set of ebuilds for checking out and installing latest zaptel-cvs, zapata-cvs, libpri-cvs and asterisk-cvs; these components make up the asterisk software pbx. Reproducible: Always Steps to Reproduce: these ebuilds use the new ECVS_ format, but are based partially on existing asterisk and zaptel builds: http://bugs.gentoo.org/show_bug.cgi?id=30873 http://bugs.gentoo.org/show_bug.cgi?id=30531 (and the included patches to zaptel-cvs are taken directly from Russ' ;-) example install (unmerge _pre asterisk and zaptel builds if already installed!) cd /usr/local/portage bzcat /path/to/asterisk_cvs.tar.bz2 | tar xf - ACCEPT_KEYWORDS="~x86" emerge -pv zaptel-cvs zapata-cvs libpri-cvs asterisk-cvs comments, improvements, and insults are welcomed. NB!!! as a result of the INSTALL_PREFIX trickery needed to install inside the sandbox, three symlinks are incorrect after the asterisk-cvs build completes: >>> /var/lib/asterisk/sounds/vm -> /var/tmp/portage/asterisk-cvs-0.1/image//var/spool/asterisk/vm (should point to /var/spool/asterisk/vm) >>> /var/lib/asterisk/sounds/voicemail -> /var/tmp/portage/asterisk-cvs-0.1/image//var/spool/asterisk/voicemail (should point to /var/spool/asterisk/voicemail) >>> /var/spool/asterisk/vm -> /var/tmp/portage/asterisk-cvs-0.1/image//var/spool/asterisk/voicemail/default (should point to /var/spool/asterisk/voicemail/default)
Created attachment 20660 [details] tarball of the four cvs ebuilds asterisk_cvs.tar.bz2 contains ebuilds for: zaptel-cvs zapata-cvs libpri-cvs asterisk-cvs
Created attachment 20699 [details] asterisk_cvs.tar.bz2 obsoletes initial tarball: now in addition to builds for zaptel-cvs, zapata-cvs, libpri-cvs and asterisk-cvs, now is added ebuild for gastman-cvs.
Created attachment 20710 [details] asterisk.tar.bz2 contains asterisk-cvs-0.2 ebuild ONLY. according to http://bugs.digium.com/bug_view_page.php?bug_id=0000200 asterisk Makefile supports DESTDIR correctly. this means that the abovementioned three install-time symlinking issues are resolved, so the ebuild has been accordingly updated (and tested successfully, yes) Makefile for each of zaptel, zapata, libpri, gastman does not yet support DESTDIR; these ebuilds remain the same.
Created attachment 20718 [details] zaptel-cvs-0.2.tar.bz2 INSTALL_PREFIX typo in zaptel Makefile has been fixed in cvs: http://bugs.digium.com/bug_view_page.php?bug_id=0000530 so one of the ebuild patches is no longer necessary. zaptel-cvs-0.2 ebuild is attached.
Created attachment 23655 [details] zaptel-cvs-0.4.tbz2 asterisk codebase appears to have been cleaned up quite a bit in preparation for the release of version 0.7 tomorrow. one new feature is that zaptel finally compiles (and seems to run well enough, so far) under linux 2.6. the only difference between this ebuild and the previous zaptel-cvs ebuild is that it does make linux26 || die instead of make || die i.e. don't use this build if you are on 2.4, continue using zaptel-cvs-0.2.ebuild. probably the least smooth build yet, this is only for people who are dying to move to 2.6 with zaptel... 1. where ever your source tree is will have to be linked to /usr/src/linux-2.6, e.g. # cd /usr/src # ln -s linux linux-2.6 2. at build time, the makefile evaluates 'arch' as x86, so a small link is in order: # cd /usr/src/linux/arch/ # ln -s i386 x86 3. some building inside the kernel source tree happens, so I was lazy and disabled sandbox to get it working: # cd /usr/local/portage/net-misc/zaptel-cvs # FEATURES="-sandbox" emerge -pv zaptel-cvs-0.4.ebuild the sandbox thing could probably be avoided, if your src/linux tree were to be owned by, say, portage instead of root....
Got the following when trying the builds today... * Applying zaptel-cvs-modulesd.patch... * Failed Patch: zaptel-cvs-modulesd.patch! * * Include in your bugreport the contents of: * * /var/tmp/portage/zaptel-cvs-0.4/temp/zaptel-cvs-modulesd.patch-11698.out !!! ERROR: net-misc/zaptel-cvs-0.4 failed. !!! Function epatch, Line 360, Exitcode 0 !!! Failed Patch: zaptel-cvs-modulesd.patch! ***** zaptel-cvs-modulesd.patch ***** ===================================== PATCH COMMAND: patch -p0 < /usr/local/portage/net-misc/zaptel-cvs/files/zaptel-cvs-modulesd.patch ===================================== can't find file to patch at input line 3 Perhaps you used the wrong -p or --strip option? The text leading up to this was: -------------------------- |--- zaptel/Makefile.orig 2003-10-10 13:30:43.000000000 -0400 |+++ zaptel/Makefile 2003-10-10 13:30:09.000000000 -0400 -------------------------- No file to patch. Skipping patch. 3 out of 3 hunks ignored ===================================== PATCH COMMAND: patch -p1 < /usr/local/portage/net-misc/zaptel-cvs/files/zaptel-cvs-modulesd.patch ===================================== patching file Makefile Hunk #1 FAILED at 80. Hunk #2 FAILED at 265. Hunk #3 FAILED at 299. 3 out of 3 hunks FAILED -- saving rejects to file Makefile.rej ===================================== PATCH COMMAND: patch -p2 < /usr/local/portage/net-misc/zaptel-cvs/files/zaptel-cvs-modulesd.patch ===================================== missing header for unified diff at line 3 of patch can't find file to patch at input line 3 Perhaps you used the wrong -p or --strip option? The text leading up to this was: -------------------------- |--- zaptel/Makefile.orig 2003-10-10 13:30:43.000000000 -0400 |+++ zaptel/Makefile 2003-10-10 13:30:09.000000000 -0400 -------------------------- No file to patch. Skipping patch. 3 out of 3 hunks ignored ===================================== PATCH COMMAND: patch -p3 < /usr/local/portage/net-misc/zaptel-cvs/files/zaptel-cvs-modulesd.patch ===================================== missing header for unified diff at line 3 of patch can't find file to patch at input line 3 Perhaps you used the wrong -p or --strip option? The text leading up to this was: -------------------------- |--- zaptel/Makefile.orig 2003-10-10 13:30:43.000000000 -0400 |+++ zaptel/Makefile 2003-10-10 13:30:09.000000000 -0400 -------------------------- No file to patch. Skipping patch. 3 out of 3 hunks ignored ===================================== PATCH COMMAND: patch -p4 < /usr/local/portage/net-misc/zaptel-cvs/files/zaptel-cvs-modulesd.patch ===================================== missing header for unified diff at line 3 of patch can't find file to patch at input line 3 Perhaps you used the wrong -p or --strip option? The text leading up to this was: -------------------------- |--- zaptel/Makefile.orig 2003-10-10 13:30:43.000000000 -0400 |+++ zaptel/Makefile 2003-10-10 13:30:09.000000000 -0400 -------------------------- No file to patch. Skipping patch. 3 out of 3 hunks ignored
Created attachment 26131 [details, diff] zaptel-cvs-modulesd.patch looks like the makefile changed a little ... not drastically, just enough to upset this patch. here's an updated version which patched ok when I tested it.
Created attachment 28486 [details, diff] zaptel-cvs-modulesd.patch same patch failed against latest cvs, this version works though.
Created attachment 28488 [details] asterisk-cvs-1.0.ebuild this uses the 1.0 stable cvs branch, that's the only difference between it and the existing cvs ebuild. it's as simple as: ECVS_BRANCH="v1-0_stable" the 1.0 branch is getting bug fixes only, while the regular cvs still gets new features as well (e.g. iaxy firmware updater) -- personally I don't use 1.0 branch but some people might want it.
As of April 17 the zaptel-cvs-modulesd.patch fails with the next message ***** zaptel-cvs-modulesd.patch ***** ===================================== PATCH COMMAND: patch -p0 < /usr/portage/net-misc/zaptel-cvs/files/zaptel-cvs-modulesd.patch ===================================== can't find file to patch at input line 3 Perhaps you used the wrong -p or --strip option? The text leading up to this was: -------------------------- |--- zaptel/Makefile.orig 2004-04-01 00:04:53.685599734 -0500 |+++ zaptel/Makefile 2004-04-01 00:06:25.890868491 -0500 -------------------------- No file to patch. Skipping patch. 3 out of 3 hunks ignored ===================================== PATCH COMMAND: patch -p1 < /usr/portage/net-misc/zaptel-cvs/files/zaptel-cvs-modulesd.patch ===================================== patching file Makefile Hunk #1 FAILED at 33. Hunk #2 succeeded at 273 (offset 8 lines). Hunk #3 succeeded at 298 (offset 8 lines). 1 out of 3 hunks FAILED -- saving rejects to file Makefile.rej ===================================== PATCH COMMAND: patch -p2 < /usr/portage/net-misc/zaptel-cvs/files/zaptel-cvs-modulesd.patch ===================================== missing header for unified diff at line 3 of patch can't find file to patch at input line 3 Perhaps you used the wrong -p or --strip option? The text leading up to this was: -------------------------- |--- zaptel/Makefile.orig 2004-04-01 00:04:53.685599734 -0500 |+++ zaptel/Makefile 2004-04-01 00:06:25.890868491 -0500 -------------------------- No file to patch. Skipping patch. 3 out of 3 hunks ignored ===================================== PATCH COMMAND: patch -p3 < /usr/portage/net-misc/zaptel-cvs/files/zaptel-cvs-modulesd.patch ===================================== missing header for unified diff at line 3 of patch can't find file to patch at input line 3 Perhaps you used the wrong -p or --strip option? The text leading up to this was: -------------------------- |--- zaptel/Makefile.orig 2004-04-01 00:04:53.685599734 -0500 |+++ zaptel/Makefile 2004-04-01 00:06:25.890868491 -0500 -------------------------- No file to patch. Skipping patch. 3 out of 3 hunks ignored ===================================== PATCH COMMAND: patch -p4 < /usr/portage/net-misc/zaptel-cvs/files/zaptel-cvs-modulesd.patch ===================================== missing header for unified diff at line 3 of patch can't find file to patch at input line 3 Perhaps you used the wrong -p or --strip option? The text leading up to this was: -------------------------- |--- zaptel/Makefile.orig 2004-04-01 00:04:53.685599734 -0500 |+++ zaptel/Makefile 2004-04-01 00:06:25.890868491 -0500 -------------------------- No file to patch. Skipping patch. 3 out of 3 hunks ignored
Created attachment 29530 [details, diff] zaptel-cvs-modulesd.patch noticed that a few days ago, here's yet another version of the modulesd patch...
Created attachment 31721 [details, diff] Fix music on hold volume (for mpg123-0.59s-r3) Asterisk's music on hold does only work with >=media-sound/mpg123-0.59s-r3 or mpg123-0.59r (not in portage). This patch fixes the low volume with 0.59s-r3.
moh worked fine here with 0.59s-r3 but then I always used loud instead of quietmp3 setting...'loud' seems a little quiet recently so I think you may be correct. you should also open a bug and submit your patch at bugs.digium.com ! nb: the relevant 1.0-stable vs. cvs-head code are very different for that file, the above patch will only work for cvs-head. earlier ebuilds of 0.59s actually *did* work -- but the mp3 files in the moh folder had to be in the exact mono 8khz cbr format that the asterisk documentation specifies, otherwise: crash city. with 0.59s-r3 I can once again use e.g. vbr stereo 44.1khz mp3, and asterisk won't crash.
Are there any plans at getting this into Portage? As discussed in the asterisk mailing lists, all the development happens on CVS head and it would be much more convinient to have a -cvs ebuild. Thanks!
Created attachment 33265 [details] zaptel-cvs-0.5.ebuild that's not a bad idea -- for a while I was running 1.0-stable until I realized that cvs head is actually much more stable. so here's a start: new zaptel-cvs build based partly on my previous one and partly on gentoo's decent zaptel-0.9.1 build -- supports both 2.4 and 2.6 (no devfs for 2.6 tho) and I also submitted a (very minor) bug to digium, so that with luck zaptel-cvs-mkdir-usrincludelinux.patch should no longer be necessary sometime in the near future. the modulesd patch will probably have to stick around tho (which is annoying as it obviously needs tweaking from time to time...)
Created attachment 33266 [details] asterisk-cvs-0.2.ebuild this is the ebuild I currently use for asterisk cvs head. no patches needed, nothing special. although in src_compile() I had to chmod some scripts (mkdep, make_build_h, editline/configure) as executable to make it work. (can files in cvs be marked +x?? if so, that's another small tweak to submit to digium...)
Comment on attachment 33265 [details] zaptel-cvs-0.5.ebuild (changed mimetype to text/plain)
btw, grats to the gentoo devs for the addwrite and MY_ARCH trickery: building with sandbox turned off and symlinking no longer necessary with this zaptel-cvs-0.5.ebuild. thanks guys!!
Isn't it possible to set the permissions right with chmod from the ebuild script? Ideally we could ask Digium to correct it but in the meantime...... :)
Created attachment 34817 [details] zaptel-cvs-0.6.ebuild just a minor change: zaptel-cvs-mkdir-usrincludelinux.patch no longer necessary since the problem has been fixed in digium's cvs.
as I mentioned, the 0.2 ebuild does in fact chmod +x the three files that need it. however, it looks as if digium fixed the mode on those files in cvs on their own, so the chmod is redundant now ^_^
Created attachment 35277 [details] asterisk-cvs-0.3.ebuild This removes zaptel as a dependency, zaptel is NOT required to run asterisk. It instead makes zaptel-cvs and libpri-cvs USE flags. This update also adds an einfo with links to documentation.
Created attachment 35279 [details] zaptel-cvs-0.7.ebuild Adds a warning that 2.6.x Zaptel is beta and to report bugs to digium.
Created attachment 35280 [details] asterisk-cvs-0.4.ebuild Fixed typo.
Created attachment 35281 [details] zaptel-cvs-0.8.ebuild Masked...
Created attachment 35282 [details] libpri-cvs-0.1.ebuild First ebuild for libpri, a USE flag for asterisk.
Cuban's CVS ebuilds compile perfectly on mine, however with the obvious practical problems that no conf.d, init.d or /etc/asterisk files are installed. I'm currently using a rather oddball Cyrix C3 Eden (low thermal output processor) system which is somewhat similar to a P2
Yes, I will be updating the asterisk ebuild to create an init script. If you would like documentation use +docs
I ran ufed to add doc to my list of use-flags, unmerged and then re-emerged asterisk-cvs. There was still nothing in the /etc/asterisk folder.
cool stuff, thanks. I would also recommend: (a) wrap the ewarn in zaptel-cvs to only happen if they are building against 2.6, no need to unnecessarily annoy 2.4 users. (2) if you're going to be updating asterisk-cvs soon, please take out the chmod statements in src_compile -- as noted, they are no longer necessary. (vii) to get config files etc., it's not 'make progdocs' but 'make samples' -- however, samples and docs should be two separate things: you probably only want to install the sample config once, same goes for the initscript (yes, there's already one in asterisk/contrib/init.d), so that should be a separate new USE flag; and then the docs obviously go to the doc USE flag. (12) maybe change the static dodocs in zaptel-cvs to run only if the doc USE flag is enabled...? my personal recommendation is still to use inittab to run asterisk, or possibly something like daemontools that can keep an eye on it.
Incidentally, the Asterisk folks have now released release candidate 1 for Asterisk, zaptel, etc, according to the asterisk-users mailing list. This is a new attempt at creating a stable 1.0 release. It might better suit what I percieve as portage policy if we can have 1.0rc1 ebuilds if possible.
Everything happens in CVS, many users will always use HEAD for their servers. I believe we should make a 1.0 ebuild when it comes out but continue to use the cvs ebuild.
since 1.0rc1 is a released tarball rather than a cvs branch like 1.0-stable was, it would seem to make sense that 1.0rc1 should exist as an increment to gentoo's current *official* asterisk ebuilds. also, I agree that (in the past, at least) cvs head is where stuff "happens" -- 1.0-stable was not all that stable, while at the same time there were fixes and stability going into head.
I would like to add one vote for replacing the 0.9 ebuild in portage with the 1.0RC1. If people want to use cvs version ok but as a gentoo n00b I prefer to just use an ebuild from the portage tree instead of something from here. In the end we will end up adding 1.0 to the portage tree anyway so why not start now :)
Then make an ebuild for 1.0. Discussing the 1.0 ebuild is offtopic of this bug. This bug is about the CVS ebuild.
Created attachment 36163 [details] Adds a "samples" USE flag, fixes the "docs" USE flags and adds an init script.
Created attachment 36164 [details] asterisk-cvs-0.6.ebuild Adds a "samples" USE flag, fixes the "docs" USE flags and adds an init script. Fixes a typo from 0.5.
Created attachment 36165 [details] asterisk.rc6 This goes in /usr/local/portage/net-misc/asterisk-cvs/files This is the init script, a modified 0.9.0 init script to follow Digium CVS's new debian init script.
Also, I removed those chmods that weren't needed.
This bug is going no where with portage so I've opened 58439. It's been opened since november and no responses.
er, what I meant was this bug has been opened since november without much response.
There is a typo in asterisk-cvs-0.6.ebuild : the IUSE flag used for the documentation is "doc" as usually seen in ebuilds, but anywhere else in this ebuild it appears as "docs". It is advisable to use "doc" anywhere.
Asterisk 1.0 Final is released :)
Should this bug be closed?
The repository is subversion now. http://www.asterisk.org/asterisk-converts-to-subversion Create a new bug, or change this summary.
this bug is old and asterisk releases often enough by now. Thanks
Asterisk 1.4.0 released! URL: http://www.asterisk.org/node/48266 Please, add ebuild to this version of asterisk ! There is many fixes and new features! :) I can help with testing ...