Since DETECT_DEADLOCKS is now split from DEBUG_THREADS, it must
always be included in buildopts.h instead of only when
ADD_CFLAGS_TO_BUILDOPTS_H is defined. A SEGV will result otherwise.
Resolves: #719
(cherry picked from commit af3dd9e2e3)
Although `make_xml_documentation`'s `print_dependencies` command was
corrected by the previous fix (#461) for #142, the `create_xml` was
not properly handling `LOCAL_MOD_SUBDIRS` XML documentation.
(cherry picked from commit 91c733bc69)
If LOCAL_MOD_SUBDIRS contains absolute paths, do not prefix them with
the path to Asterisk's source tree.
Fixes#142
(cherry picked from commit 2293edffd0)
The current STIR/SHAKEN implementation is not currently usable due
to encryption issues. Rather than trying to futz with OpenSSL and
the the current code, we can take advantage of the existing
capabilities of libjwt but we first need to add it to the
third-party infrastructure already in place for jansson and
pjproject.
A few tweaks were also made to the third-party infrastructure as
a whole. The jansson "dest" install directory was renamed "dist"
to better match convention, and the third-party Makefile was updated
to clean all product directories not just the ones currently in
use.
Resolves: #349
(cherry picked from commit a9d4175e1d)
Previously, DETECT_DEADLOCKS depended on DEBUG_THREADS.
Unfortunately, DEBUG_THREADS adds a lot of lock tracking overhead
to all of the lock lifecycle calls whereas DETECT_DEADLOCKS just
causes the lock calls to loop over trylock in 200us intervals until
the lock is obtained and spits out log messages if it takes more
than 5 seconds. From a code perspective, the only reason they were
tied together was for logging. So... The ifdefs in lock.c were
refactored to allow DETECT_DEADLOCKS to be enabled without
also enabling DEBUG_THREADS.
Resolves: #321
UserNote: You no longer need to select DEBUG_THREADS to use
DETECT_DEADLOCKS. This removes a significant amount of overhead
if you just want to detect possible deadlocks vs needing full
lock tracing.
(cherry picked from commit 7e2243f9e1)
The previous behavior of make_buildopts_h was to not add the
non-ABI-breaking MENUSELECT_CFLAGS like DETECT_DEADLOCKS,
REF_DEBUG, etc. to the buildopts.h file because "it caused
ccache to invalidate files and extended compile times". They're
only defined by passing them on the gcc command line with '-D'
options. In practice, including them in the include file rarely
causes any impact because the only time ccache cares is if you
actually change an option so the hit occurrs only once after
you change it.
OK so why would we want to include them? Many IDEs follow the
include files to resolve defines and if the options aren't in an
include file, it can cause the IDE to mark blocks of "ifdeffed"
code as unused when they're really not.
So...
* Added a new menuselect compile option ADD_CFLAGS_TO_BUILDOPTS_H
which tells make_buildopts_h to include the non-ABI-breaking
flags in buildopts.h as well as the ABI-breaking ones. The default
is disabled to preserve current behavior. As before though,
only the ABI-breaking flags appear in AST_BUILDOPTS and only
those are used to calculate AST_BUILDOPT_SUM.
A new AST_BUILDOPT_ALL define was created to capture all of the
flags.
* make_version_c was streamlined to use buildopts.h and also to
create asterisk_build_opts_all[] and ast_get_build_opts_all(void)
* "core show settings" now shows both AST_BUILDOPTS and
AST_BUILDOPTS_ALL.
UserNote: The "Build Options" entry in the "core show settings"
CLI command has been renamed to "ABI related Build Options" and
a new entry named "All Build Options" has been added that shows
both breaking and non-breaking options.
(cherry picked from commit 06da7b342e)
* Fixed issue with the script not parsing the new tag format for
certified releases. The format changed from certified/18.9-cert5
to certified-18.9-cert5.
* Fixed issue where the asterisk version wasn't being considered
when looking for cached versions.
Resolves: #263
(cherry picked from commit 715594767d)
* Remove .gitreview and switch to pulling the main asterisk branch
version from configure.ac instead.
* Replace references to JIRA with GitHub.
* Other minor cleanup found along the way.
Resolves: #39
(cherry picked from commit 5c6d5ea38f)
* All of the code that used subversion has been removed.
* When Asterisk is checked out from a tag or commit instead
of one of the regular branches, git would emit messages like
"fatal: ref HEAD is not a symbolic ref" which weren't fatal
at all. Those are now suppressed.
Change-Id: I2a11bc9ebbaf6dfa50f53516ede50a6bac65ca3c
(cherry picked from commit bbec5d1a99)
make_version now silently checks if the required git commands will
fail. If they do, then return UNKNOWN__git_check_fail to
distinguish this failure from other UNKNOWN__ version failures
Makefile checks for this value on install and exits out with
instructions
ASTERISK-30029
Change-Id: If8f10cac8f509c08981120f17555762342020221
get_sourceable_makeopts wasn't handling variables with embedded
double quotes in them very well. One example was the DOWNLOAD
variable when curl was being used instead of wget. Rather than
trying to fix get_sourceable_makeopts, it's just been removed.
ASTERISK-29986
Reported by: Stefan Ruijsenaars
Change-Id: Idf2a90902228c2558daa5be7a4f8327556099cd2
Added functions to open, close, and apply XML Stylesheets
to XML documents. Although the presence of libxslt was already
being checked by configure, it was only happening if xmldoc was
enabled. Now it's checked regardless.
Added ability to parse a string consisting of comma separated
name/value pairs into an ast_variable list. The reverse of
ast_variable_list_join().
Change-Id: I1e1d149be22165a1fb8e88e2903a36bba1a6cf2e
Added the missing xml-stylesheet and Xinclude namespace
declarations in pjsip_config.xml and pjsip_manager.xml.
Updated make_xml_documentation to show detailed errors when
xmlstarlet is the validator. It's now run once with the '-q'
option to suppress harmless/expected messages and if it actually
fails, it's run again without '-q' but with '-e' to show
the actual errors.
Change-Id: I4bdc9d2ea6741e8d2e5eb82df60c68ccc59e1f5e
Moved the xmldoc build logic from the top-level Makefile into
its own script "make_xml_documentation" in the build_tools
directory.
Created a new utility script "get_sourceable_makeopts", also in
the build_tools directory, that dumps the top-level "makeopts"
file in a format that can be "sourced" from shell sscripts.
This allows scripts to easily get the values of common make
build variables such as the location of the GREP, SED, AWK, etc.
utilities as well as the AST* and library *_LIB and *_INCLUDE
variables.
Besides moving logic out of the Makefile, some optimizations
were done like removing "third-party" from the list of
subdirectories to be searched for documentation and changing some
assignments from "=" to ":=" so they're only evaluated once.
The speed increase is noticeable.
The makeopts.in file was updated to include the paths to
REALPATH and DIRNAME. The ./conifgure script was setting them
but makeopts.in wasn't including them.
So...
With this change, you can now place documentation in any"c"
source file AND you can now place it in a separate XML file
altogether. The following are examples of valid locations:
res/res_pjsip.c
Using the existing /*** DOCUMENTATION ***/ fragment.
res/res_pjsip/pjsip_configuration.c
Using the existing /*** DOCUMENTATION ***/ fragment.
res/res_pjsip/pjsip_doc.xml
A fully-formed XML file. The "configInfo", "manager",
"managerEvent", etc. elements that would be in the "c"
file DOCUMENTATION fragment should be wrapped in proper
XML. Example for "somemodule.xml":
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE docs SYSTEM "appdocsxml.dtd">
<docs>
<configInfo>
...
</configInfo>
</docs>
It's the "appdocsxml.dtd" that tells make_xml_documentation
that this is a documentation XML file and not some other XML file.
It also allows many XML-capable editors to do formatting and
validation.
Other than the ".xml" suffix, the name of the file is not
significant.
As a start... This change also moves the documentation that was
in res_pjsip.c to 2 new XML files in res/res_pjsip:
pjsip_config.xml and pjsip_manager.xml. This cut the number of
lines in res_pjsip.c in half. :)
Change-Id: I486c16c0b5a44d7a8870008e10c941fb19b71ade
Recap from earlier commit: If you have a development branch for a
major project that will receive gerrit reviews it'll probably be
named something like "development/16/newproject" or a work branch
based on that "development" branch. That will necessitate
setting "defaultbranch=development/16/newproject" in .gitreview.
The make_version script uses that variable to construct the
asterisk version however, which results in versions
like "GIT-development/16/newproject-ee582a8c7b" which is probably
not what you want. It also constructs the URLs for downloading
external modules with that version, which will fail.
Fast-forward:
The earlier attempt at adding a "basebranch" variable to
.gitreview didn't work out too well in practice because changes
were made to .gitreview, which is a checked-in file. So, if
you wanted to rebase your work branch on the base branch, rebase
would attempt to overwrite your .gitreview with the one from
the base branch and complain about a conflict.
This is a slighltly different approach that adds three methods to
determine the mainline branch:
1. --- MAINLINE_BRANCH from the environment
If MAINLINE_BRANCH is already set in the environment, that will
be used. This is primarily for the Jenkins jobs.
2. --- .develvars
Instead of storing the basebranch in .gitreview, it can now be
stored in a non-checked-in ".develvars" file and keyed by the
current branch. So, if you were working on a branch named
"new-feature-work" based on "development/16/new-feature" and wanted
to push to that branch in Gerrit but wanted to pull the external
modules for 16, you'd create the following .develvars file:
[branch "new-feature-work"]
mainline-branch = 16
The .gitreview file would still look like:
[gerrit]
defaultbranch=development/16/new-feature
...which would cause any reviews pushed from "new-feature-work" to
go to the "development/16/new-feature" branch in Gerrit.
The key is that the .develvars file is NEVER checked in (it's been
added to .gitignore).
3. --- Well Known Development Branch
If you're actually working in a branch named like
"development/<mainline_branch>/some-feature", the mainline branch
will be parsed from it.
4. --- .gitreview
If none of the earlier conditions exist, the .gitreview
"defaultbranch" variable will be used just as before.
Change-Id: I1cdeeaa0944bba3f2e01d7a2039559d0c266f8c9
If you have a development branch for a major project that
will receive gerrit reviews it'll probably be named something
like "development/16/newproject". That will necessitate setting
"defaultbranch=development/16/newproject" in .gitreview. The
make_version script uses that variable to construct the asterisk
version however, which results in versions like
"GIT-development/16/newproject-ee582a8c7b" which is probably not
what you want. Worse, since the download_externals script uses
make_version to construct the URL to download the binary codecs
or DPMA. Since it's expecting a simple numeric version, the
downloads will fail.
To get this to work, a new variable "basebranch" has been added
to .gitreview and make_version has been updated to use that instead
of defaultversion:
.gitreview:
defaultbranch=development/16/myproject
basebranch=16
Now git-review will send the reviews to the proper branch
(development/16/myproject) but the version will still be
constructed using the simple branch number (16).
If "basebranch" is missing from .gitreview, make_version will
fall back to using "defaultbranch".
Change-Id: I2941a3b21e668febeb6cfbc1a7bb51a67726fcc4
Fix the sed(1) invocation used to process git-svn-id not to use "\s"
that is a GNU-ism and is not supported by NetBSD sed. As a result,
this call did not work properly and make_version did output the full
git-svn-id line rather than the revision.
ASTERISK-29852
Change-Id: Ie4b406e2748920643446851a0a252a4ca7245772
download_externals: Add check for i686 and i386 (in addition
to the current x86_64) and exit if not one of the three.
ASTERISK-26497
Change-Id: Ia4d429fcefa5b2f5b6e99159d4607de8e8325b2f
As described in the issue, /tmp is not a suitable location for a
large amount of cached media files, since most distributions make
/tmp a RAM-based tmpfs mount with limited capacity.
I opted for a location that can be configured separately, as opposed
to using a subdirectory of spooldir, given the different storage
profile (transient files vs files that might stay there indefinitely).
This commit just makes the cache directory configurable, and changes
the default location from /tmp to /var/cache/asterisk.
ASTERISK-29143
Change-Id: Ic54e95199405abacd9e509cef5f08fa14c510b5d
This new option can be passed for ./configure or
./tests/CI/buildAsterisk.sh to prevent download/install of binary
modules.
Normally enabling the categories MENUSELECT_CODECS or MENUSELECT_RES
will result in binary modules being enabled even if the build target is
incompatible with those modules. This includes CI scripts which enable
categories before disabling specific modules.
If more binary modules are offered in the future this will help avoid
accidentally downloading them if unwanted or incompatible. Adding a
binary module will only require creating a new menuselect entry similar
to the existing ones, it will not be necessary to modify the CI scripts.
Change-Id: I6b1bd1c75a2e48f05b8b8a45b7a7a2d00a079166
Changing any Menuselect option in the `Compiler Flags` section causes a
full rebuild of the Asterisk source tree. Every enabled option causes
a #define to be added to buildopts.h, thus breaking ccache caching for
every source file that includes "asterisk.h". In most cases each option
only applies to one or two files. Now we only define those options for
the specific sources which use them, this causes much better cache
matching when working with multiple builds. For example testing code
with an without MALLOC_DEBUG will now use just over half the ccache
size, only main/astmm.o will have two builds cached instead of every
file.
Reorder main/Makefile so _ASTCFLAGS set on specific object files are all
together, sorted by filename. Stop adding -DMALLOC_DEBUG to CFLAGS of
bundled pjproject, this define is no longer used by any header so only
serves to break cache.
The only code change is a slight adjustment to how main/astmm.c is
initialized. Initialization functions always exist so main/asterisk.c
can call them unconditionally. Additionally rename the astmm
initialization functions so they are not exported.
Change-Id: Ie2085237a964f6e1e6fff55ed046e2afff83c027
Remove AST_DEVMODE from the AST_BUILDOPTS list and the AST_BUILDOPTS_SUM
calculation as it no longer affects API/ABI compatibility.
Change-Id: Id5bd6dfade173a53b3a49f715586b86e3fb24acb
Use .gitreview defaultbranch setting to determine the mainline branch.
This allows the script to be used against other directories which might
not be on the same defaultbranch. This can be used by CI scripts to
report the testsuite version being used:
./build_tools/make_version ${TESTSUITE_DIR}
Change-Id: Ifdad4a9d8a26138c41bc6b630ecc3e34ea1c2758
I have removed the STATIC_BUILD option immediately as it has not
been maintained in many years and is non-functional.
ASTERISK-27965
Change-Id: I64783d017b86dba9ee3c7bcfb97e59889a3f76d7
Asterisk does not need the development package of libltdl, because it does not
use any symbol of -lltdl directly. Instead, it uses the runtime package via the
shared library -lodbc. On the supported platforms, that shared library declares
its dependency on -lltdl correctly, otherwise AST_EXT_LIB_CHECK would have
failed.
ASTERISK-27745
Change-Id: Icd315809b8e7978203431f3afb66240dd3a040ba
This replaces AST_INLINE_API allocators in utils.h with real functions
implemented in astmm.c. Associated macro's are also moved from utils.h
to astmm.h.
Remove menuselect conflicts between MALLOC_DEBUG and DEBUG_CHAOS as they
can now be combined.
This has multiple benefits:
* Simplifies asterisk/utils.h by removing inline functions and use of
the logger.
* Removal of these inline functions decreases size of Asterisk and
module binaries by 1% or more.
* Puts memory management functions together with and without
MALLOC_DEBUG enabled, simplifying management of the code.
* Enables DEBUG_CHAOS for ASTMM_REDIRECT and bundled pjproject.
Change-Id: If9df4377f74bdbb627461b27a473123e05525887
This allows asterisk to be compiled with MALLOC_DEBUG to load modules
built without MALLOC_DEBUG. Now pre-compiled third-party modules will
still work regardless of MALLOC_DEBUG being enabled or not.
Change-Id: Ic07ad80b2c2df894db984cf27b16a69383ce0e10
Because of a copy-and-paste from the script build_tools/download_externals,
the script build_tools/list_valid_installed_externals got its local variables.
However in the latter, three variables were not used actually.
Change-Id: I252de5a98c17ea54459174875357c22c2eebe8d5
OSX does not support 'readlink -f' or 'sed -r'. Replace readlink with
the GNU make macro 'realpath'. Replace sed with grep in one place, cut
in the other.
ASTERISK-27332
Change-Id: I5d34ecca905384decb22ead45c913ae5e8aff748
This check is being added to make it easier for end-users of third party
open source Opus modules. This was removed by ASTERISK-26426 but only
the module needed to be removed.
Change-Id: I62b9cd0c4fa8a77596ab0e042948a643a1152677
Beanstalkd is a simple to use job queue. It provides a means to
create multiple job queues called "tubes". Each tube can store
multiple jobs, with varying priorities with the queue. Queue
processing is available via a simple TCP socket or via well defined
libraries, avaialble at
https://github.com/kr/beanstalkd/wiki/client-libraries
This module is based upon the beanstalk-client library, available
for download at: https://github.com/deepfryed/beanstalk-client
Change-Id: I5fe4089a34ab3b39230786d9bbfddafa56715f48
The codec/dpma downloader wasn't handling curl correctly. The logic
that transforms makeopts into a bash-sourceable file wasn't
handling the make 'or' command in DOWNLOAD_TIMEOUT so bash was
looking for an 'or' command.
That logic has been eliminated. Instead of trying to transform
and source makeopts, the downloader now calls a make scriptlet
to print the value of a specific variable. This way, make handles
the ors (or any other make construct that happens to creep into
that file).
ASTERISK-27202
Reported by: Sean McCord
Change-Id: Iadfb6693528e4d4da7b8bb201fa66da2c71c7f99
George Joseph
Sean Bright
Sean Bright
Mike Bradeen
George Joseph
Sean Bright
Michał Górny
Josh Soref
Joshua C. Colp
lvl
Corey Farrell
Richard Mudgett
Joshua Colp
Jenkins2
Alexander Traud
Nir Simionovich