development Mailing List
sr-dev@lists.kamailio.org
May have something to do with recent two commits.
-- Juha
---
CC (gcc) [M pv.so] pv_trans.o
pv_trans.c: In function 'tr_eval_string':
pv_trans.c:539:22: warning: pointer targets in passing argument 1 of
Module: kamailio
Branch: master
Commit: 51d95e0e7cce0483c517e2bcfc5000b60a7e35be
URL:
https://github.com/kamailio/kamailio/commit/51d95e0e7cce0483c517e2bcfc5000b60a7e35be
Author: Juha Heinanen
Committer: Juha Heinanen
Date: 2020-03-06T09:08:12+02:00
modukles/tm: added t_clean() script
I tried to implement dynamic registering/unregistering of siptrace
module sl callbacks, but it turned out that sl module lacks API function
to unregister a callback.
-- Juha
___
Kamailio (SER) - Development Mailing List
sr-dev@lists.kamailio.org
Daniel-Constantin Mierla writes:
> There is also a rpc command to turn tracinf on/off, but it doesn't
> affect the trace_mode=1. Again, for the next major release we should
> revise and make those options work for all variants: mirroring or save
> to db, with control via modparams and rpc...
I
sergey-safarov writes:
> Now unit contains
> ```
> RuntimeDirectory=kamailio
> ```
...
> Look as current definition correct
Yes, RuntimeDirectory is correct, but not this:
PIDFile=/var/run/kamailio/kamailio.pid
since it hardwires /var/run rather than using $RUNTIME_DIRECTORY.
-- Juha
Daniel-Constantin Mierla writes:
> Does this one still needs to be addressed, or are we fine with the
> version included in the deb/rpm packages right now?
It works right now (with the syslog message), but according systemd
documentation, the dir is /run, not /var/run:
Looks like systemd service pid files should nowadays be placed in /run
dir instead of /var/run. I noticed this when I got complaint about
fail2ban to syslog:
Oct 17 20:59:35 lab systemd[1]: /lib/systemd/system/fail2ban.service:12:
PIDFile= references path below legacy directory /var/run/,
vinod mn writes:
> For a cancel message from a user kamailio is send "200-canceling" so I want
> to supress this for some reasons.
Relay the request statelessly and then on reply route drop the 200.
-- Juha
___
Kamailio (SER) - Development Mailing
vinod mn writes:
> how to remove the "200-canceling" response from kamailio it is very
> urgent.
So you ask, how to make Kamailio break RFC3261?
-- Juha
___
Kamailio (SER) - Development Mailing List
sr-dev@lists.kamailio.org
Daniel-Constantin Mierla writes:
> The callbacks for tm are used also for database storage, those from the
> core for packet-in/packet-out are for hep only at this moment.
What if only messages are traced, i.e., not transactions nor
dialogs. Are callbacks still used?
-- Juha
Daniel-Constantin Mierla writes:
> There are transaction callbacks enabled because some of them can be set
> only at startup, detection whether siptrace should be done or not for
> that transaction being done later in that function, once the pointer to
> the message is figure out.
My
Using 5.3 branch, I have
sip-proxy_ctl> siptrace.status check
Disabled
but still these kind of errors appear in syslog:
Oct 16 12:07:10 salmon /usr/bin/sip-proxy[26459]: ERROR: siptrace
[siptrace.c:1250]: trace_onreq_out(): retransmission
Oct 16 12:07:11 salmon /usr/bin/sip-proxy[26459]:
Henning Westerholt writes:
> tls: switch to TLSv1.2 in example cfg, usage of TLS 1.0 is not
> recommended anymore
How about TLSv1.3? baresip, for example, would like to use it. Isn't
there TLSv1.2+ value?
-- Juha
___
Kamailio (SER) - Development
README still needs clatifications. In Overview:
* by setting “trace_mode” to mirror all traffic.
=>
* by setting “trace_mode” to mirror all traffic if HEP mode is on.
or has this now changed in master?
3.18. trace_mode (integer)
If set to 1, the module uses core events triggered when
Daniel-Constantin Mierla writes:
> The db storage and hep mirroring were developed independently, not
> expecting to have them in pair always, but based on this discussion, it
> seems you would like some of the capabilities from one tracing type to
> be also in the other.
Then also trace_mode
Daniel-Constantin Mierla writes:
> given the changes done in the module to track the message, transaction
> or dialog, it should be fine to do this right now, because it is the
> testing period for 5.3, where it is supposed to clean up of issue and
> make coherent the new changes for this
Daniel-Constantin Mierla writes:
> There is a parameter trace_mode which can turn on automatic capture of
> all sip messages, but is for mirroring the traffic. Probably, for the
> next major release, it should be extended to do the save in the database
> as well, if that is wanted via some
Kamailio Dev writes:
> If set to 1, the module uses core events triggered when receiving or
> sending SIP traffic to mirror traffic to a SIP capture server using
> - HEP. It will automatically do the mirroring of all traffic, no need to
> - set the siptrace flag per request.
> +
Daniel-Constantin Mierla writes:
> The variable is used in my version of code from master branch. Can you
> check and see if you don't have a modified version you did to disable
> mysql connection ping?
That may have been the case. Sorry again. Now build works fine.
-- Juha
Juha Heinanen writes:
> For me it would be best and simplest if tracing of all request and
> responses would happen automatically when siptrace module is loaded and
> suitable parameters are set and when tracing is turned on via an rpc
> command, i.e., without any function calls i
Juha Heinanen writes:
> It would be nice if devs would try to compile the code after making changes.
I apologize. It was stupid statement.
-- Juha
___
Kamailio (SER) - Development Mailing List
sr-dev@lists.kamailio.org
https://lists.kamailio.org/
Daniel-Constantin Mierla writes:
> Can you try with latest master? I pushed a commit that hopefully fixes
> it.
It works now.
> On the other hand, the use of flag looks like no longer necessary if
> tracing mode parameter is set to 't'. Would it be acceptable to just add
> a new function
With latest master, I got this:
CC (gcc) [M db_mysql.so]km_dbase.o
km_dbase.c: In function 'db_mysql_submit_query':
km_dbase.c:71:9: warning: unused variable 't' [-Wunused-variable]
time_t t;
^
It would be nice if devs would try to compile the code after making
Daniel-Constantin Mierla writes:
> It is about storing in database? Or also mirroring?
Storing to db. Don't know about mirroring.
-- Juha
___
Kamailio (SER) - Development Mailing List
sr-dev@lists.kamailio.org
In 5.2
setflag(TRACE);
sip_trace();
traces both requests and responses. In master, only requests.
Please fix.
-- Juha
___
Kamailio (SER) - Development Mailing List
sr-dev@lists.kamailio.org
I build libmariadbclient18 with symbols (below) and got a new backtrace
of the hanging process.
I guess the only choice is to try with the newer lib that comes with
Debian Buster or give up use of db_cluster module with MySQL.
-- Juha
---start 24717
1 - 100 of 225 matches
Mail list logo
