summaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorThomas Haller <thaller@redhat.com>2019-06-25 18:17:44 +0200
committerThomas Haller <thaller@redhat.com>2019-06-26 09:53:54 +0200
commitbf6e902c90129a86efd2e2ae223256ca30267450 (patch)
treec76a1be80d5b6f773ca67d7fb2fade13b9667a0d
parente4ce9bd7af6a39677ff1a1380906d18062abb24a (diff)
downloadNetworkManager-bf6e902c90129a86efd2e2ae223256ca30267450.tar.gz
CONTRIBUTING: update section about assertions in NetworkManager
-rw-r--r--CONTRIBUTING49
1 files changed, 32 insertions, 17 deletions
diff --git a/CONTRIBUTING b/CONTRIBUTING
index 68026a81b5..febfdf0439 100644
--- a/CONTRIBUTING
+++ b/CONTRIBUTING
@@ -55,51 +55,66 @@ There are different kind of assertions. Use the one that is appropriate.
1) g_return_*() from glib. This is usually enabled in release builds and
can be disabled with G_DISABLE_CHECKS define. This uses g_log() with
G_LOG_LEVEL_CRITICAL level (which allows the program to continue,
- until G_DEBUG=fatal-criticals or G_DEBUG=fatal-warnings is set). As such,
- this is the preferred way for assertions that are commonly enabled.
-
- Make a mild attempt to work around such assertion failure, but don't try
- to hard. A failure of g_return_*() assertion might allow the process
- to continue, but there is no guarantee.
+ unless G_DEBUG=fatal-criticals or G_DEBUG=fatal-warnings is set). As such,
+ this is usually the preferred way for assertions that are supposed to be
+ enabled by default.
+
+ Optimally, after a g_return_*() failure the program can still continue. This is
+ also the reason why g_return_*() is preferable over g_assert().
+ For example, that is often not given for functions that return a GError, because
+ g_return_*() will return failure without setting the error output. That often leads
+ to a crash immidiately after, because the caller requires the GError to be set.
+ Make a reasonable effort so that an assertion failure may allow the process
+ to proceed. But don't put too much effort in it. After all, it's an assertion
+ failure that is not supposed to happen either way.
2) nm_assert() from NetworkManager. This is disabled by default in release
builds, but enabled if you build --with-more-assertions. See "WITH_MORE_ASSERTS"
define. This is preferred for assertions that are expensive to check or
- nor necessary to check frequently (stuff that is really not expected to
- fail). Use this deliberately and assume it's not present in production builds.
+ nor necessary to check frequently. It's also for conditions that can easily
+ verified to be true and where future refactoring is unlikley to break that
+ condition.
+ Use this deliberately and assume it is removed from production builds.
3) g_assert() from glib. This is used in unit tests and commonly enabled
in release builds. It can be disabled with G_DISABLE_ASSERT assert
define. Since this results in a hard crash on assertion failure, you
- should almost always prefer g_return_*() over this (except unit tests).
+ should almost always prefer g_return_*() over this (except in unit tests).
4) assert() from <assert.h>. It is usually enabled in release builds and
can be disabled with NDEBUG define. Don't use it in NetworkManager,
it's basically like g_assert().
-5) g_log() from glib. These are always compiled in, depending on the levels
+5) g_log() from glib. These are always compiled in, depending on the logging level
these are assertions too. G_LOG_LEVEL_ERROR aborts the program, G_LOG_LEVEL_CRITICAL
logs a critical warning (like g_return_*(), see G_DEBUG=fatal-criticals)
and G_LOG_LEVEL_WARNING logs a warning (see G_DEBUG=fatal-warnings).
G_LOG_LEVEL_DEBUG level is usually not printed, unless G_MESSAGES_DEBUG environment
is set.
- In general, avoid using g_log() in NetworkManager. We have nm-logging instead.
+ In general, avoid using g_log() in NetworkManager. We have nm-logging instead
+ which logs to syslog/systemd-journald.
From a library like libnm it might make sense to log warnings (if someting
is really wrong) or debug messages. But better don't. If it's important,
find a way to report the notification via the API to the caller. If it's
not important, keep silent.
+ In particular, don't use levels G_LOG_LEVEL_CRITICAL and G_LOG_LEVEL_WARNING because
+ these are effectively assertions and we want to run with G_DEBUG=fatal-warnings.
6) g_warn_if_*() from glib. These are always compiled in and log a G_LOG_LEVEL_WARNING
warning. Don't use this.
7) G_TYPE_CHECK_INSTANCE_CAST() from glib. Unless building with "WITH_MORE_ASSERTS",
- we disable G_DISABLE_CAST_CHECKS. This means, cast macros like NM_DEVICE(ptr)
- translate to plain C pointer casts. Use the cast macros deliberately, in production
- code they are cheap, with debugging enabled they assert that the pointer is valid.
+ we set G_DISABLE_CAST_CHECKS. This means, cast macros like NM_DEVICE(ptr)
+ translate to plain C pointer casts. Use such cast macros deliberately, in production
+ code they are cheap, with more asserts enabled the check that the pointer type is
+ suitable.
+
+Of course, every assertion failure is a bug, and calling it must have no side effects.
-Of course, every assertion failure is a bug, and they must not have side effects.
Theoretically, you are welcome to disable G_DISABLE_CHECKS and G_DISABLE_ASSERT
in production builds. In practice, nobody tests such a configuration, so beware.
-For testing, you also want to run NetworkManager with G_DEBUG=fatal-warnings
-to crash un critical and warn g_log() messages.
+For testing, you also want to run NetworkManager with environment variable
+G_DEBUG=fatal-warnings to crash upon G_LOG_LEVEL_CRITICAL and G_LOG_LEVEL_WARNING
+g_log() message. NetworkManager won't use these levels for regular logging
+but for assertions.