| Summary: | ASTERISK-05548: Choppy Audio When Using > 20 MS Audio Frames | ||
| Reporter: | Chih-Wei Huang (cwhuang) | Labels: | |
| Date Opened: | 2005-11-10 03:36:14.000-0600 | Date Closed: | 2011-06-07 14:10:45 |
| Priority: | Minor | Regression? | No |
| Status: | Closed/Complete | Components: | Applications/app_meetme |
| Versions: | Frequency of Occurrence | ||
| Related Issues: | |||
| Environment: | Attachments: | ( 0) app_meetme_dyna_buff.diff ( 1) meetme-dyna-buff.diff ( 2) write.diff ( 3) write2.diff | |
| Description: | When some participants use G.723.1 codec in a conference, the audio quality becomes bad. If participant A uses G.723.1 while B use other codec (e.g. G.729/alaw), A can hear clear sound from B, but B hears very choppy sound from A. The problem was independent of the channel types. (I have tested SIP, H.323 and MGCP) If nobody use G.723.1, the audio quality is very good. The Asterisk version under test is 1.2.0-rc1 and today CVS-HEAD. The problem doesn't occur in 1.2-beta2 or prior. So I believe it it related to recently fixed of delay problems for app_meetme. BTW, the G.723.1 and G.729 codec are from Intel IPP library. | ||
| Comments: | By: Tilghman Lesher (tilghman) 2005-11-10 08:35:12.000-0600 The G.723.1 codec is not distributed with Asterisk, as it is under patent protection. We cannot help you with this issue. I suggest that you contact whoever licensed this code to you. By: Chih-Wei Huang (cwhuang) 2005-11-10 20:30:00.000-0600 I feel very sad about your comment. The codec I used has no problem. Asterisk CVS-HEAD works with it very fine, unless in a conference of Meetme. As I said, version before 1.2-beta2 (included) also works with the codec without any problem. So it MUST BE a problem of Asterisk CVS-HEAD itself, very possiblely in recently changes of app_meetme. The fact is, Asterisk CVS-HEAD doesn't work with some codecs, includes G.723.1 but not limites to it. You should at least take a look at the problem instead of just ignoring it. By: Kevin P. Fleming (kpfleming) 2005-11-10 20:43:38.000-0600 How can we look at the problem? The codec involved is not something we distribute or maintain, and we cannot provide any patches or fixes to its source code. In addition, it is illegal to use it anywhere that the Intel 'non-commercial use' license applies (disregarding the patent protection), so none of us can legally run it even for testing. You will need to take this up with the parties that provided the codec to you. By: Chih-Wei Huang (cwhuang) 2005-11-11 04:48:08.000-0600 I don't understand you. Do you mean even for testing is NOT a way of 'non-commercial use'? Anyway, I have found a way to fix it. As I said, the fix is in Asterisk itself (app_meetme.c), not the codec. Since you are not interesting in fixing the problem, I think I don't have to submit it. By: Serge Vecher (serge-v) 2005-11-11 09:24:18.000-0600 cwhuang, can you please post your patch to meetme for the benefit of the community? Also, please confirm if you have disclaimer on file. By: Kevin P. Fleming (kpfleming) 2005-11-11 11:16:34.000-0600 Who said we were not interested in fixing the problem? Why does everyone have to be so dramatic about things... I only said that there was nothing we could to help you find and solve the problem, since we can't use the codec. And to answer your question: my job is working on Asterisk, so any testing that I do on Asterisk is by definition 'commercial use'. Many other community volunteers are in the same situation so it's much safer for us to not risk any problems by avoiding the situation. Now, if you have determined that the fault here is in app_meetme, we will gladly consider your solution and determine whether it is the correct solution or not. If the patch is small (a few lines), then there isn't even a need for a disclaimer, so please post it so we can help you. By: damin (damin) 2005-11-12 11:24:01.000-0600 I would second a request for you to post the patch to app_meetme. I am NOT a commercially employed developer of Asterisk, so I do a lot of testing w/ a variety of equipment and codecs. I would be happy to review your patch and test it against the Digium distributed g729 binaries to ensure that it works with both of them. Please consider posting the patch for the benefit of the community. By: Chih-Wei Huang (cwhuang) 2005-11-16 21:36:06.000-0600 OK, I uploaded the patch. I compared the CVS and 1.2-beta2, and found there is a change careful_write() => write() in about line 1307. There is a comment that said it should not use careful_write here, but I decided to give a try. Magically, it works great! So my patch is simple. If the channel native format is G.723.1, use careful_write(), other use write(). By: Chih-Wei Huang (cwhuang) 2005-11-16 21:54:00.000-0600 I studied further why careful_write() works better than write(). There are some zaptel stuffs that I don't understand. But basically careful_write() does one thing more than write(): it flushes the buffer by a while loop. By adding more debug messages, I found reason. The datalen of a G.723.1 frame is 320, while other codecs (e.g., G.729) are just 160. I don't know why, maybe someone can give an explanation. A call of write() just writes 160 bytes. That means half data of a G.723.1 frame are not written. That explains all the situations I mentioned above. So I just copied careful_write() to flush the buffer, removed zaptel stuffs. It seems to work great! No significant delay in one hour of conversation. This is write2.diff. By: damin (damin) 2005-11-16 22:59:44.000-0600 g723.1 uses (a minimum of) 30 ms frames, hence the increased size of the data. By: Tony Mountifield (softins) 2005-11-18 06:26:15.000-0600 I have just discovered the same issue with the change from careful_write() to write() to the pseudo-fd. And I am using the aLaw codec. My SIP provider's switch sends the RTP stream to us in 60ms frames, not 20ms. This means that every ast_frame received by meetme has 960 bytes of data. When the 960 bytes are passed to write(), it returns indicating that only 320 bytes were written. Reverting to careful_write() restored the audio quality by ensuring that all data was written. What is the purpose of the ioctl with ZT_IOMUX_WRITE | ZT_IOMUX_SIGEVENT inside careful_write()? By: Tony Mountifield (softins) 2005-11-18 06:28:34.000-0600 I meant to ask if a bug marshall could change the title of this bug, since it is not specific to the G.723.1 codec - it occurs with any source that provides frames larger than 20ms. By: damin (damin) 2005-11-18 08:27:37.000-0600 Bug title updated. Please verify that this is a more accurate description. By: Tony Mountifield (softins) 2005-11-18 08:42:06.000-0600 Looks fine - thanks! By: BJ Weschke (bweschke) 2005-11-18 17:06:11.000-0600 I think we need to understand why the write isn't writing the whole frame and address that issue, because while this may fix this issue, I think it will bring back the problem of audio delay. Is it because the zap mux is only accepting slinear chunks that are 20ms in size and that's it? By doing careful_write or looping to insure that all frames in are written back out, we are now back to introducing more audio into the zap mux then there is time for when doing name announces and the delay recently fixed with this change will be back again. By: Tony Mountifield (softins) 2005-11-19 03:36:25.000-0600 I agree it would be good to understand why the pseudo is only accepting 20ms of data at a time. However.... When the pseudo-fd is opened, it is set to non-blocking. This should mean that if there is no buffer available for a write(), it returns EAGAIN (=EWOULDBLOCK) instead of waiting (we should verify that this is the case). Note that both careful_write() and cwhuang's inline loop break out when EAGAIN is returned. So there should be no difference between using careful_write() and having the pseudo accept the whole buffer in one go. I'd still like someone to explain the purpose of the ioctl in careful_write(). In my experience, any delay introduced by playing announcements to the conf chan are best eliminated by the threaded approach I proposed in bug ASTERISK-3519, and the gradually increasing delay is cured by the async patch in ASTERISK-5230. This is while still retaining careful_write() in the conf loop. By: dea (dea) 2005-11-19 18:38:50.000-0600 I regularly use 30ms frames in MeetMe with no issues. Courtesy of bugid 0005162. I have tested 20,30,40 and 60ms frames with ulaw and G729 to and from chan_sip and chan_ooh323. Quality is each case is excellent. This system also has the asynch patch applied. No Zaptel hardware, just ztdummy on 2.6.14 kernel with RTC support. The server even has a poor RTC, with zttest reporting --- Results after 47 passes --- Best: 99.963379 -- Worst: 99.938965 -- Average: 99.957665 and quality still beats an obscenely expensive propriatary solution I amd trying to replace. ***Completely bogus post. The above was true with beta1, not the release*** By: Tony Mountifield (softins) 2005-11-20 06:50:08.000-0600 DEA: please could you confirm what version of app_meetme.c you are using with no issues? The issue in question here (choppy audio) would occur with 1.125 or later of app_meetme.c when using >20ms frames. By: dea (dea) 2005-11-20 12:19:18.000-0600 I am running 1.2.0 release, which includes app_meetme.c version 1.126. Usage on the server has been light since the upgrade, so I'll schedule a test on Monday to generate a load to see if it falls over at some point. By: Tony Mountifield (softins) 2005-11-20 12:25:50.000-0600 DEA: perhaps the patch in ASTERISK-5026 is always delivering to meetme in 20ms frames. If possible, could you test >20ms frames with patch 5162 removed? By: dea (dea) 2005-11-20 18:11:06.000-0600 I don't think 5162 is only sending 20ms to meetme. I am using Cisco phones and 'rtp debug' from the Asterisk cli. The phones show that they are receiving 30ms, and the cli output shows 30ms in each rtp packet. By: Chih-Wei Huang (cwhuang) 2005-11-21 03:04:19.000-0600 First, I'm glad to see the problem has been confirmed to be a general problem instead of for a particular patented codec. I agree with bweschke that we should figure out why and when not whole frame are written. DEA: Could you modify the following code to see how many data are really read and written? (app_meetme.c about line 1307) int res = write(fd, f->data, f->datalen); ast_log(LOG_NOTICE, "datalen=%d, written %d byte(s)", datalen, res); By: dea (dea) 2005-11-21 11:39:06.000-0600 Odd. The results do show that only 20ms is being passed. Nov 21 09:36:53 NOTICE[2265]: app_meetme.c:1318 conf_run: datalen=480, written 320 byte Yet I have not had any noticable quality issues, which with 33% of the audio being either delayed until the next write, or worse dropped, I would expect to hear something 'bad'. By: dea (dea) 2005-11-21 13:55:57.000-0600 OK, I maybe be out in left field, but poking at this to try to understand. In conf_run at line 993(ish): bi.bufsize = CONF_SIZE/2; Line 206: #define CONF_SIZE 320 I set CONF_SIZE to 640 and now the debugging step cwhuang suggested shows the write() call is handling all of the data. So, is CONF_SIZE too small or inflexible (inregards to variable frame sizes)? By: dea (dea) 2005-11-21 23:12:51.000-0600 I think I got it. In app_meetme starting near line 992 there is an ioctl that setup the buffers in the zaptel driver. The current code divides CONF_SIZE (default 320) by 2 and passes is and numbufs (default 32) to the pseudo fd. Inside zaptel, ZT_SET_BUFINFO calls zt_reallocbufs to allocate memory for the buffers. With the defaults, zaptel allocates (320/2 * 2 * 32) /*(SIZE_CONF/2) * 2 * (numbufs)*/ then divides the memory up into 32 readbuf and 32 writebuf, each half the size of CONF_SIZE. That's as far as I could decipher. Logic tells me that there must be a loop to handle reading two buffers (320 out however many bytes passed), but I cannot find it. Increasing SIZE_CONF might not be the correct way, but it creates buffers big enough to handle greater than 20ms. The only thing that makes sense is if somehow the numbufs gets reset to the default of 2, but again I don't see how. By: jht (jht) 2005-11-22 11:19:10.000-0600 Just another data point: I'm running Asterisk 1.2.0-rc2 Meetme audio quality was very poor for calls from Linksys ATA user, but fine for other callers. As suggested by DEA I changed in app_meetme.c: #define CONF_SIZE 320 to #define CONF_SIZE 640 That fixed the quality which is now good for all callers. By: dea (dea) 2005-11-22 11:38:11.000-0600 I think that adjusting CONF_SIZE is a short-term fix. I still do not fully understand the code, but it is pretty clear that the problem is that meetme assumes that it will only handle 20ms of audio. Not only in how it sets up buffers. Setting CONF_SIZE to 960 should fix softins issue, but would waste memory for most uses. Dynamically resizing the buffers when a new caller joins and has a larger payload might be possible. Another option would to remove the constants that limit meetme to two samples per read/write based on the payload. Or the easiest, and quickest would be to make a config option of CONF_SIZE and put it in meetme.conf to give people a choice to work-around it. By: dea (dea) 2005-11-23 13:07:46.000-0600 Badness all around. Adjusting CONF_SIZE only works if ALL participants are using the same codec payload. A caller with a 20ms payload will get poor audio if CONF_SIZE is tuned for 30ms (480). I implimented a quick and dirty hack to resize the buffers if a payload does not match the default. I have a number of unrelated hacks in app_meetme, so here is the crux of the changes: In conf_run, around line 830- int dtmf + int dyna_buff = CONF_SIZE; + char **buf = NULL; + size_t rbufsize = CONF_SIZE + AST_FRIENDLY_OFFSET; ZT_BUFFERINFO bi; around line 995- memset(&bi, 0, sizeof(bi)); - bi.bufsize = CONF_SIZE/2; + bi.bufsize = dyna_buff/2; bi.txbufpolicy = ZT_POLICY_IMMEDIATE; around line 1295- f = ast_read(c); if (!f) break; + /* If we get more than 20ms of audio, resize the buffers for this participant */ + if ( f->datalen != dyna_buff){ + dyna_buff = f->datalen; + goto zapretry; + } around line 1509- } else if (outfd > -1) { + if (!buf || (rbufsize < dyna_buff) ){ + ast_log(LOG_NOTICE, "Read Buffer size: %d Write Buffer size: %d\n", rbufsize, dyna_buff); + rbufsize = dyna_buff + AST_FRIENDLY_OFFSET - (dyna_buff % AST_FRIENDLY_OFFSET); + buf = realloc(buf, (rbufsize+1)*sizeof(char *)); + } - res = read(outfd, buf, CONF_SIZE); + res = read(outfd, buf, dyna_buff); if (res > 0) { If someone can test these small changes and confirm success, I'll update a clean copy of app_meetme.c and post a proper patch. **Edit -forgot to change the read buffer By: Chih-Wei Huang (cwhuang) 2005-11-24 00:15:43.000-0600 DEA: I believe your changes doesn't work because of res = read(outfd, buf, dyna_buff); but buf is a fixed length buffer: char __buf[CONF_SIZE + AST_FRIENDLY_OFFSET]; char *buf = __buf + AST_FRIENDLY_OFFSET; A quick test confirmed my suspect. Asterisk core dumpped immediately. I have a phone with datelen=960: conf_run: datalen=960, written 960 byte(s) By: dea (dea) 2005-11-24 01:44:28.000-0600 Good catch. I needed to track that one down, as I have a wireless phone that is not smooth in it's delivery, and two 30ms 729 frames will be delivered at the same time. So, either buf needs to be a much bigger buffer, or replace the static buffer with one malloc'd to the size needed. Another bug with the concept is that VAD can trigger a resize to 0, which is obviously a bad thing. I've fixed that one already, and will look at *buf tomorrow. I'd be interested in your findings on smaller audio samples. I am working with 30ms mostly, with a little 20ms in the mix. The results so far have been quite good. By: dea (dea) 2005-11-24 18:26:54.000-0600 The read buffer is now dynamically sized. No more segfaults with 60ms audio samples. I think the code is ready for me to submit a real patch. One weakness of the current code is that it does nothing to correct app_meetme's assumption that it has two samples of 10ms. I simply split the audio in half and resize the buffers. I think the proper final solution will be to determine the correct number of 10ms samples. Not so much for app_meetme, but the mux in zaptel. By: Geoffrey Cleaves (gcleaves) 2005-11-25 04:31:38.000-0600 Ole ole! Looking forward to your patch DEA, as this issue has kept me from upgrading to 1.2 . By: Anthony Minessale (anthm) 2005-11-25 11:27:49.000-0600 You can use ilbc to debug this too since it also has 30 ms frames ignoring the problem because the example was g723.1 is not neccessary. The reporter clearly demonstrates a real problem and shunning him is not the proper way to presuade him to report any new issues he may encounter. The bigger framesize in RTP would not show this problem unless you are also using a 30ms codec since it's just a series of frames in 1 packet. The bottom line is in this environment where you are having a problem, you have many channels all possibly using different codecs trying to syncronize thier audio. in order to succeed you must normalze the data to 8khz raw signed linear. you can count on this being done because the app sets the channels read/write to slinear which means regardless of the actual codec you will always read/write uncompressed frames. the only thing these frames have in common at this point is the samples len which is a value that tells you the length in time the frame represents both in compressed and uncompressed formats. Everything in asterisk operates at 8000hz (8000 samples per second) which is limiting but true. because you know the speed and you are sure all channels are using raw audio the only difference you have to consider is the duration of the frames. Zap is going to be 20ms per frame which is 160 samples or 320 bytes of data. When you introduce a channel that is sending 30ms you will get 480 bytes every 30ms from the remote end. Again remember the actual codec is irrelivant here it's all raw now. since the zap channel wants to get 320 bytes every 20 ms it will make the audio choppy because on a timeline at the 20ms mark the channel is ready for 320 bytes but it will be dissappointed to find that the audio will arrive 10ms late and be too much at that. 10ms later now 40ms on the timeline it has the extra 160 bytes already from the previous packet but that is 160 too small and it now must wait again (hense the gaps) the best solution is to keep the data in a buffer and only write frames in the excact size it expects. recv 480, write 320, save the other 160 in a buffer, read 480 more, now you have 640 which is enuf for 2 320 byte frames so write them both which will make up for the lost time not sending the extra 160 the first time. Just FYI By: Tony Mountifield (softins) 2005-11-25 11:53:40.000-0600 anthm: compressed codecs no doubt work on individual frames, but alaw and ulaw are a per-sample mapping without fixed size frames. My SIP provider sends me 60ms of alaw at a time in RTP, and each 60ms arrives in meetme as a single ast_frame with 960 bytes of slin data. Why buffer audio within meetme, when it can be done in zaptel? We just need to be sure that zaptel accepts data immediately if it can, and returns EAGAIN without blocking if it can't. If that currently happens, then careful_write() should not cause a problem. Even better would be for zt_write to accept all the data given to it, if it can, rather than just accepting 20ms worth. I'm not convinced DEA's dynamic buffering is the correct solution, but haven't had time to examine stuff in detail before commenting. By: dea (dea) 2005-11-25 12:39:18.000-0600 I don't have access to G.723.1, but as Anthm points out any codec using more than 20ms exhibits this problem. The reason I choose a dynamic buffer approach, other than being a complete neophyte in regards to the Asterisk internals, is that app_meetme issues an ioctl to set the Zaptel buffers. To the untrained eye that made it look like perhaps Zaptel would/could deal with increased buffer sizes. With the holiday I am only able to do basic testing by myself. I have tried ulaw and G.729(licensed version) up to 60ms, over IAX2 and chan_ooh323. I also threw a copy of Firefly into the mix, using SIP, only to note that regardless of the codec choosen, it sends 33ms(GSM?) of audio per packet. The functions write() and read() don't seem to care about the sample size, as long as the buffers are appropriately sized. By: Anthony Minessale (anthm) 2005-11-25 13:55:02.000-0600 I just thought I would tell you what I *think* the problem is based on the info you provided in the bug report as a gesture to try and save you some headaches. I didnt say you needed to make a new buffer I just tried to describe how to solve the problem with a buffer so if you can make zaptel do it, that's fine. =D By: dea (dea) 2005-11-25 14:45:07.000-0600 Dynamic buffer patch uploaded. It does not currently shrink the read buffer, but would be trivial to do so. Anthm- I do appreciate the advice. I had already worked this patch out, and it seems to work, so the headache had come and gone. Disclaimer on file. By: Chih-Wei Huang (cwhuang) 2005-11-27 21:15:39.000-0600 DEA: First, do you free the dynamically allocated buffer? I didn't see it. Second, why buf is defined to be char** and allocates size of memory XXX*sizeof(char*)? I don't think the date type to be read are char*. Am I missed something? By: dea (dea) 2005-11-27 21:56:55.000-0600 I used the alloca() function which releases memory automatically when exiting the function it was called from. I got the idea to use it from other instances of buffer code in Asterisk. The original buffer code was a char* sized for an array of chars. When looking over other buffer code in Asterisk, most seemed to be char**, so that's what I used. I made sure to test a variety of packetization settings with SIP, H323 and IAX(20ms only obviously). No segfaults in the case of 60ms of audio, and quality seemed good to very good, but the quality tests were not very scientific. By: Chih-Wei Huang (cwhuang) 2005-11-28 01:49:24.000-0600 I don't see where you used alloca(), except this line /* buf = alloca((rbufsize)*sizeof(char *)); */ But it is a comment. Doesn't it? Besides, if you used alloca(), you cannot use realloc to adjust the buffer size. alloca() is very different than the standard allocate function malloc/calloc/realloc since normally gcc translates calls to alloca() by inlined code(at least in Linux). See man alloca. Also note the manpage said it's discouraged to use alloca() since on many systems its implementation is buggy. I think f->datalen and your dyna_buff are calculated in size of bytes intead of size of pointers to byte(char*). Surely you won't have segfaults since you allocated a buffer approximately quad-size than necessary (sizeof(char*)=4 in 32-bit CPU). In fact, I do think dynamically allocated buffer for this problem is unnecessary. It just introduces superfluous complexities to the problem. Why not just use a local buffer that is large enough, say 500ms or even 1000ms, to cover almost all cases? By: dea (dea) 2005-11-28 02:07:54.000-0600 Yout correct, I planned to use alloca, but had problems with realloc when allocated with alloca. I found that fixed size buffers cause problems during write. If the buffer was larger than needed it introduced quality problems. The read buffer could be statically sized, I think, but I'm not positive that wouldn't also introduce quality degradation. cwhuang, I am travelling this week, would you be able to test the large static buffer and changing the realloc to rbufsize*sizeof(char)? What are your thoughts about the dynamic write buffer sizing? By: Chih-Wei Huang (cwhuang) 2005-11-30 04:48:20.000-0600 I uploaded a patch modified from DEA's. The differences are: * Use a local buffer with size about ZT_MAX_BUF_SPACE / ZT_DEFAULT_NUM_BUFS. This is the maximim size that the zaptel interface allows. * Add a check to skip too large frame. Sometime I got a very large frame, say 2400 or 1920, that would cause IOCTL() to return EINVAL. I don't know why. I've tested my patch with frame size 160, 480, 960. So far so good. :) By: dea (dea) 2005-12-02 19:16:58.000-0600 One comment about the static buffer size- ZT_MAX_BUF_SPACE / ZT_DEFAULT_NUM_BUFS = 1024 AST_FRIENDLY_OFFSET = 64 I believe that the original intent of AST_FRIENDLY_OFFSET was to ensure memory alignement, and is not needed if you statically assign the buffer to 1024. The dynamic resize code considered that and adjusted the offset depending on the target buffer size. By: Mark Spencer (markster) 2005-12-20 03:27:27.000-0600 Fixed in SVN trunk ASTERISK-7347 the "right way"... By: dea (dea) 2005-12-20 04:07:25.000-0600 For educational purposes, can you elaborate on what the 'Right Way' is? I'm not going to be able to check the SVN repository for some time (I know, not your problem), and a short dscription of the implimented change would be enlightening By: Tony Mountifield (softins) 2005-12-20 04:15:45.000-0600 From reading the svn-commits posting from the change (svn-commits is a good list to subscribe to!), the "right way" is apparently to revert to careful_write(), but passing it a parameter to tell it to skip calling the ioctl(). This is the ioctl(), the purpose of which I asked about further up this bug, but nobody bothered to explain. From Mark's comment in the code, it apparently causes the write() withing careful_write() to block, but I'm not sure why. By: Russell Bryant (russell) 2005-12-20 10:19:52.000-0600 I'm moving this from pending stable to closed since kpfleming has already backported this into 1.2. If anyone has any more questions about the code, the best place would be the asterisk-dev mailing list or the #asterisk-dev IRC channel. Thanks! By: Digium Subversion (svnbot) 2008-01-15 16:07:39.000-0600 Repository: asterisk Revision: 7542 U trunk/apps/app_meetme.c ------------------------------------------------------------------------ r7542 | markster | 2008-01-15 16:07:38 -0600 (Tue, 15 Jan 2008) | 2 lines Fix choppy audio with > 20ms audio frames (bug ASTERISK-5548) ------------------------------------------------------------------------ http://svn.digium.com/view/asterisk?view=rev&revision=7542 By: Digium Subversion (svnbot) 2008-01-15 16:07:42.000-0600 Repository: asterisk Revision: 7546 U branches/1.2/apps/app_meetme.c ------------------------------------------------------------------------ r7546 | kpfleming | 2008-01-15 16:07:41 -0600 (Tue, 15 Jan 2008) | 2 lines backport fix for larger-than-20ms-frames from trunk (bug ASTERISK-5548) ------------------------------------------------------------------------ http://svn.digium.com/view/asterisk?view=rev&revision=7546 By: Digium Subversion (svnbot) 2008-01-15 16:07:49.000-0600 Repository: asterisk Revision: 7553 _U team/russell/agent_function/ U team/russell/agent_function/apps/app_chanspy.c U team/russell/agent_function/apps/app_dial.c U team/russell/agent_function/apps/app_directed_pickup.c U team/russell/agent_function/apps/app_meetme.c U team/russell/agent_function/build_tools/make_svn_branch_name U team/russell/agent_function/cdr.c U team/russell/agent_function/channel.c U team/russell/agent_function/channels/Makefile U team/russell/agent_function/channels/chan_iax2.c U team/russell/agent_function/channels/chan_mgcp.c U team/russell/agent_function/channels/chan_sip.c U team/russell/agent_function/codecs/Makefile U team/russell/agent_function/codecs/codec_gsm.c U team/russell/agent_function/doc/README.ael U team/russell/agent_function/file.c U team/russell/agent_function/frame.c U team/russell/agent_function/include/asterisk/channel.h U team/russell/agent_function/include/asterisk/frame.h U team/russell/agent_function/include/asterisk/linkedlists.h U team/russell/agent_function/include/asterisk/rtp.h U team/russell/agent_function/rtp.c ------------------------------------------------------------------------ r7553 | russell | 2008-01-15 16:07:49 -0600 (Tue, 15 Jan 2008) | 129 lines Merged revisions 7509,7511-7512,7514,7516,7518,7520,7522,7524,7528,7538-7542,7547-7548,7551 via svnmerge from https://origsvn.digium.com/svn/asterisk/trunk ................ r7509 | tilghman | 2005-12-16 20:07:44 -0500 (Fri, 16 Dec 2005) | 3 lines Merged revisions 7508 via svnmerge from /branches/1.2 ................ r7511 | kpfleming | 2005-12-16 21:21:36 -0500 (Fri, 16 Dec 2005) | 2 lines block a commit to a module that no longer exists in trunk ................ r7512 | kpfleming | 2005-12-16 21:22:24 -0500 (Fri, 16 Dec 2005) | 10 lines Merged revision 7510 via svnmerge from https://origsvn.digium.com/svn/asterisk/branches/1.2 ........ r7510 | kpfleming | 2005-12-16 20:20:04 -0600 (Fri, 16 Dec 2005) | 2 lines fix some buglet when building team branch version strings ........ ................ r7514 | kpfleming | 2005-12-16 22:45:25 -0500 (Fri, 16 Dec 2005) | 10 lines Merged revisions 7513 via svnmerge from https://origsvn.digium.com/svn/asterisk/branches/1.2 ........ r7513 | kpfleming | 2005-12-16 21:44:30 -0600 (Fri, 16 Dec 2005) | 2 lines forcibly expire previous subscriptions from a peer when they resubscribe (keeps them from building up and waiting for expiration, and stops us sending unwanted NOTIFY messages to devices) ........ ................ r7516 | kpfleming | 2005-12-16 22:59:27 -0500 (Fri, 16 Dec 2005) | 10 lines Merged revisions 7515 via svnmerge from https://origsvn.digium.com/svn/asterisk/branches/1.2 ........ r7515 | kpfleming | 2005-12-16 21:59:05 -0600 (Fri, 16 Dec 2005) | 2 lines Max-Forwards headers must only be present on requests, not responses ........ ................ r7518 | tilghman | 2005-12-17 12:22:24 -0500 (Sat, 17 Dec 2005) | 3 lines Merged revisions 7517 via svnmerge from /branches/1.2 ................ r7520 | tilghman | 2005-12-17 13:58:57 -0500 (Sat, 17 Dec 2005) | 3 lines Merged revisions 7519 via svnmerge from /branches/1.2 ................ r7522 | tilghman | 2005-12-19 00:42:55 -0500 (Mon, 19 Dec 2005) | 3 lines Merged revisions 7521 via svnmerge from /branches/1.2 ................ r7524 | tilghman | 2005-12-19 14:08:42 -0500 (Mon, 19 Dec 2005) | 3 lines Merged revisions 7523 via svnmerge from /branches/1.2 ................ r7528 | russell | 2005-12-19 18:41:53 -0500 (Mon, 19 Dec 2005) | 4 lines - add note on required values of sip_methods struct - remove duplicate function prototype - remove duplicate ast_mutex_lock (issue ASTERISK-5866) ................ r7538 | russell | 2005-12-20 02:45:05 -0500 (Tue, 20 Dec 2005) | 3 lines allow forcing the build to exclude PRI support using WITHOUT_PRI, similar to how we already have WITHOUT_ZAPTEL (issue ASTERISK-5829) ................ r7539 | russell | 2005-12-20 03:16:53 -0500 (Tue, 20 Dec 2005) | 2 lines use the system libgsm if available (issue ASTERISK-5287, modified to still work with builtin libgsm) ................ r7540 | markster | 2005-12-20 04:39:31 -0500 (Tue, 20 Dec 2005) | 2 lines Fix reload of peer contexts (bug ASTERISK-5848) ................ r7541 | markster | 2005-12-20 04:56:55 -0500 (Tue, 20 Dec 2005) | 2 lines Fix segfault on directed pickup when no CDR is available (bug ASTERISK-5839) ................ r7542 | markster | 2005-12-20 05:26:53 -0500 (Tue, 20 Dec 2005) | 2 lines Fix choppy audio with > 20ms audio frames (bug ASTERISK-5548) ................ r7547 | markster | 2005-12-20 08:07:02 -0500 (Tue, 20 Dec 2005) | 3 lines Major peformance improvements to meetme ................ r7548 | markster | 2005-12-20 09:14:01 -0500 (Tue, 20 Dec 2005) | 2 lines Don't bother decode on muted participants ................ r7551 | markster | 2005-12-20 12:52:31 -0500 (Tue, 20 Dec 2005) | 3 lines Major RTP fixes for using inbound SDP on outbound connection, get rid of old local rtp stuff... ................ ------------------------------------------------------------------------ http://svn.digium.com/view/asterisk?view=rev&revision=7553 By: Digium Subversion (svnbot) 2008-01-15 16:07:59.000-0600 Repository: asterisk Revision: 7564 _U team/crichter/0.3.0/ U team/crichter/0.3.0/apps/app_chanspy.c U team/crichter/0.3.0/apps/app_dial.c U team/crichter/0.3.0/apps/app_directed_pickup.c U team/crichter/0.3.0/apps/app_meetme.c U team/crichter/0.3.0/build_tools/make_svn_branch_name U team/crichter/0.3.0/cdr.c U team/crichter/0.3.0/channel.c U team/crichter/0.3.0/channels/Makefile U team/crichter/0.3.0/channels/chan_agent.c U team/crichter/0.3.0/channels/chan_iax2.c U team/crichter/0.3.0/channels/chan_mgcp.c U team/crichter/0.3.0/channels/chan_sip.c U team/crichter/0.3.0/cli.c U team/crichter/0.3.0/codecs/Makefile U team/crichter/0.3.0/codecs/codec_gsm.c U team/crichter/0.3.0/doc/README.ael U team/crichter/0.3.0/file.c U team/crichter/0.3.0/frame.c U team/crichter/0.3.0/include/asterisk/channel.h U team/crichter/0.3.0/include/asterisk/frame.h U team/crichter/0.3.0/include/asterisk/linkedlists.h U team/crichter/0.3.0/include/asterisk/rtp.h U team/crichter/0.3.0/include/asterisk/strings.h U team/crichter/0.3.0/res/res_agi.c U team/crichter/0.3.0/rtp.c U team/crichter/0.3.0/utils.c ------------------------------------------------------------------------ r7564 | crichter | 2008-01-15 16:07:58 -0600 (Tue, 15 Jan 2008) | 144 lines Merged revisions 7509,7511-7512,7514,7516,7518,7520,7522,7524,7528,7538-7542,7547-7548,7551,7554,7556 via svnmerge from https://origsvn.digium.com/svn/asterisk/trunk ................ r7509 | tilghman | 2005-12-17 02:07:44 +0100 (Sa, 17 Dez 2005) | 3 lines Merged revisions 7508 via svnmerge from /branches/1.2 ................ r7511 | kpfleming | 2005-12-17 03:21:36 +0100 (Sa, 17 Dez 2005) | 2 lines block a commit to a module that no longer exists in trunk ................ r7512 | kpfleming | 2005-12-17 03:22:24 +0100 (Sa, 17 Dez 2005) | 10 lines Merged revision 7510 via svnmerge from https://origsvn.digium.com/svn/asterisk/branches/1.2 ........ r7510 | kpfleming | 2005-12-16 20:20:04 -0600 (Fri, 16 Dec 2005) | 2 lines fix some buglet when building team branch version strings ........ ................ r7514 | kpfleming | 2005-12-17 04:45:25 +0100 (Sa, 17 Dez 2005) | 10 lines Merged revisions 7513 via svnmerge from https://origsvn.digium.com/svn/asterisk/branches/1.2 ........ r7513 | kpfleming | 2005-12-16 21:44:30 -0600 (Fri, 16 Dec 2005) | 2 lines forcibly expire previous subscriptions from a peer when they resubscribe (keeps them from building up and waiting for expiration, and stops us sending unwanted NOTIFY messages to devices) ........ ................ r7516 | kpfleming | 2005-12-17 04:59:27 +0100 (Sa, 17 Dez 2005) | 10 lines Merged revisions 7515 via svnmerge from https://origsvn.digium.com/svn/asterisk/branches/1.2 ........ r7515 | kpfleming | 2005-12-16 21:59:05 -0600 (Fri, 16 Dec 2005) | 2 lines Max-Forwards headers must only be present on requests, not responses ........ ................ r7518 | tilghman | 2005-12-17 18:22:24 +0100 (Sa, 17 Dez 2005) | 3 lines Merged revisions 7517 via svnmerge from /branches/1.2 ................ r7520 | tilghman | 2005-12-17 19:58:57 +0100 (Sa, 17 Dez 2005) | 3 lines Merged revisions 7519 via svnmerge from /branches/1.2 ................ r7522 | tilghman | 2005-12-19 06:42:55 +0100 (Mo, 19 Dez 2005) | 3 lines Merged revisions 7521 via svnmerge from /branches/1.2 ................ r7524 | tilghman | 2005-12-19 20:08:42 +0100 (Mo, 19 Dez 2005) | 3 lines Merged revisions 7523 via svnmerge from /branches/1.2 ................ r7528 | russell | 2005-12-20 00:41:53 +0100 (Di, 20 Dez 2005) | 4 lines - add note on required values of sip_methods struct - remove duplicate function prototype - remove duplicate ast_mutex_lock (issue ASTERISK-5866) ................ r7538 | russell | 2005-12-20 08:45:05 +0100 (Di, 20 Dez 2005) | 3 lines allow forcing the build to exclude PRI support using WITHOUT_PRI, similar to how we already have WITHOUT_ZAPTEL (issue ASTERISK-5829) ................ r7539 | russell | 2005-12-20 09:16:53 +0100 (Di, 20 Dez 2005) | 2 lines use the system libgsm if available (issue ASTERISK-5287, modified to still work with builtin libgsm) ................ r7540 | markster | 2005-12-20 10:39:31 +0100 (Di, 20 Dez 2005) | 2 lines Fix reload of peer contexts (bug ASTERISK-5848) ................ r7541 | markster | 2005-12-20 10:56:55 +0100 (Di, 20 Dez 2005) | 2 lines Fix segfault on directed pickup when no CDR is available (bug ASTERISK-5839) ................ r7542 | markster | 2005-12-20 11:26:53 +0100 (Di, 20 Dez 2005) | 2 lines Fix choppy audio with > 20ms audio frames (bug ASTERISK-5548) ................ r7547 | markster | 2005-12-20 14:07:02 +0100 (Di, 20 Dez 2005) | 3 lines Major peformance improvements to meetme ................ r7548 | markster | 2005-12-20 15:14:01 +0100 (Di, 20 Dez 2005) | 2 lines Don't bother decode on muted participants ................ r7551 | markster | 2005-12-20 18:52:31 +0100 (Di, 20 Dez 2005) | 3 lines Major RTP fixes for using inbound SDP on outbound connection, get rid of old local rtp stuff... ................ r7554 | russell | 2005-12-20 20:56:52 +0100 (Di, 20 Dez 2005) | 2 lines add AGENT function, similar to SIPPEER or IAXPEER (issue ASTERISK-5383) ................ r7556 | russell | 2005-12-20 21:20:04 +0100 (Di, 20 Dez 2005) | 7 lines - move the string join() function to utils.c since it is used in both cli.c and res_agi.c - reimplement ast_join to be of linear effieciency instead of quadratic - remove some useless checks for "if (e)" - reorder checks for strings starting with '_' to avoid a useless call to ast_join() - check array bounds when parsing arguments to AGI (issue ASTERISK-5715) ................ ------------------------------------------------------------------------ http://svn.digium.com/view/asterisk?view=rev&revision=7564 By: Digium Subversion (svnbot) 2008-01-15 16:13:30.000-0600 Repository: asterisk Revision: 7934 _U team/oej/metermaids/ U team/oej/metermaids/apps/app_voicemail.c U team/oej/metermaids/channels/Makefile U team/oej/metermaids/channels/chan_sip.c A team/oej/metermaids/channels/misdn/mISDN.patch A team/oej/metermaids/channels/misdn/mISDNuser.patch U team/oej/metermaids/doc/README.variables U team/oej/metermaids/pbx/pbx_spool.c U team/oej/metermaids/res/res_agi.c ------------------------------------------------------------------------ r7934 | oej | 2008-01-15 16:13:29 -0600 (Tue, 15 Jan 2008) | 145 lines Merged revisions 7490,7517,7529,7546,7550,7552,7557,7580,7586,7595,7605,7641,7663,7706,7738,7771,7792,7812,7870-7871,7898-7900,7904,7908 via svnmerge from https://origsvn.digium.com/svn/asterisk/branches/1.2 ........ r7490 | crichter | 2005-12-15 11:52:30 +0100 (Thu, 15 Dec 2005) | 9 lines * Added mISDN/mISDNuser Echo cancel Patch * Fixed Makefiles so that chan_misdn can be compiled again * added some hints, that mISDN cannot be compiled against gcc-4, SMP, Spinlock Debug * fixed some Minor issues in chan_misdn, regarding Type Of Number and Presentation ........ r7517 | tilghman | 2005-12-17 18:19:32 +0100 (Sat, 17 Dec 2005) | 2 lines Bug 6009 - off by one error ........ r7529 | russell | 2005-12-20 00:47:23 +0100 (Tue, 20 Dec 2005) | 5 lines I messed up and accidently committed this to the trunk first ... - add note on required values of sip_methods struct - remove duplicate function prototype - remove duplicate ast_mutex_lock (issue ASTERISK-5866) ........ r7546 | kpfleming | 2005-12-20 13:58:37 +0100 (Tue, 20 Dec 2005) | 2 lines backport fix for larger-than-20ms-frames from trunk (bug ASTERISK-5548) ........ r7550 | russell | 2005-12-20 18:34:00 +0100 (Tue, 20 Dec 2005) | 2 lines backport fix for segfault on directed pickup when no CDR is available (issue ASTERISK-5839) ........ r7552 | russell | 2005-12-20 19:05:45 +0100 (Tue, 20 Dec 2005) | 2 lines backport fix for reloading peer context (issue ASTERISK-5848) ........ r7557 | russell | 2005-12-20 21:21:26 +0100 (Tue, 20 Dec 2005) | 2 lines check array bounds when parsing arguments to AGI (issue ASTERISK-5715) ........ r7580 | bweschke | 2005-12-21 20:53:49 +0100 (Wed, 21 Dec 2005) | 3 lines Bug ASTERISK-5882 - Documentation correction ........ r7586 | twisted | 2005-12-21 23:23:39 +0100 (Wed, 21 Dec 2005) | 3 lines Actually put in the per-peer settings for sip video, as they didn't make it in at astricon somehow, and I've been too busy up until now to redo it. ........ r7595 | russell | 2005-12-22 17:17:43 +0100 (Thu, 22 Dec 2005) | 2 lines remove stray unlock (issue ASTERISK-5800) ........ r7605 | bweschke | 2005-12-23 01:00:11 +0100 (Fri, 23 Dec 2005) | 3 lines Another app documentation tweak. ........ r7641 | kpfleming | 2005-12-27 01:07:45 +0100 (Tue, 27 Dec 2005) | 2 lines backport fix to ensure that DSP is never enabled on pseudo channels ........ r7663 | russell | 2005-12-27 22:07:08 +0100 (Tue, 27 Dec 2005) | 2 lines backport fix for permissions of created recordings (issue ASTERISK-5909) ........ r7706 | bweschke | 2006-01-02 03:04:14 +0100 (Mon, 02 Jan 2006) | 3 lines Fix compiler warnings. ........ r7738 | kpfleming | 2006-01-03 18:00:01 +0100 (Tue, 03 Jan 2006) | 2 lines backport rport scanning fix from trunk (bug ASTERISK-5913) ........ r7771 | bweschke | 2006-01-04 06:27:38 +0100 (Wed, 04 Jan 2006) | 3 lines Fix the 'if' clause to be true under the right conditions. Bug ASTERISK-5969 ........ r7792 | oej | 2006-01-04 22:43:14 +0100 (Wed, 04 Jan 2006) | 2 lines Fixing typo in XML for video updates. ........ r7812 | oej | 2006-01-05 10:13:21 +0100 (Thu, 05 Jan 2006) | 2 lines Fix copyright of changed file ........ r7870 | russell | 2006-01-09 05:52:16 +0100 (Mon, 09 Jan 2006) | 2 lines backport fix for unnecessary unlock (issue ASTERISK-6014) ........ r7871 | russell | 2006-01-09 06:11:44 +0100 (Mon, 09 Jan 2006) | 2 lines fix seg fault when using greek syntax in VoicemMailMain (issue ASTERISK-5878) ........ r7898 | kpfleming | 2006-01-09 19:08:07 +0100 (Mon, 09 Jan 2006) | 2 lines fix breakage introduced in revision 7871 ........ r7899 | kpfleming | 2006-01-09 19:09:53 +0100 (Mon, 09 Jan 2006) | 2 lines backport fix from revision 7856 of trunk ........ r7900 | kpfleming | 2006-01-09 19:11:23 +0100 (Mon, 09 Jan 2006) | 2 lines commit user/group-related changes from trunk ........ r7904 | tilghman | 2006-01-09 19:37:50 +0100 (Mon, 09 Jan 2006) | 2 lines Update variable documentation to match the code ........ r7908 | tilghman | 2006-01-09 21:08:24 +0100 (Mon, 09 Jan 2006) | 2 lines Bug 6157 - Memory leak ........ ------------------------------------------------------------------------ http://svn.digium.com/view/asterisk?view=rev&revision=7934 By: Digium Subversion (svnbot) 2008-01-15 16:13:35.000-0600 Repository: asterisk Revision: 7938 _U team/oej/managerstuff/ U team/oej/managerstuff/apps/app_voicemail.c U team/oej/managerstuff/channels/chan_iax2.c U team/oej/managerstuff/channels/chan_sip.c A team/oej/managerstuff/channels/misdn/mISDN.patch A team/oej/managerstuff/channels/misdn/mISDNuser.patch U team/oej/managerstuff/doc/README.variables U team/oej/managerstuff/file.c U team/oej/managerstuff/pbx/pbx_spool.c ------------------------------------------------------------------------ r7938 | oej | 2008-01-15 16:13:35 -0600 (Tue, 15 Jan 2008) | 155 lines Merged revisions 7490,7517,7529,7546,7550,7552,7557,7580,7586,7595,7605,7641,7663,7706,7738,7771,7792,7812,7870-7871,7898-7900,7904,7908,7915,7917 via svnmerge from https://origsvn.digium.com/svn/asterisk/branches/1.2 ........ r7490 | crichter | 2005-12-15 11:52:30 +0100 (Thu, 15 Dec 2005) | 9 lines * Added mISDN/mISDNuser Echo cancel Patch * Fixed Makefiles so that chan_misdn can be compiled again * added some hints, that mISDN cannot be compiled against gcc-4, SMP, Spinlock Debug * fixed some Minor issues in chan_misdn, regarding Type Of Number and Presentation ........ r7517 | tilghman | 2005-12-17 18:19:32 +0100 (Sat, 17 Dec 2005) | 2 lines Bug 6009 - off by one error ........ r7529 | russell | 2005-12-20 00:47:23 +0100 (Tue, 20 Dec 2005) | 5 lines I messed up and accidently committed this to the trunk first ... - add note on required values of sip_methods struct - remove duplicate function prototype - remove duplicate ast_mutex_lock (issue ASTERISK-5866) ........ r7546 | kpfleming | 2005-12-20 13:58:37 +0100 (Tue, 20 Dec 2005) | 2 lines backport fix for larger-than-20ms-frames from trunk (bug ASTERISK-5548) ........ r7550 | russell | 2005-12-20 18:34:00 +0100 (Tue, 20 Dec 2005) | 2 lines backport fix for segfault on directed pickup when no CDR is available (issue ASTERISK-5839) ........ r7552 | russell | 2005-12-20 19:05:45 +0100 (Tue, 20 Dec 2005) | 2 lines backport fix for reloading peer context (issue ASTERISK-5848) ........ r7557 | russell | 2005-12-20 21:21:26 +0100 (Tue, 20 Dec 2005) | 2 lines check array bounds when parsing arguments to AGI (issue ASTERISK-5715) ........ r7580 | bweschke | 2005-12-21 20:53:49 +0100 (Wed, 21 Dec 2005) | 3 lines Bug ASTERISK-5882 - Documentation correction ........ r7586 | twisted | 2005-12-21 23:23:39 +0100 (Wed, 21 Dec 2005) | 3 lines Actually put in the per-peer settings for sip video, as they didn't make it in at astricon somehow, and I've been too busy up until now to redo it. ........ r7595 | russell | 2005-12-22 17:17:43 +0100 (Thu, 22 Dec 2005) | 2 lines remove stray unlock (issue ASTERISK-5800) ........ r7605 | bweschke | 2005-12-23 01:00:11 +0100 (Fri, 23 Dec 2005) | 3 lines Another app documentation tweak. ........ r7641 | kpfleming | 2005-12-27 01:07:45 +0100 (Tue, 27 Dec 2005) | 2 lines backport fix to ensure that DSP is never enabled on pseudo channels ........ r7663 | russell | 2005-12-27 22:07:08 +0100 (Tue, 27 Dec 2005) | 2 lines backport fix for permissions of created recordings (issue ASTERISK-5909) ........ r7706 | bweschke | 2006-01-02 03:04:14 +0100 (Mon, 02 Jan 2006) | 3 lines Fix compiler warnings. ........ r7738 | kpfleming | 2006-01-03 18:00:01 +0100 (Tue, 03 Jan 2006) | 2 lines backport rport scanning fix from trunk (bug ASTERISK-5913) ........ r7771 | bweschke | 2006-01-04 06:27:38 +0100 (Wed, 04 Jan 2006) | 3 lines Fix the 'if' clause to be true under the right conditions. Bug ASTERISK-5969 ........ r7792 | oej | 2006-01-04 22:43:14 +0100 (Wed, 04 Jan 2006) | 2 lines Fixing typo in XML for video updates. ........ r7812 | oej | 2006-01-05 10:13:21 +0100 (Thu, 05 Jan 2006) | 2 lines Fix copyright of changed file ........ r7870 | russell | 2006-01-09 05:52:16 +0100 (Mon, 09 Jan 2006) | 2 lines backport fix for unnecessary unlock (issue ASTERISK-6014) ........ r7871 | russell | 2006-01-09 06:11:44 +0100 (Mon, 09 Jan 2006) | 2 lines fix seg fault when using greek syntax in VoicemMailMain (issue ASTERISK-5878) ........ r7898 | kpfleming | 2006-01-09 19:08:07 +0100 (Mon, 09 Jan 2006) | 2 lines fix breakage introduced in revision 7871 ........ r7899 | kpfleming | 2006-01-09 19:09:53 +0100 (Mon, 09 Jan 2006) | 2 lines backport fix from revision 7856 of trunk ........ r7900 | kpfleming | 2006-01-09 19:11:23 +0100 (Mon, 09 Jan 2006) | 2 lines commit user/group-related changes from trunk ........ r7904 | tilghman | 2006-01-09 19:37:50 +0100 (Mon, 09 Jan 2006) | 2 lines Update variable documentation to match the code ........ r7908 | tilghman | 2006-01-09 21:08:24 +0100 (Mon, 09 Jan 2006) | 2 lines Bug 6157 - Memory leak ........ r7915 | russell | 2006-01-09 23:07:26 +0100 (Mon, 09 Jan 2006) | 2 lines add missing unlock (issue ASTERISK-5954) ........ r7917 | kpfleming | 2006-01-09 23:48:48 +0100 (Mon, 09 Jan 2006) | 2 lines re-initialize _all_ sequence numbers when transfer completes ........ ------------------------------------------------------------------------ http://svn.digium.com/view/asterisk?view=rev&revision=7938 By: Digium Subversion (svnbot) 2008-01-15 16:15:22.000-0600 Repository: asterisk Revision: 8054 _U team/crichter/0.3.0/ ------------------------------------------------------------------------ r8054 | crichter | 2008-01-15 16:15:22 -0600 (Tue, 15 Jan 2008) | 110 lines Merged revisions 7490,7517,7529,7546,7550,7552,7557,7580,7586,7595,7605,7641,7663,7706,7738,7771,7792,7812 via svnmerge from https://origsvn.digium.com/svn/asterisk/branches/1.2 ........ r7490 | crichter | 2005-12-15 11:52:30 +0100 (Do, 15 Dez 2005) | 9 lines * Added mISDN/mISDNuser Echo cancel Patch * Fixed Makefiles so that chan_misdn can be compiled again * added some hints, that mISDN cannot be compiled against gcc-4, SMP, Spinlock Debug * fixed some Minor issues in chan_misdn, regarding Type Of Number and Presentation ........ r7517 | tilghman | 2005-12-17 18:19:32 +0100 (Sa, 17 Dez 2005) | 2 lines Bug 6009 - off by one error ........ r7529 | russell | 2005-12-20 00:47:23 +0100 (Di, 20 Dez 2005) | 5 lines I messed up and accidently committed this to the trunk first ... - add note on required values of sip_methods struct - remove duplicate function prototype - remove duplicate ast_mutex_lock (issue ASTERISK-5866) ........ r7546 | kpfleming | 2005-12-20 13:58:37 +0100 (Di, 20 Dez 2005) | 2 lines backport fix for larger-than-20ms-frames from trunk (bug ASTERISK-5548) ........ r7550 | russell | 2005-12-20 18:34:00 +0100 (Di, 20 Dez 2005) | 2 lines backport fix for segfault on directed pickup when no CDR is available (issue ASTERISK-5839) ........ r7552 | russell | 2005-12-20 19:05:45 +0100 (Di, 20 Dez 2005) | 2 lines backport fix for reloading peer context (issue ASTERISK-5848) ........ r7557 | russell | 2005-12-20 21:21:26 +0100 (Di, 20 Dez 2005) | 2 lines check array bounds when parsing arguments to AGI (issue ASTERISK-5715) ........ r7580 | bweschke | 2005-12-21 20:53:49 +0100 (Mi, 21 Dez 2005) | 3 lines Bug ASTERISK-5882 - Documentation correction ........ r7586 | twisted | 2005-12-21 23:23:39 +0100 (Mi, 21 Dez 2005) | 3 lines Actually put in the per-peer settings for sip video, as they didn't make it in at astricon somehow, and I've been too busy up until now to redo it. ........ r7595 | russell | 2005-12-22 17:17:43 +0100 (Do, 22 Dez 2005) | 2 lines remove stray unlock (issue ASTERISK-5800) ........ r7605 | bweschke | 2005-12-23 01:00:11 +0100 (Fr, 23 Dez 2005) | 3 lines Another app documentation tweak. ........ r7641 | kpfleming | 2005-12-27 01:07:45 +0100 (Di, 27 Dez 2005) | 2 lines backport fix to ensure that DSP is never enabled on pseudo channels ........ r7663 | russell | 2005-12-27 22:07:08 +0100 (Di, 27 Dez 2005) | 2 lines backport fix for permissions of created recordings (issue ASTERISK-5909) ........ r7706 | bweschke | 2006-01-02 03:04:14 +0100 (Mo, 02 Jan 2006) | 3 lines Fix compiler warnings. ........ r7738 | kpfleming | 2006-01-03 18:00:01 +0100 (Di, 03 Jan 2006) | 2 lines backport rport scanning fix from trunk (bug ASTERISK-5913) ........ r7771 | bweschke | 2006-01-04 06:27:38 +0100 (Mi, 04 Jan 2006) | 3 lines Fix the 'if' clause to be true under the right conditions. Bug ASTERISK-5969 ........ r7792 | oej | 2006-01-04 22:43:14 +0100 (Mi, 04 Jan 2006) | 2 lines Fixing typo in XML for video updates. ........ r7812 | oej | 2006-01-05 10:13:21 +0100 (Do, 05 Jan 2006) | 2 lines Fix copyright of changed file ........ ------------------------------------------------------------------------ http://svn.digium.com/view/asterisk?view=rev&revision=8054 By: Digium Subversion (svnbot) 2008-01-15 16:15:24.000-0600 Repository: asterisk Revision: 8055 _U team/crichter/frame-data/ U team/crichter/frame-data/apps/app_voicemail.c U team/crichter/frame-data/channel.c U team/crichter/frame-data/channels/chan_agent.c U team/crichter/frame-data/configs/voicemail.conf.sample ------------------------------------------------------------------------ r8055 | crichter | 2008-01-15 16:15:24 -0600 (Tue, 15 Jan 2008) | 246 lines Merged revisions 8053-8054 via svnmerge from https://origsvn.digium.com/svn/asterisk/team/crichter/0.3.0 ................ r8053 | crichter | 2006-01-13 12:03:35 +0100 (Fr, 13 Jan 2006) | 125 lines Merged revisions 7870-7871,7898-7900,7963,7973-7975,7977-8051 via svnmerge from https://origsvn.digium.com/svn/asterisk/branches/1.2 ........ r7870 | russell | 2006-01-09 05:52:16 +0100 (Mo, 09 Jan 2006) | 2 lines backport fix for unnecessary unlock (issue ASTERISK-6014) ........ r7871 | russell | 2006-01-09 06:11:44 +0100 (Mo, 09 Jan 2006) | 2 lines fix seg fault when using greek syntax in VoicemMailMain (issue ASTERISK-5878) ........ r7898 | kpfleming | 2006-01-09 19:08:07 +0100 (Mo, 09 Jan 2006) | 2 lines fix breakage introduced in revision 7871 ........ r7899 | kpfleming | 2006-01-09 19:09:53 +0100 (Mo, 09 Jan 2006) | 2 lines backport fix from revision 7856 of trunk ........ r7900 | kpfleming | 2006-01-09 19:11:23 +0100 (Mo, 09 Jan 2006) | 2 lines commit user/group-related changes from trunk ........ r7963 | mogorman | 2006-01-11 05:38:07 +0100 (Mi, 11 Jan 2006) | 2 lines Minor typo refrenced in 6191 ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ r7986 | russell | 2006-01-11 20:08:53 +0100 (Mi, 11 Jan 2006) | 2 lines move variable to correct scope (issue ASTERISK-6040) ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ r7999 | tilghman | 2006-01-12 07:14:22 +0100 (Do, 12 Jan 2006) | 2 lines Bug 6211 - Add option deletevoicemail as equivalent to option delete for Realtime ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ ........ r8047 | russell | 2006-01-13 07:07:39 +0100 (Fr, 13 Jan 2006) | 2 lines fix spelling errors (issue ASTERISK-6069) ........ ........ ........ ........ ........ ................ r8054 | crichter | 2006-01-13 14:00:07 +0100 (Fr, 13 Jan 2006) | 110 lines Merged revisions 7490,7517,7529,7546,7550,7552,7557,7580,7586,7595,7605,7641,7663,7706,7738,7771,7792,7812 via svnmerge from https://origsvn.digium.com/svn/asterisk/branches/1.2 ........ r7490 | crichter | 2005-12-15 11:52:30 +0100 (Do, 15 Dez 2005) | 9 lines * Added mISDN/mISDNuser Echo cancel Patch * Fixed Makefiles so that chan_misdn can be compiled again * added some hints, that mISDN cannot be compiled against gcc-4, SMP, Spinlock Debug * fixed some Minor issues in chan_misdn, regarding Type Of Number and Presentation ........ r7517 | tilghman | 2005-12-17 18:19:32 +0100 (Sa, 17 Dez 2005) | 2 lines Bug 6009 - off by one error ........ r7529 | russell | 2005-12-20 00:47:23 +0100 (Di, 20 Dez 2005) | 5 lines I messed up and accidently committed this to the trunk first ... - add note on required values of sip_methods struct - remove duplicate function prototype - remove duplicate ast_mutex_lock (issue ASTERISK-5866) ........ r7546 | kpfleming | 2005-12-20 13:58:37 +0100 (Di, 20 Dez 2005) | 2 lines backport fix for larger-than-20ms-frames from trunk (bug ASTERISK-5548) ........ r7550 | russell | 2005-12-20 18:34:00 +0100 (Di, 20 Dez 2005) | 2 lines backport fix for segfault on directed pickup when no CDR is available (issue ASTERISK-5839) ........ r7552 | russell | 2005-12-20 19:05:45 +0100 (Di, 20 Dez 2005) | 2 lines backport fix for reloading peer context (issue ASTERISK-5848) ........ r7557 | russell | 2005-12-20 21:21:26 +0100 (Di, 20 Dez 2005) | 2 lines check array bounds when parsing arguments to AGI (issue ASTERISK-5715) ........ r7580 | bweschke | 2005-12-21 20:53:49 +0100 (Mi, 21 Dez 2005) | 3 lines Bug ASTERISK-5882 - Documentation correction ........ r7586 | twisted | 2005-12-21 23:23:39 +0100 (Mi, 21 Dez 2005) | 3 lines Actually put in the per-peer settings for sip video, as they didn't make it in at astricon somehow, and I've been too busy up until now to redo it. ........ r7595 | russell | 2005-12-22 17:17:43 +0100 (Do, 22 Dez 2005) | 2 lines remove stray unlock (issue ASTERISK-5800) ........ r7605 | bweschke | 2005-12-23 01:00:11 +0100 (Fr, 23 Dez 2005) | 3 lines Another app documentation tweak. ........ r7641 | kpfleming | 2005-12-27 01:07:45 +0100 (Di, 27 Dez 2005) | 2 lines backport fix to ensure that DSP is never enabled on pseudo channels ........ r7663 | russell | 2005-12-27 22:07:08 +0100 (Di, 27 Dez 2005) | 2 lines backport fix for permissions of created recordings (issue ASTERISK-5909) ........ r7706 | bweschke | 2006-01-02 03:04:14 +0100 (Mo, 02 Jan 2006) | 3 lines Fix compiler warnings. ........ r7738 | kpfleming | 2006-01-03 18:00:01 +0100 (Di, 03 Jan 2006) | 2 lines backport rport scanning fix from trunk (bug ASTERISK-5913) ........ r7771 | bweschke | 2006-01-04 06:27:38 +0100 (Mi, 04 Jan 2006) | 3 lines Fix the 'if' clause to be true under the right conditions. Bug ASTERISK-5969 ........ r7792 | oej | 2006-01-04 22:43:14 +0100 (Mi, 04 Jan 2006) | 2 lines Fixing typo in XML for video updates. ........ r7812 | oej | 2006-01-05 10:13:21 +0100 (Do, 05 Jan 2006) | 2 lines Fix copyright of changed file ........ ................ ------------------------------------------------------------------------ http://svn.digium.com/view/asterisk?view=rev&revision=8055 | ||