Qt help files are used by Qt Creator and KDevelop, for example, to support
the development of Qt-based applications and libraries.
Generating api documentation in Qt help format is controlled by two
user specific options named --enable-qt-help and --with-qchdir (autotools)
and -DENABLE_QT_HELP and -DINSTALL_QCH_DIR (cmake).
Marking targets with a component would only be useful if we
marked every target with a component in a consistent way,
but because we don't do that, it's pointless to have it
in just a few places.
For cleaning purpose the event list members are initialized with
WSA_INVALID_EVENT. The cleanup code detects and handles the
case that the event list has been created from calloc ().
The function _dbus_poll() has been split into two functions,
_dbus_poll_events() and _dbus_poll_select(), each containing the
corresponding implementation.
_dbus_poll() now calls the corresponding function.
Some of the command-lines that we print as diagnostics contain newlines,
which will cause warnings or errors under a strict TAP parser (and one of
them wasn't correctly prefixed with '#' anyway). TAP parsers only parse
stdout, not stderr, so we can use stderr for these diagnostic messages.
[smcv: Expand commit message]
Signed-off-by: Simon McVittie <smcv@collabora.com>
Otherwise, Coverity will diagnose this as a resource leak,
because it doesn't understand that our assertions end up guaranteeing
that the result is freed if and only if it's non-`NULL`.
Coverity CID: 354884
Assertions can be disabled, but in test code the assertions are the
whole point, so use checks that can't be disabled instead.
Because there's a lot of test code, I haven't done this globally, only
in the tests that I recently converted from "embedded" to "modular".
Signed-off-by: Simon McVittie <smcv@collabora.com>
Assertions can be disabled; but in test code the assertions are the
entire point, so we don't want to disable them. Use _dbus_test_fatal()
instead.
test-service is actually a test helper rather than a test, so use its
pre-existing die() function instead.
Signed-off-by: Simon McVittie <smcv@collabora.com>
These previously relied on embedding test-specific code in libdbus,
but they actually only need public APIs, private interfaces that get
exported anyway for the benefit of dbus-daemon, and the TAP helpers;
so we can run them even in production builds.
Signed-off-by: Simon McVittie <smcv@collabora.com>
This lets us run a subset of the tests that previously relied on extra
test-only code being compiled into libdbus.
Signed-off-by: Simon McVittie <smcv@collabora.com>
This avoids unexpected precedence when used as the body of an "if" or
"else" without being wrapped in {}.
Signed-off-by: Simon McVittie <smcv@collabora.com>
Debian 8 'jessie' has reached EOL for mainstream security support,
which puts it outside our usual distribution/OS support policy.
Signed-off-by: Simon McVittie <smcv@collabora.com>
Previously, we built on Debian 9 'stretch' by default, and on
Debian 10 'buster' only on request. Let's reverse that so that we get
more modern toolchains, before Debian 9 'stretch' reaches EOL.
Signed-off-by: Simon McVittie <smcv@collabora.com>
POSIX.1-2001 and POSIX.1-2008 specifies include <poll.h> so use that
rather than the non-standard/legacy include <sys/poll.h>.
This fixes the following warnings when building with musl libc:
1 | #warning redirecting incorrect #include <sys/poll.h> to <poll.h>
| ^~~~~~~
Signed-off-by: Natanael Copa <ncopa@alpinelinux.org>
Use getrandom(2) and fall back to /dev/urandom if it is missing or if it
fails some any reason.
This solves problem where dbus-uuidgen is called from a chroot which
lacks /dev/urandom.
Signed-off-by: Natanael Copa <ncopa@alpinelinux.org>
As suggested on !143. Instead of remembering to free it in every error
condition, let's move its cleanup to the "out" phase so that it's done
every time.
Change the iterator variable tmp to be const so that it's obvious we
aren't meant to free that too.
Signed-off-by: Simon McVittie <smcv@collabora.com>
If we ran out of memory while handling connect() errors, we didn't
free the linked list of struct addrinfo. Move their cleanup to the
"out" phase of the function so that we always do it.
While I'm there, change the iterator variable tmp to be const, to make
it more obvious that we aren't meant to free it.
This is similar to commit 00badeba (!143) in the corresponding Windows
code path, but with some refactoring.
Signed-off-by: Simon McVittie <smcv@collabora.com>
If there is no more memory available within the mentiond function, e.g.,
when checking memory management, the release of memory allocated by
getaddrinfo() is missing.
Coverity CID: 354880
