[Buildroot] [PATCH 1/1] package/ltrace: fix musl build

[Buildroot] [PATCH 1/1] package/ltrace: fix musl build

[Fabrice Fontaine]

Fix the following musl build failure raised since commit
bf9583a50276f52edbc37d9f85df5f2cc7fdb5dc:

checking host system type... mips-buildroot-linux-musl

Fixes:
 - http://autobuild.buildroot.org/results/a3a5c46e8562d3f091a9b4b205322168fbf9d16b

Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
---
 package/ltrace/0004-fix-musl-build.patch | 71 ++++++++++++++++++++++++
 1 file changed, 71 insertions(+)
 create mode 100644 package/ltrace/0004-fix-musl-build.patch

diff --git a/package/ltrace/0004-fix-musl-build.patch b/package/ltrace/0004-fix-musl-build.patch
new file mode 100644
index 0000000000..0dceb08ec3
--- /dev/null
+++ b/package/ltrace/0004-fix-musl-build.patch
@@ -0,0 +1,71 @@
+From 2070162fef54fceeb62fd2a6989dde4948f807d7 Mon Sep 17 00:00:00 2001
+From: Fabrice Fontaine <fontaine.fabrice@gmail.com>
+Date: Thu, 2 Nov 2023 16:59:41 +0100
+Subject: [PATCH] fix musl build
+
+Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
+Upstream: unsent yet
+---
+ configure.ac                   | 1 +
+ sysdeps/linux-gnu/mips/plt.c   | 2 ++
+ sysdeps/linux-gnu/ppc/regs.c   | 2 ++
+ sysdeps/linux-gnu/xtensa/plt.c | 2 ++
+ 4 files changed, 7 insertions(+)
+
+diff --git a/configure.ac b/configure.ac
+index c83dd02..8015f0d 100644
+--- a/configure.ac
++++ b/configure.ac
+@@ -33,6 +33,7 @@ AC_CANONICAL_HOST
+ 
+ case "${host_os}" in
+     linux-gnu*) HOST_OS="linux-gnu" ;;
++    linux-musl*) HOST_OS="linux-gnu" ;;
+     linux-uclibc*) HOST_OS="linux-gnu" ;;
+     *)		AC_MSG_ERROR([unkown host-os ${host_os}]) ;;
+ esac
+diff --git a/sysdeps/linux-gnu/mips/plt.c b/sysdeps/linux-gnu/mips/plt.c
+index c7c10ac..ab96961 100644
+--- a/sysdeps/linux-gnu/mips/plt.c
++++ b/sysdeps/linux-gnu/mips/plt.c
+@@ -23,7 +23,9 @@
+ 
+ #include <sys/ptrace.h>
+ #include <errno.h>
++#ifdef __GLIBC__
+ #include <error.h>
++#endif
+ #include <gelf.h>
+ #include <inttypes.h>
+ #include <string.h>
+diff --git a/sysdeps/linux-gnu/ppc/regs.c b/sysdeps/linux-gnu/ppc/regs.c
+index 40d7e7a..71a0d35 100644
+--- a/sysdeps/linux-gnu/ppc/regs.c
++++ b/sysdeps/linux-gnu/ppc/regs.c
+@@ -27,7 +27,9 @@
+ #include <sys/ptrace.h>
+ #include <asm/ptrace.h>
+ #include <errno.h>
++#ifdef __GLIBC__
+ #include <error.h>
++#endif
+ 
+ #include "proc.h"
+ #include "common.h"
+diff --git a/sysdeps/linux-gnu/xtensa/plt.c b/sysdeps/linux-gnu/xtensa/plt.c
+index dd0a0f1..b281510 100644
+--- a/sysdeps/linux-gnu/xtensa/plt.c
++++ b/sysdeps/linux-gnu/xtensa/plt.c
+@@ -19,7 +19,9 @@
+  */
+ 
+ #include <errno.h>
++#ifdef __GLIBC__
+ #include <error.h>
++#endif
+ #include <gelf.h>
+ #include <inttypes.h>
+ #include <string.h>
+-- 
+2.42.0
+
-- 
2.42.0

_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

Re: [Buildroot] [PATCH 1/1] package/ltrace: fix musl build

[Thomas Petazzoni via buildroot]

On Sat,  4 Nov 2023 16:58:57 +0100
Fabrice Fontaine <fontaine.fabrice@gmail.com> wrote:

> Fix the following musl build failure raised since commit
> bf9583a50276f52edbc37d9f85df5f2cc7fdb5dc:
> 
> checking host system type... mips-buildroot-linux-musl
> 
> Fixes:
>  - http://autobuild.buildroot.org/results/a3a5c46e8562d3f091a9b4b205322168fbf9d16b
> 
> Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>

Thanks for solving this issue.

> +Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com>
> +Upstream: unsent yet

Doesn't that defeat the purpose of the Upstream: flag entirely? Isn't
the idea precisely to strongly encourage (not to say "require") that
patches be sent upstream?

Thanks!

Thomas
-- 
Thomas Petazzoni, co-owner and CEO, Bootlin
Embedded Linux and Kernel engineering and training
https://bootlin.com
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot

Re: [Buildroot] [PATCH 1/1] package/ltrace: fix musl build

[Thomas Petazzoni via buildroot]

Hello Fabrice,

On Sat, 4 Nov 2023 18:41:42 +0100
Thomas Petazzoni via buildroot <buildroot@buildroot.org> wrote:

> Doesn't that defeat the purpose of the Upstream: flag entirely? Isn't
> the idea precisely to strongly encourage (not to say "require") that
> patches be sent upstream?

There has still been no feedback from you on this, so I looked into
this a bit more, and it's messy.

ltrace.org points to https://gitlab.com/cespedes/ltrace as being the
Git repo, and this Git repo has changes from a year ago, but absolutely
no tags.

On the other side, Buildroot uses https://github.com/dkogan/ltrace/tree/master
as the upstream... which hasn't seen any commit for the last 8 years.
Though this one has tags, including a 0.7.3 tag, which it the latest
version according to ltrace.org.

And when trying to figure if one if a fork from the other... it's not
clear.

Regarding your patch, you made the inclusion of <error.h> conditional
on __GLIBC__, but what if the code is using functions from <error.h>?
If it builds without <error.h>, then presumably it means the code is no
longer using any function from <error.h>, and those includes can be
dropped?

Alpine had a different patch, that adds a configure.ac check for
<error.h>: https://git.alpinelinux.org/aports/plain/main/ltrace/musl.patch

However, their patch is for 0.7.3, while we use the "latest" commit
from master.

So overall, quite messy stuff, so I decided to mark ltrace as
unavailable for musl:

  https://gitlab.com/buildroot.org/buildroot/-/commit/859bd545821a838454577c4c1c9106837cda51cb

Best regards,

Thomas
-- 
Thomas Petazzoni, co-owner and CEO, Bootlin
Embedded Linux and Kernel engineering and training
https://bootlin.com
_______________________________________________
buildroot mailing list
buildroot@buildroot.org
https://lists.buildroot.org/mailman/listinfo/buildroot