| Message ID | 20191209222746.3440986-3-fontaine.fabrice@gmail.com |
|---|---|
| State | Superseded |
| Headers | show |
| Series | None | expand |
Hello Fabrice, On Mon, 9 Dec 2019 23:27:46 +0100 Fabrice Fontaine <fontaine.fabrice@gmail.com> wrote: > +As a result, $ERRVALFILE does not contain the expected values because > +the expected value is not on the same line than gaierrval: > + > + gaierrval: > + # 130 "conftest.c" 3 4 > + -3 I am not seeing this issue here, the relevant part of the .i file looks like this: # 129 "conftest.c" 2 int main () { gaierrval: -11 ; return 0; } It's quite odd that you get this # 130 conftest.c line in the middle of the .c file dump. What are we doing differently ? Is it a compiler version issue ? Which toolchain are you using when you're having the issue ? Thanks, Thomas
Hello Thomas, Le sam. 14 déc. 2019 à 22:55, Thomas Petazzoni <thomas.petazzoni@bootlin.com> a écrit : > > Hello Fabrice, > > On Mon, 9 Dec 2019 23:27:46 +0100 > Fabrice Fontaine <fontaine.fabrice@gmail.com> wrote: > > > +As a result, $ERRVALFILE does not contain the expected values because > > +the expected value is not on the same line than gaierrval: > > + > > + gaierrval: > > + # 130 "conftest.c" 3 4 > > + -3 > > I am not seeing this issue here, the relevant part of the .i file looks > like this: > > # 129 "conftest.c" 2 > int > main () > { > > > gaierrval: -11 > > > > ; > return 0; > } > > It's quite odd that you get this # 130 conftest.c line in the middle of > the .c file dump. What are we doing differently ? Is it a compiler > version issue ? > > Which toolchain are you using when you're having the issue ? I got this build failure with the defconfig that I put in the "Fixes:" statement (i.e. toolchain-external-linaro-aarch64 from http://autobuild.buildroot.org/results/8a9/8a92d5257ba3e6fd0ad56ddb0afad8b6edd40ae3/defconfig). I have the same error with br-x86-64-musl-2019.05.1.tar.bz2 from http://autobuild.buildroot.org/results/434/434393662f73118efb4b7113a48aaf40e9484a3c/defconfig. So I don't really understand what's going on. If you want, I can send a v2 without this patch and we'll see if build failures are raised on the autobuilders. > > Thanks, > > Thomas > -- > Thomas Petazzoni, CTO, Bootlin > Embedded Linux and Kernel engineering > https://bootlin.com Best Regards, Fabrice
diff --git a/package/dante/0002-compiler.m4-do-not-remove-g-flag.patch b/package/dante/0002-compiler.m4-do-not-remove-g-flag.patch index 0b41de6192..a56fe297cb 100644 --- a/package/dante/0002-compiler.m4-do-not-remove-g-flag.patch +++ b/package/dante/0002-compiler.m4-do-not-remove-g-flag.patch @@ -13,6 +13,8 @@ after -g negates the -g, and having -ggdb after -g is enough to produce debugging information for GDB. Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com> +[Fabrice: update for 1.4.2] +Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com> --- compiler.m4 | 6 ------ 1 file changed, 6 deletions(-) @@ -21,18 +23,18 @@ diff --git a/compiler.m4 b/compiler.m4 index 3a23406..621372c 100644 --- a/compiler.m4 +++ b/compiler.m4 -@@ -378,15 +378,9 @@ else +@@ -443,15 +443,9 @@ else gcc) if test x"$aixldbug" != x; then #disable debug info -- if echo $CFLAGS | grep -- "-g" >/dev/null; then -- CFLAGS="`echo $CFLAGS | sed -e 's/-g//g'`" +- if echo $CFLAGS | grep -- "-g " >/dev/null; then +- CFLAGS="`echo $CFLAGS | sed -e 's/-g //g'`" - fi CFLAGS="$CFLAGS${CFLAGS:+ }-g0" else #use -ggdb also when not debugging -- if echo $CFLAGS | grep -- "-g" >/dev/null; then -- CFLAGS="`echo $CFLAGS | sed -e 's/-g//g'`" +- if echo $CFLAGS | grep -- "-g " >/dev/null; then +- CFLAGS="`echo $CFLAGS | sed -e 's/-g //g'`" - fi CFLAGS="$CFLAGS${CFLAGS:+ }-ggdb" fi diff --git a/package/dante/0003-osdep-m4-Remove-getaddrinfo-too-low-checks.patch b/package/dante/0003-osdep-m4-Remove-getaddrinfo-too-low-checks.patch new file mode 100644 index 0000000000..af23d46d11 --- /dev/null +++ b/package/dante/0003-osdep-m4-Remove-getaddrinfo-too-low-checks.patch @@ -0,0 +1,35 @@ +osdep.m4: Remove getaddrinfo() too low checks + +dante runs AC_PREPROC_IFELSE then it fills $ERRVALFILE by running the +following command: + +cat conftest.i | grep gaierrval: >>$2 + +As a result, $ERRVALFILE does not contain the expected values because +the expected value is not on the same line than gaierrval: + + gaierrval: + # 130 "conftest.c" 3 4 + -3 + +So drop these checks + +Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com> + +diff -Nuar dante-1.4.2-orig/osdep.m4 dante-1.4.2/osdep.m4 +--- dante-1.4.2-orig/osdep.m4 2019-12-09 21:28:38.936003218 +0100 ++++ dante-1.4.2/osdep.m4 2019-12-09 21:49:48.764019371 +0100 +@@ -956,13 +956,7 @@ + unset UNIQUEVALS UNIQUESYMBOLS + if test -s $ERRVALFILE; then + UNIQUEVALS=`sort $ERRVALFILE | uniq | wc -l | awk '{ print $1 }'` +- if test $UNIQUEVALS -le 1; then +- AC_MSG_FAILURE([error: getaddrinfo() error value count too low]) +- fi + UNIQUESYMBOLS=`cat $ERRVALFILE | wc -l | awk '{ print $1 }'` +- if test $UNIQUESYMBOLS -le 1; then +- AC_MSG_FAILURE([error: getaddrinfo() error symbol count too low]) +- fi + + if test $ERRNOCNT -ne $UNIQUESYMBOLS; then + AC_MSG_FAILURE([internal error: errno symbol count mismatch]) diff --git a/package/dante/dante.hash b/package/dante/dante.hash index 38dea61b0a..cb87cdc04a 100644 --- a/package/dante/dante.hash +++ b/package/dante/dante.hash @@ -1,2 +1,5 @@ -# Locally computed -sha256 b6d232bd6fefc87d14bf97e447e4fcdeef4b28b16b048d804b50b48f261c4f53 dante-1.4.1.tar.gz +# From https://www.inet.no/dante/download.html +sha256 4c97cff23e5c9b00ca1ec8a95ab22972813921d7fbf60fc453e3e06382fc38a7 dante-1.4.2.tar.gz + +# Hash for license file +sha256 954ab6dbcf994711e28e603d8657d6eceef333cd6f9ca6705f0e9d118ab7d69a LICENSE diff --git a/package/dante/dante.mk b/package/dante/dante.mk index 35b785dc70..14b21b13bf 100644 --- a/package/dante/dante.mk +++ b/package/dante/dante.mk @@ -4,7 +4,7 @@ # ################################################################################ -DANTE_VERSION = 1.4.1 +DANTE_VERSION = 1.4.2 DANTE_SITE = http://www.inet.no/dante/files DANTE_LICENSE = BSD-3-Clause DANTE_LICENSE_FILES = LICENSE
- Add hash for license file - Refresh second patch - Add third patch - This bump fixes the build failure with latest libminiupnpc Fixes: - http://autobuild.buildroot.org/results/8a92d5257ba3e6fd0ad56ddb0afad8b6edd40ae3 Signed-off-by: Fabrice Fontaine <fontaine.fabrice@gmail.com> --- ...002-compiler.m4-do-not-remove-g-flag.patch | 12 ++++--- ...m4-Remove-getaddrinfo-too-low-checks.patch | 35 +++++++++++++++++++ package/dante/dante.hash | 7 ++-- package/dante/dante.mk | 2 +- 4 files changed, 48 insertions(+), 8 deletions(-) create mode 100644 package/dante/0003-osdep-m4-Remove-getaddrinfo-too-low-checks.patch