Hi,
GWLIB's Connection module has a bug relating to how it handles large
data reads over SSLed connections.
The issue lies in the function unlocked_read() in conn.c, which is
called in several places to read from the underlying socket/file
descriptor. Typical usage in HTTPS transactions is as
Hi,
In bb_boxc.c, both smsbox_start() and wapbox_start() initialise
box_allow_ip/box_deny_ip. Which means if you have WAP and SMS running,
you leak a few bytes. Caught this using GCC LSAN. Quick fix was/is to
only initialise them if they are NULL.
P.
but don't know exactly when.
And for debugging just try to enable memory debug in configure or use
valgrind.
Thanks,
Alexander Malysh
Am 16.02.2011 um 14:33 schrieb Paul Bagyenda:
Hi guys,
I have a strange crash in my application that uses gwlib (from 1.4.2). Every
so often, we see
Hi guys,
I have a strange crash in my application that uses gwlib (from 1.4.2). Every
so often, we see:
PANIC: gwlib/thread.c:139: mutex_lock_real: Mutex failure! (Called from
gwlib/conn.c:202:lock_out.)
or
PANIC: gwlib/conn.c:200: lock_out: Assertion `gwthread_self() ==
Definitely an improvement IMO. And It does not break what we had before.
On Nov 11, 2009, at 18:36, Alexander Malysh wrote:
Nikos,
hmm seems you not reading anything just posting back? I come to this
conclusion
because I posted description of the issue together with a patch and
if you
True indeed.
Updated one (attached) does not.
x86_64-osx-2.patch
Description: Binary data
On Nov 09, 2009, at 12:30, Alexander Malysh wrote:
Hi,
this patch will break old OSX support and 32bit compilation.
Thanks,
Alexander Malysh
Am 06.11.2009 um 12:24 schrieb Paul Bagyenda:
Kannel
You want to look at the openss7 project at openss7.org. Or to be more
exact, watch it. As another person has written, SMS over SS7 means
implementing MAP. And TCAP. And SCCP. (Or the SIGTRAN equivalents.)
There are no open source implementations of these. Openss7.org has
them in the works,
concern weather the feature can be added
on kannel or not. And what are the advantages and disadvanges of the
above feature in comparision to SMPP.
Regards,
Manas
On Thu, Mar 26, 2009 at 2:30 PM, Paul Bagyenda
bagye...@dsmagic.com wrote:
You want to look at the openss7 project at openss7.org
date_parse_iso in gwlib/date.c is a little broken. It will not
correctly parse ISO dates of the form YYYMMDDTHHMMSS (e.g.
20090121T150501). It works just fine on -MM-DD...
The attached patch remedies this without loss of functionality
P.
date.c.diff
Description: Binary data
There is a need for a small change to the way multipart/related MIME
messages are parsed. Attached patch should help take care of issues
where the Content-ID for the start element is enclosed in '' ''
Kindly review, etc.
Thanks
P.
mime.diff
Description: Binary data
On Apr 12, 2007, at 13:51, Guillaume Cottenceau wrote:
Paul Bagyenda bagyenda 'at' dsmagic.com writes:
+ /* First, strip the if any. XXX some mime coders produce
such messiness! */
There's no evidence in the RFC it should be called a messiness..
Not sure I understand what you
it. */
use dict_remove instead.
9) please also supply userguide patch
Paul Bagyenda wrote:
Following your comments, here is a MO concatenation patch that now
resides in bb_smscconn.c, and uses the key as advised below. Same
code, largely, just integrated in a different place. There is also
better
Following your comments, here is a MO concatenation patch that now
resides in bb_smscconn.c, and uses the key as advised below. Same
code, largely, just integrated in a different place. There is also
better handling of the message stores...
Tests, comments, votes.
P.
Ps. If you are
On Jan 08, 2007, at 17:04, Stipe Tolj wrote:
Alexander Malysh wrote:
Hi Paul,
+1 for the at2 part.
-1 for the reassemble part. Your patch will never work reliable
because not only msisdn + refnum should be considered.
Reassemble should have triple as key: SMSC, msisdn, refnum.
IMO
I think Alex is talking about something different. You are talking
about MT concatenation.
To answer Alex's question: Yes I submitted the patch a while back.
AFAIK it is not yet on CVS. More votes in favour from you and others
might change that...
Paul.
On Jan 05, 2007, at 10:11, Enver
Hi,
There is a slight problem with the multipart parsing in Kannel's lib:
- when parsing an enclosed multipart element fails, the module crashes
- body elements that end in CRLF cause some problems.
Attached is a fix for this.
Cheers and happy new year.
mime.diff
Description: Binary data
Hi people,
Just came across a small discrepancy between how the SMPP module
decodes MO SMS UDH and how the AT2 module does the same. (To be
precise, somebody was testing the incoming concatenation patch I
posted earlier, and tripped on this.)
The short story is that the SMPP module
Hi,
There is a small but nasty bug in gwlib/mime.c which causes it to
crash when parsing multi-part messages. This is caused by free-ing
an Octstr and then returning a pointer to the same (which then gets
used)!
The small diff attached fixes the problem.
mime.diff
Description: Binary
Hi,
There is a minor problem with the parsing of quoted strings in wap/
wsp_headers.c -- some stacks (nokia/symbian 9 at least) use a literal
quote, which then gets missed by the parser. The net effect is that
the string returned still retains the quote (without the
corresponding end
If you poke around in Mbuni libs, there is code for doing HTTP POST
(enctype = www-url-encoded and multipart/form-data) handling nicely,
so you could use that too instead of XML. XML adds the over-head of
the other side having to parse that rather than merely have to handle
URL
This patch logs all the message parts received as it tries to
concatenate them. Run Kannel in full debug mode and pass on
kannel.log for inspection so we can see where the bug potentially is,
if any.
P.
On Sep 27, 2006, at 11:43, Boardmain wrote:
Hi, i used this patch
Hi All,
Responding to Alejandro's comment on the above subject (to quote:
IMHO this should be enabled/disabled from the config files.),
attached is the same patch for handling incoming concatenated
messages in smsbox. What I've done is add a boolean config variable
in the core group:
Hi,
Attached is a proposal for addressing the above issue. Stipe and I
have had this discussion for a while, and I finally got round to
doing it. The main driver for this has been the need for Mbuni MMS
Gateway (www.mbuni.org), which uses Kannel's libs, to be able to
receive MMS
Hi,
Attached is a new diff to update the mime libs somewhat, after my
last updates. This one reduces the amount of copying that goes on as
the interface passes objects back and forth
Cheers
Paul.
mime.diff
Description: Binary data
With subject change. Apologies
Hi,
Attached is a new diff to update the mime libs somewhat, after my
last updates. This one reduces the amount of copying that goes on
as the interface passes objects back and forth
Cheers
Paul.
mime.diff
Description: Binary data
Hi, Mbuni MMSC/MMS VAS Gateway makes heavy use of Kannel's GWLIB, which has resulted partly in some patches to Kannel from our side, to take care of missing stuff (or stuff we thought looked faulty). Many of these are now merged into CVS. One area we've been talking about for a while now
in HTTP applications. (There was a small problem with generation of
MIME boundaries.)
Cheers
Paul.
Edward C.A. Tromp Wrote:
Hi,
After some troubleshooting and with the help of both Paul Bagyenda
nad
Paul
Komkoff.
together we found out what the problem was with the OMA not working
Edward,
In gw/ota_prov.c there is a function ota_pack_udh() -- it sets the
UDH, which contains the source port. Try changing the source port to 23
f0 and see if that helps. I doubt the issue is with the HMAC
computation, since nokia seems to understand it fine, and the other
Paul's code seems
Take a look at Mbuni (http://www.mbuni.org) instead. It does what you
want.
On Mar 01, 2005, at 04:45, Stipe Tolj wrote:
Henry Wessel wrote:
Hi there,
sorry to bother all of you, I am new to the whole SMS gateway subject
and I am wondering whether there is a way to receive and send MMS via
the
Peter,
Hopefully this has been resolved with the new mbuni patch now at
http://www.mbuni.org/ ...
Thanks
On Feb 08, 2005, at 10:00, Peter Löfman wrote:
Hi list,
I have just compiled 1.4.0 trying to send some OTA configurations.
However, when I send the message smsbox crashes with this error:
in). We also patch the makefile
so that WAP libs (and headers) are installed/exposed. One interim
option might be to fold the patch into Kannel.
On Feb 08, 2005, at 05:49, Davy Chan wrote:
**Date: Mon, 07 Feb 2005 21:19:04 +0100
**From: Stipe Tolj [EMAIL PROTECTED]
**To: Paul Bagyenda [EMAIL
Hello All,
As well as putting the Mbuni MMS Gateway out, I have updated the
patches,
previously posted here, to enable Kannel send OTA settings using OMA
formats. Please see
mbuni.org for details.
Thank you
Something that was mentioned earlier but is still not fixed in v1.3.2:
In gw/smsbox.c function obey_request() -- when 'type' is TRANS_FILE or
TRANS_EXECUTE I suggest that a call to:
octstr_url_decode(pattern);
be inserted just after the case statement. It must be causing some
unnecessary
]
To: [EMAIL PROTECTED]
Guys,
I'm planning to use OMA-OTA as a provision mechanism for phones that
support
this feature. I've search on the mailing list and saw that Paul
Bagyenda did
do some work regarding this. I was trying to add his patch to the
current
CVS but with no succes i think i need the cvs
On 12.3.2004, at 15:51, Stipe Tolj wrote:
Paul Bagyenda schrieb:
Attached are some diffs to implement OMA-style sending of OTA
settings.
Refer to OMA-WAP-ProvCont-v1_1-20021112-C (at
http://www.openmobilealliance.org/tech/docs/) for details on how this
works.
The changes include all (most) the HMAC
Hi,
I have located a phone that should work, so I can test and finalise
before you begin working on integration. Will get back to you.
P.
-
Paul A. Bagyenda
Digital Solutions
P.O.Box 23833
58 Bukoto Str.
Kampala, UGANDA
Tel: +256-31-262117
Mob: +256-77-507743
40 or 60 Nokia (6800 and above, 6600
and above), SonyEricsson P900 or some of the new Motorolla phones to
test with. I have none of these! I had assumed all Series 60 support
this, but nada.
Let me know if you need more info.
P.
On 12 Mar 2004, at 17:51, Stipe Tolj wrote:
Paul Bagyenda
Attached are some diffs to implement OMA-style sending of OTA settings.
Refer to OMA-WAP-ProvCont-v1_1-20021112-C (at
http://www.openmobilealliance.org/tech/docs/) for details on how this
works.
The changes include all (most) the HMAC computation stuff required for
authentication. What I've
to test with one of the supported phones and provide some
feedback?
Thanks
On 13 Feb 2004, at 18:16, Paul Bagyenda wrote:
Attached are some diffs to implement OMA-style sending of OTA
settings. Refer to OMA-WAP-ProvCont-v1_1-20021112-C (at
http://www.openmobilealliance.org/tech/docs
Hi,
sms box now supports url as well as text and exec methods for
generating responses to sms. However the use of parameters for the exec
cases is somewhat flawed: smsbox performs url encoding on the command
parameters, which is probably undesirable for most applications. A
quick but dirty
Hi,
Over time I've made some changes to some of the kannel libraries that
I thought I might contribute for inclusion into the current CVS.
Attached is a unified diff of these against the latest CVS.
Short explanation:
1. date.c - The function for making iso dates seems to have not been
41 matches
Mail list logo