| Message ID | 20240517120057.3736694-1-yegorslists@googlemail.com |
|---|---|
| State | New |
| Headers | show |
| Series | package/ntp: fix build with GCC 14.x | expand |
Yegor, All, On 2024-05-17 14:00 +0200, yegorslists--- via buildroot spake thusly: > From: Yegor Yefremov <yegorslists@googlemail.com> > > Fix pthread_detach() detection that is broken when building > with GCC 14.x. > > The patch is taken from a Gentoo repository: > > https://gitweb.gentoo.org/repo/gentoo.git/diff/net-misc/ntp/files/ntp-4.2.8_p15-configure-clang16.patch?id=56ef5ce33891feec34cd0c39a0a774a81a587dc7 > > Bonus: refresh the 0002-ntp-syscalls-fallback.patch file. Sure, but it's relatively obvious, by looking and the changeset, that the patch was modified; the interesting information to have in the commit log is why it was refreshed: Addtionally, refresh patch 0002 to fix the offset of hunk 1. > Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com> > --- > package/ntp/0002-ntp-syscalls-fallback.patch | 9 ++++--- > .../ntp/0003-fix-pthread-detach-check.patch | 27 +++++++++++++++++++ $ make check-package package/ntp/0003-fix-pthread-detach-check.patch:0: missing Upstream in the header (https://nightly.buildroot.org/#_additional_patch_documentation) Yes, you do provide the location where you grabbed the patch from, and this is very good. Still, we need to know what the upstream status for that patch is: has it been submitted? If no, why? If yes, what is the URL to the PR/MR, bug report, email? It is a bit unfortunate that ntp uses bitkeeper as a VCS, which makes it a bit more difficult to contribute to, but we still need that in formation... I see that the last commits dates from just a few days ago, so the project is not dead. Could you please check whether this patch was submitted? Maybe there is some info about that in the gentoo history for that patch? > diff --git a/package/ntp/0002-ntp-syscalls-fallback.patch b/package/ntp/0002-ntp-syscalls-fallback.patch > index 535fc93dfd..46704b9d3f 100644 > --- a/package/ntp/0002-ntp-syscalls-fallback.patch > +++ b/package/ntp/0002-ntp-syscalls-fallback.patch > @@ -8,10 +8,11 @@ Signed-off-by: James Knight <james.knight@rockwellcollins.com> > configure.ac | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > -diff --git a/configure.ac b/configure.ac > ---- a/configure.ac > -+++ b/configure.ac > -@@ -1523,11 +1523,11 @@ AC_CACHE_CHECK( > +Index: ntp-4.2.8p17/configure.ac > +=================================================================== > +--- ntp-4.2.8p17.orig/configure.ac > ++++ ntp-4.2.8p17/configure.ac > +@@ -1472,11 +1472,11 @@ AC_CACHE_CHECK( Just fixing the line would have made for a smaller delta, with less useless changes in the diff metadata. Also, the patch was (almost) git-formatted, which is the preferred form for bundled patches. Unless it is submitted to, or backported from the upstream VCS, in which case it should be in a format supported by that VCS. ;-) Regards, Yann E. MORIN.
Hi Yann, On Fri, May 17, 2024 at 8:25 PM Yann E. MORIN <yann.morin.1998@free.fr> wrote: > > Yegor, All, > > On 2024-05-17 14:00 +0200, yegorslists--- via buildroot spake thusly: > > From: Yegor Yefremov <yegorslists@googlemail.com> > > > > Fix pthread_detach() detection that is broken when building > > with GCC 14.x. > > > > The patch is taken from a Gentoo repository: > > > > https://gitweb.gentoo.org/repo/gentoo.git/diff/net-misc/ntp/files/ntp-4.2.8_p15-configure-clang16.patch?id=56ef5ce33891feec34cd0c39a0a774a81a587dc7 > > > > Bonus: refresh the 0002-ntp-syscalls-fallback.patch file. > > Sure, but it's relatively obvious, by looking and the changeset, that > the patch was modified; the interesting information to have in the > commit log is why it was refreshed: > > Addtionally, refresh patch 0002 to fix the offset of hunk 1. > > > Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com> > > --- > > package/ntp/0002-ntp-syscalls-fallback.patch | 9 ++++--- > > .../ntp/0003-fix-pthread-detach-check.patch | 27 +++++++++++++++++++ > > $ make check-package > package/ntp/0003-fix-pthread-detach-check.patch:0: missing Upstream in the header (https://nightly.buildroot.org/#_additional_patch_documentation) > > Yes, you do provide the location where you grabbed the patch from, and > this is very good. Still, we need to know what the upstream status for > that patch is: has it been submitted? If no, why? If yes, what is the > URL to the PR/MR, bug report, email? > > It is a bit unfortunate that ntp uses bitkeeper as a VCS, which makes it > a bit more difficult to contribute to, but we still need that in > formation... > > I see that the last commits dates from just a few days ago, so the > project is not dead. > > Could you please check whether this patch was submitted? Maybe there is > some info about that in the gentoo history for that patch? > > > diff --git a/package/ntp/0002-ntp-syscalls-fallback.patch b/package/ntp/0002-ntp-syscalls-fallback.patch > > index 535fc93dfd..46704b9d3f 100644 > > --- a/package/ntp/0002-ntp-syscalls-fallback.patch > > +++ b/package/ntp/0002-ntp-syscalls-fallback.patch > > @@ -8,10 +8,11 @@ Signed-off-by: James Knight <james.knight@rockwellcollins.com> > > configure.ac | 4 ++-- > > 1 file changed, 2 insertions(+), 2 deletions(-) > > > > -diff --git a/configure.ac b/configure.ac > > ---- a/configure.ac > > -+++ b/configure.ac > > -@@ -1523,11 +1523,11 @@ AC_CACHE_CHECK( > > +Index: ntp-4.2.8p17/configure.ac > > +=================================================================== > > +--- ntp-4.2.8p17.orig/configure.ac > > ++++ ntp-4.2.8p17/configure.ac > > +@@ -1472,11 +1472,11 @@ AC_CACHE_CHECK( > > Just fixing the line would have made for a smaller delta, with less > useless changes in the diff metadata. > > Also, the patch was (almost) git-formatted, which is the preferred form > for bundled patches. Unless it is submitted to, or backported from the > upstream VCS, in which case it should be in a format supported by that > VCS. ;-) Now, it is official: https://bugs.ntp.org/show_bug.cgi?id=3926 I'll wait till the patch is committed and then will make v2. Yegor
Yegor, All, On 2024-05-18 12:00 +0200, Yegor Yefremov spake thusly: > Now, it is official: https://bugs.ntp.org/show_bug.cgi?id=3926 That's great, thanks! And upstream looks quite happy with it. :-) > I'll wait till the patch is committed and then will make v2. You don't need to wait for the patch to be applied upstream; just add an upstream tag to the patch, pointing to that bug report: Upstream: https://bugs.ntp.org/show_bug.cgi?id=3926 Signed-off-by: You That way, the next person to bump ntp can assess whether that patch needs to be carried/rebased/dropped, by checking the upstream status. Thanks! Regards, Yann E. MORIN.
Yann, All, On Sat, May 18, 2024 at 12:16 PM Yann E. MORIN <yann.morin.1998@free.fr> wrote: > > Yegor, All, > > On 2024-05-18 12:00 +0200, Yegor Yefremov spake thusly: > > Now, it is official: https://bugs.ntp.org/show_bug.cgi?id=3926 > > That's great, thanks! And upstream looks quite happy with it. :-) > > > I'll wait till the patch is committed and then will make v2. > > You don't need to wait for the patch to be applied upstream; just add an > upstream tag to the patch, pointing to that bug report: > > Upstream: https://bugs.ntp.org/show_bug.cgi?id=3926 > Signed-off-by: You > > That way, the next person to bump ntp can assess whether that patch > needs to be carried/rebased/dropped, by checking the upstream status. We are almost there. The fix landed at the maintainer's personal repo. Another show stopper with GCC 14.x is lrzsz. There are a lot of issues described in this one [1] and without any feedback :-( [1] https://github.com/UweOhse/lrzsz/issues/3 Yegor
diff --git a/package/ntp/0002-ntp-syscalls-fallback.patch b/package/ntp/0002-ntp-syscalls-fallback.patch index 535fc93dfd..46704b9d3f 100644 --- a/package/ntp/0002-ntp-syscalls-fallback.patch +++ b/package/ntp/0002-ntp-syscalls-fallback.patch @@ -8,10 +8,11 @@ Signed-off-by: James Knight <james.knight@rockwellcollins.com> configure.ac | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) -diff --git a/configure.ac b/configure.ac ---- a/configure.ac -+++ b/configure.ac -@@ -1523,11 +1523,11 @@ AC_CACHE_CHECK( +Index: ntp-4.2.8p17/configure.ac +=================================================================== +--- ntp-4.2.8p17.orig/configure.ac ++++ ntp-4.2.8p17/configure.ac +@@ -1472,11 +1472,11 @@ AC_CACHE_CHECK( [ntp_cv_var_ntp_syscalls], [ ntp_cv_var_ntp_syscalls=no diff --git a/package/ntp/0003-fix-pthread-detach-check.patch b/package/ntp/0003-fix-pthread-detach-check.patch new file mode 100644 index 0000000000..c61b6cc736 --- /dev/null +++ b/package/ntp/0003-fix-pthread-detach-check.patch @@ -0,0 +1,27 @@ +fix pthread_detach check + +Refrence: + https://gitweb.gentoo.org/repo/gentoo.git/diff/net-misc/ntp/files/ntp-4.2.8_p15-configure-clang16.patch?id=56ef5ce33891feec34cd0c39a0a774a81a587dc7 + +Signed-off-by: Yegor Yefremov <yegorslists@googlemail.com> +--- + sntp/m4/openldap-thread-check.m4 | 6 ++---- + 1 file changed, 2 insertions(+), 4 deletions(-) + +Index: ntp-4.2.8p17/sntp/m4/openldap-thread-check.m4 +=================================================================== +--- ntp-4.2.8p17.orig/sntp/m4/openldap-thread-check.m4 ++++ ntp-4.2.8p17/sntp/m4/openldap-thread-check.m4 +@@ -262,10 +262,8 @@ pthread_rwlock_t rwlock; + dnl save the flags + AC_LINK_IFELSE([AC_LANG_PROGRAM([[ + #include <pthread.h> +-#ifndef NULL +-#define NULL (void*)0 +-#endif +-]], [[pthread_detach(NULL);]])],[ol_cv_func_pthread_detach=yes],[ol_cv_func_pthread_detach=no]) ++pthread_t thread; ++]], [[pthread_detach(thread);]])],[ol_cv_func_pthread_detach=yes],[ol_cv_func_pthread_detach=no]) + ]) + + if test $ol_cv_func_pthread_detach = no ; then