| Message ID | 1438771429-15145-1-git-send-email-joerg.krause@embedded.rocks |
|---|---|
| State | Accepted |
| Headers | show |
Dear Jörg Krause, On Wed, 5 Aug 2015 12:43:49 +0200, Jörg Krause wrote: > Add a patch from the Alpine Linux project [1] to fix a musl build issue with > gcc 5: > > ``` > error: temporary of non-literal type ‘pthread_mutex_t’ in a constant expression > constexpr PosixMutex():mutex(PTHREAD_MUTEX_INITIALIZER) {} > ``` > > Problem has been reported by the Alpine team upstream and was closed by the MPD > maintainer with WONTFIX: > http://bugs.musicpd.org/view.php?id=4387 > http://bugs.musicpd.org/view.php?id=4110 > > However... > > POSIX does not permit using PTHREAD_COND_INITIALIZER except for static > initialization, and certainly does not permit using it as a value. > > Also POSIX does not specify the type of the object (it's opaque) so if > there are any types for which their code would be invalid C++, then their > code is invalid. > > Also, volatile in the type is necessary. without that, LTO can break the code. > > [1] > http://git.alpinelinux.org/cgit/aports/log/main/mpd/musl-gcc5-fixes.patch > > Signed-off-by: Jörg Krause <joerg.krause@embedded.rocks> > --- > package/mpd/0002-musl-gcc5-fixes.patch | 56 ++++++++++++++++++++++++++++++++++ > 1 file changed, 56 insertions(+) > create mode 100644 package/mpd/0002-musl-gcc5-fixes.patch Applied, thanks. To be honest, I hesitated quite a bit before applying this patch. A Buildroot rule is that we don't go against upstream. And here upstream clearly say that they don't want to do the necessary effort to support musl, so what we should have done instead is to also make mpd not available with the musl C library. But the patch is quite simple, so I made an exception. However, if more patches are needed in mpd to fix musl build problems, I will reject them and mark the package as not available in musl. Unhappy users can go complain to the mpd developers that their decision of not supporting musl by doing a configure-time check is stupid. Thanks! Thomas
diff --git a/package/mpd/0002-musl-gcc5-fixes.patch b/package/mpd/0002-musl-gcc5-fixes.patch new file mode 100644 index 0000000..a4c65cc --- /dev/null +++ b/package/mpd/0002-musl-gcc5-fixes.patch @@ -0,0 +1,56 @@ +musl gcc5 fixes + +Fetch from: +http://git.alpinelinux.org/cgit/aports/plain/main/mpd/musl-gcc5-fixes.patch + +Problem has been reported upstream and closed with WONTFIX: +http://bugs.musicpd.org/view.php?id=4387 +http://bugs.musicpd.org/view.php?id=4110 + +however... + +POSIX does not permit using PTHREAD_COND_INITIALIZER except for static +initialization, and certainly does not permit using it as a value + +also POSIX does not specify the type of the object (it's opaque) so if +there are any types for which their code would be invalid C++, then their +code is invalid + +also, volatile in the type is necessary. without that, LTO can break the code. + +Signed-off-by: Jörg Krause <joerg.krause@embedded.rocks> + +--- ./src/notify.hxx.orig ++++ ./src/notify.hxx +@@ -28,7 +28,7 @@ + Cond cond; + bool pending; + +-#if !defined(WIN32) && !defined(__NetBSD__) && !defined(__BIONIC__) ++#if defined(__GLIBC__) + constexpr + #endif + notify():pending(false) {} +--- ./src/thread/PosixCond.hxx.orig ++++ ./src/thread/PosixCond.hxx +@@ -41,7 +41,7 @@ + pthread_cond_t cond; + + public: +-#if defined(__NetBSD__) || defined(__BIONIC__) ++#if !defined(__GLIBC__) + /* NetBSD's PTHREAD_COND_INITIALIZER is not compatible with + "constexpr" */ + PosixCond() { +--- ./src/thread/PosixMutex.hxx.orig ++++ ./src/thread/PosixMutex.hxx +@@ -41,7 +41,7 @@ + pthread_mutex_t mutex; + + public: +-#if defined(__NetBSD__) || defined(__BIONIC__) ++#if !defined(__GLIBC__) + /* NetBSD's PTHREAD_MUTEX_INITIALIZER is not compatible with + "constexpr" */ + PosixMutex() { +
Add a patch from the Alpine Linux project [1] to fix a musl build issue with gcc 5: ``` error: temporary of non-literal type ‘pthread_mutex_t’ in a constant expression constexpr PosixMutex():mutex(PTHREAD_MUTEX_INITIALIZER) {} ``` Problem has been reported by the Alpine team upstream and was closed by the MPD maintainer with WONTFIX: http://bugs.musicpd.org/view.php?id=4387 http://bugs.musicpd.org/view.php?id=4110 However... POSIX does not permit using PTHREAD_COND_INITIALIZER except for static initialization, and certainly does not permit using it as a value. Also POSIX does not specify the type of the object (it's opaque) so if there are any types for which their code would be invalid C++, then their code is invalid. Also, volatile in the type is necessary. without that, LTO can break the code. [1] http://git.alpinelinux.org/cgit/aports/log/main/mpd/musl-gcc5-fixes.patch Signed-off-by: Jörg Krause <joerg.krause@embedded.rocks> --- package/mpd/0002-musl-gcc5-fixes.patch | 56 ++++++++++++++++++++++++++++++++++ 1 file changed, 56 insertions(+) create mode 100644 package/mpd/0002-musl-gcc5-fixes.patch