Starting in glibc 2.25, the macros `major` and `minor` were moved
from `<sys/types.h>` to `<sys/sysmacros.h>`, to stop leaking these
common words into the global namespace of those including <stdlib.h>
https://sourceware.org/bugzilla/show_bug.cgi?format=multiple&id=19239
The macros MAJOR_IN_MKDEV and MAJOR_IN_SYSMACROS would be defined if
necessary by autoconfig, and put in `config.h`. But osquery build
uses CMake instead of autotools, with a git-committed `config.h`.
This generally is all right across Ubuntu variations, as an automake
for one will likely make a config.h that works on others. However
the glibc change makes a "one-size-fits-all-Ubuntu" config impossible.
This mirrors detection done by autoconf/headers.m4 as CMake code.
It is a relatively simple patch, which unblocks building of librpm
on (K)ubuntu 20 while still working on Ubuntu 18.
This is a followup of PR osquery/osquery#6472,
while the code was patched we were still using the unpatched version.
This should finally fix ossfuzz build failure.
We have ported all platforms to use the source layer.
The facebook layer is not needed anymore since we build everything from source now,
the same is for the transitional source_migration layer.
Minor cleanups about the recently removed python modules.
The set_binding function is incorrectly marked as const,
since it actually modifies one of the struct members.
This was somehow not caught by older compilers,
but is failing on LLVM/Clang 10, so we create a patch to fix this.
This is meant to reduce the noise in the installation output of openssl,
since docs would easily fill the scrollback of most terminals.
Since we don't care about the docs, only install binaries.
Openssl was depending on zlib for SSL/TLS compression,
though it was trying to load it as a dynamic library.
On Windows especially this is an issue because the dll
could possibly be loaded from a insecure place,
so this can be exploited.
Moreover it's proven that compression can lead
to leaks of information; so we completely disable it
to resolve both issues.
Also separated the C from the C++ source code in a different library,
so that the C code is not compiled as C with a C++ compiler.
This would make the macOS C++ compiler fail to compile sleuthkit
since it wasn't how the original build system of the library was building it.
- Add fallback url to download the openssl tar.gz from the "old" archives
- Add URL_HASH to the openssl external project,
to avoid to redownload the archive if it has been alredy downloaded
and the integrity is verified
- Update curl_certificate table to use the newer openssl API,
so that it builds.
Remove a level of indirection when configuring and building formulas.
This should simplify working with them and also remove some issues
encountered when trying to build on Windows.
- Update libarchive to build from source on Windows and macOS
- Update yara to build from source on macOS
- Update librdkafka to build from source on macOS
- Build librdkafka with SSL and SASL_SCRAM support on Linux
- Update librpm to 4.15.1 to support the newer openssl
- Update libxml2 to build from source on Windows and macOS
- Update lzma to build from source on Windows and macOS
- Use ICU library not only for boost but libxml2 too
- Implement a workaround to have Buck builds still compile
with the old openssl version
Implements the Community ID hash that will allow correlating network connections detected by osquery with other tools that support the standard (Zeek, Suricata, etc.).
* Add sccache to the Windows job
Also fix debug symbols on googletest,
so that it doesn't try to create a separate pdb.
* Fix uploading the submodule cache
sccache needs its cache to be uploaded as normal files,
not with a tar, otherwise the cache when extracted
will contain extraneous files and folders (PaxHeader).
The Files mode though by default ignores the .git folder,
so we have to use the .artifactignore file to unignore it,
and we have to copy in the cached path for the cache task to see it.
When using Boost.Process on Windows with Boost.Asio for async_pipe,
the build fails because Asio requires BOOST_ASIO_WINDOWS define
to enable access to async_pipe.
This only partially solves the issue since it's necessary that
Asio users will also include the "sdkddkver.h" header
before the "boost/asio.hpp" header to properly define
_WIN32_WINNT_WIN7, used by osquery.
The move assignment operator of boost::basic_ipstream is bugged.
It has been fixed on upstream in commit
ca994c1972
which is not on any stable build, so we will patch boost with that commit.
