Bug#1067561: FTBFS: Error: symbol `open64' is already defined
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 On Fri, 2024-04-19 at 00:01 +0100, Peter Green wrote: > Thanks, upstream has now accepted a patch that takes a slightly different > approach to fixing the issue. > > https://github.com/canonical/lightdm/issues/352 Yes I saw. That's why I think this should have been reported and talked directly with upstream, before the NMU, but eh... > > Could we get this uploaded to sid, so that the lightweight desktops > are installable on armel/armhf again? Yes I'll look into it. Regards, - -- Yves-Alexis -BEGIN PGP SIGNATURE- iQEzBAEBCAAdFiEE8vi34Qgfo83x35gF3rYcyPpXRFsFAmYiRuIACgkQ3rYcyPpX RFsQkQgA1h8bIbWp0Sj6VqgcdG/dkwKEQ6Xrcuh6gpL7Gy2BiT2d1tY4HFN94bhG FJVz9E9QOqlZt2Mo7JQ6EBYyvcTspO0T09ZU/22VjA2NdP0eSRHTiyiFXq4hyoYf OBnPpKTxq0Wp9L7Tefta7uNPIBX+lsVlw9cdwv6PIEy6jlF3xjX+B8FoK/K/fIeY FuQh5qOw/k/nnPAM4pHxYGAfRw+BNBEnI3ADFCCWDvpQSgZ9uJ9UB3wnEo50aXub dsNsXAZx3UIX2CfDBs1jXsPc/B37csOr9sJIqDn1xJKaT/Q6Yk1A/LV8uTa38wGh MykmdfDqWLT0nK4atbVH5fxjkTNdDg== =Fysh -END PGP SIGNATURE-
Bug#1067561: FTBFS: Error: symbol `open64' is already defined
On 14/04/2024 20:21, Yves-Alexis Perez wrote: On Sat, 2024-04-13 at 16:11 +0100, Peter Green wrote: >> Hi, thanks for the patch. It looks a bit strong though, undefining stuff >> like >> that unconditionally. Do you have pointers to the Ubuntu bug or something? >> I've looked at upstream commits and issues and couldn't see anything >> there. > My understanding of the issue. Hi Peter, thanks for the details. They really should be transmitted to the upstream maintainer so I took the liberty to copy/paste it to the upstream bug at https://github.com/canonical/lightdm/issues/352 Thanks, upstream has now accepted a patch that takes a slightly different approach to fixing the issue. https://github.com/canonical/lightdm/issues/352 Could we get this uploaded to sid, so that the lightweight desktops are installable on armel/armhf again?
Bug#1067561: FTBFS: Error: symbol `open64' is already defined
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 On Sat, 2024-04-13 at 16:11 +0100, Peter Green wrote: > > Hi, thanks for the patch. It looks a bit strong though, undefining stuff > > like > > that unconditionally. Do you have pointers to the Ubuntu bug or something? > > I've looked at upstream commits and issues and couldn't see anything > > there. > > My understanding of the issue. Hi Peter, thanks for the details. They really should be transmitted to the upstream maintainer so I took the liberty to copy/paste it to the upstream bug at https://github.com/canonical/lightdm/issues/352 Regards, - -- Yves-Alexis -BEGIN PGP SIGNATURE- iQEzBAEBCAAdFiEE8vi34Qgfo83x35gF3rYcyPpXRFsFAmYcLKoACgkQ3rYcyPpX RFvszggAuBC/jVutExKct2P2GNDoxIG8UHEkD0q0nZwzY3ojcVYF4Za4+g5i9HyJ khWQgqXzwxA5aWvOn4/lxo0CUd3LyzvKf4c1bP5wEX6LO3WZmOOAwDMQSdGfF5/k jZ2tDP212sS7WRhc8cUqTagKww90WJ8O9MfhusOSzhjimArxeM5XkO2CTmKZI7/0 eU+VTahJxaqyOp8LNNVrqyJHgdkr1LdWL67crixf4sPrkIKfcibYleoaiG0pK/Oq hcm7iV956GEFTa0VIYhuskTjB2TopA85UsGNTLYMGWtS7+XZbLirpuE8ToUqVeQ2 lm5douVLT78wjf4ZBbUq58wn499Q4g== =AAyB -END PGP SIGNATURE-
Bug#1067561: FTBFS: Error: symbol `open64' is already defined
Hi, thanks for the patch. It looks a bit strong though, undefining stuff like that unconditionally. Do you have pointers to the Ubuntu bug or something? I've looked at upstream commits and issues and couldn't see anything there. My understanding of the issue. In glibc _FILE_OFFSET_BITS=64 is used to substitute the standard file handling types and functions with versions that use 64-bit file offsets. Similarly _TIME_BITS=64 is used to susbstitute time handling types and functions with versions that use 64-bit seconds counts. All of this only applies to 32-bit architectures, on 64-bit architectures the default types in glibc are 64-bit. The "time64" transition was implemented by defining _TIME_BITS=64 and _FILE_OFFSET_BITS=64 by default in both dpkg-buildflags and the compiler. That's fine for normal code, but tests/src/libsystem.c is not normal code, it's a file of "test mocks" that override functions in glibc for testing purposes. If _FILE_OFFSET_BITS=64 is defined, it ends up trying to override "open64" twice, rather than trying to override both "open" and "open64" If we undef _FILE_OFFSET_BITS then we must also undef _TIME_BITS, otherwise the glibc headers will refuse to compile. So the patch seems reasonable to me, and since this is only test code, it appears to be pretty low risk. We would appreciate having this fix in unstable so the time_t transition can move forward.
Bug#1067561: FTBFS: Error: symbol `open64' is already defined
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 On Fri, 2024-04-05 at 11:25 +0500, Andrey Rakhmatullin wrote: > > Hi, thanks for the patch. It looks a bit strong though, undefining stuff > > like > > that unconditionally. Do you have pointers to the Ubuntu bug or something? > I haven't checked their bugs manually but the changelog entry doesn't > close any: > > lightdm (1.30.0-0ubuntu12) noble; urgency=medium > > * Undefine _FILE_OFFSET_BITS and _TIME_BITS when building preload gadget > to > allow interposition of open and open64 to work. > > -- Michael Hudson-Doyle Fri, 22 Mar 2024 > 15:08:46 +1300 Thanks, I've forwarded upstream, hoping to get their insight on this. Regards, - -- Yves-Alexis -BEGIN PGP SIGNATURE- iQEzBAEBCAAdFiEE8vi34Qgfo83x35gF3rYcyPpXRFsFAmYPobQACgkQ3rYcyPpX RFuS/wf9GL2qVFHO2p9o7RusyuTFQYUxUQL/GsGCRXirSGPAV7p3n84NgZuLP7s0 exqt/MLArwQcKKejkKHWZE/BppUX72UUdCzIKM31PSZTsmfNvaZyomfO4NIYTzrM UBTCGdp/xA0nIjzHNjAyABLFsCunxVCodGIF3w2fnvcc6t0/RZKh8clsAUMgKKGx yfRhyuW+YyapuWAx1gXLOn1uL8C2K/MFnyGYilvv/8Fe10tt+HsOmRQygXKDOzHG NVHpcIL8beIR4yeLBzCu2/U255955I2RGTIcNP1mLkPfzGt9IiVbP9HdsXt/ZXTD OeePyswwLEdntKNJcjnn20MPpKPTHA== =s7QG -END PGP SIGNATURE-
Bug#1067561: FTBFS: Error: symbol `open64' is already defined
On Fri, Apr 05, 2024 at 08:09:21AM +0200, Yves-Alexis Perez wrote: > > I assume the following patch from Ubuntu fixes this: > > > > --- a/tests/src/libsystem.c > > +++ b/tests/src/libsystem.c > > @@ -1,6 +1,9 @@ > > #define _GNU_SOURCE > > #define __USE_GNU > > > > +#undef _FILE_OFFSET_BITS > > +#undef _TIME_BITS > > + > > #include > > > > #include > > Hi, thanks for the patch. It looks a bit strong though, undefining stuff like > that unconditionally. Do you have pointers to the Ubuntu bug or something? I haven't checked their bugs manually but the changelog entry doesn't close any: lightdm (1.30.0-0ubuntu12) noble; urgency=medium * Undefine _FILE_OFFSET_BITS and _TIME_BITS when building preload gadget to allow interposition of open and open64 to work. -- Michael Hudson-Doyle Fri, 22 Mar 2024 15:08:46 +1300 -- WBR, wRAR signature.asc Description: PGP signature
Bug#1067561: FTBFS: Error: symbol `open64' is already defined
-BEGIN PGP SIGNED MESSAGE- Hash: SHA256 On Mon, 2024-04-01 at 00:37 +0500, Andrey Rakhmatullin wrote: > I assume the following patch from Ubuntu fixes this: > > --- a/tests/src/libsystem.c > +++ b/tests/src/libsystem.c > @@ -1,6 +1,9 @@ > #define _GNU_SOURCE > #define __USE_GNU > > +#undef _FILE_OFFSET_BITS > +#undef _TIME_BITS > + > #include > > #include Hi, thanks for the patch. It looks a bit strong though, undefining stuff like that unconditionally. Do you have pointers to the Ubuntu bug or something? I've looked at upstream commits and issues and couldn't see anything there. Regards, - -- Yves-Alexis -BEGIN PGP SIGNATURE- iQEzBAEBCAAdFiEE8vi34Qgfo83x35gF3rYcyPpXRFsFAmYPlZEACgkQ3rYcyPpX RFtkIwf/QZUtuSvH9DPGorTFJ/RybRfdEYONU8kLD8dOFGTXiNRcQYmSCTU/fH5B 4h0VxRWFXlF8TfBN0LmvgM+Rqzx9bT+5ZR+h8SGXL4UvH/W1qPeKIEOYYCmTtFMo nI9TBJhc501maa4PolJXSJ/OYO4AAQPmKvo+kAq0X7eTCaZcN5q1AleRDY5dBmru lHoQXy+L7T2+5N+L8wo7I7rAjLbrTEKH6CkYjNNTZx23Tp/Q6bLRELUg0f4XUusb xfAjR53aav/jCXJ8pYKKYWp60N2uRGPQcw05npjM314YlZlDXPyGn3xHQuEmQfkZ jRYE5mAmrHbtOplP9P49XX4W6cjZBQ== =9cvf -END PGP SIGNATURE-
Bug#1067561: FTBFS: Error: symbol `open64' is already defined
On Sat, Mar 23, 2024 at 09:56:58PM +0500, Andrey Rakhmatullin wrote: > /bin/bash ../../libtool --tag=CC --mode=link gcc -Wall > -Wstrict-prototypes -Wnested-externs > -Werror=missing-prototypes -Werror=implicit-function- > declaration -Werror=pointer-arith > -Werror=init-self -Werror=format-security > -Werror=format=2 -Werror=missing-include-dirs > -I/usr/include/glib-2.0 -I/usr/lib/arm-linux-gnueabi/glib-2.0/include > -I/usr/include/glib-2.0 -I/usr/lib/arm-linux-gnueabi/glib-2.0/include > -I/usr/include/glib-2.0 -I/usr/lib/arm-linux-gnueabi/glib-2.0/include -pthread > -I/usr/include/libmount -I/usr/include/blkid -I/usr/include/gio-unix-2.0 > -I/usr/include/glib-2.0 -I/usr/lib/arm-linux-gnueabi/glib-2.0/include -pthread > -I/usr/include/libmount -I/usr/include/blkid -g -O2 > -Werror=implicit-function- > declaration -ffile-prefix-map=/<>=. -fstack-protector-strong > -fstack-clash-protection -Wformat -Werror=format-security -O0 -Wl,-z,relro > -Wl,-z,now -Wl,-O1 -o Xvnc Xvnc-Xvnc.o Xvnc-x-authority.o Xvnc-x-common.o > Xvnc- > x-server.o Xvnc-status.o -lgobject-2.0 -lglib-2.0 -lglib-2.0 -lgio-2.0 > -lgobject-2.0 -lglib-2.0 -lgio-2.0 -lgobject-2.0 -lglib-2.0 > /tmp/ccCHYR2t.s: Assembler messages: > /tmp/ccCHYR2t.s:2779: Error: symbol `open64' is already defined > /tmp/ccCHYR2t.s:3181: Error: symbol `creat64' is already defined > /tmp/ccCHYR2t.s:3508: Error: symbol `__stat64_time64' is already defined I assume the following patch from Ubuntu fixes this: --- a/tests/src/libsystem.c +++ b/tests/src/libsystem.c @@ -1,6 +1,9 @@ #define _GNU_SOURCE #define __USE_GNU +#undef _FILE_OFFSET_BITS +#undef _TIME_BITS + #include #include -- WBR, wRAR signature.asc Description: PGP signature
Bug#1067561: FTBFS: Error: symbol `open64' is already defined
Source: lightdm Version: 1.32.0-4 Severity: serious Tags: ftbfs https://buildd.debian.org/status/fetch.php?pkg=lightdm=armel=1.32.0-4%2Bb3=1711185127=0 /bin/bash ../../libtool --tag=CC --mode=link gcc -Wall -Wstrict-prototypes -Wnested-externs -Werror=missing-prototypes -Werror=implicit-function- declaration -Werror=pointer-arith -Werror=init-self -Werror=format-security -Werror=format=2 -Werror=missing-include-dirs -I/usr/include/glib-2.0 -I/usr/lib/arm-linux-gnueabi/glib-2.0/include -I/usr/include/glib-2.0 -I/usr/lib/arm-linux-gnueabi/glib-2.0/include -I/usr/include/glib-2.0 -I/usr/lib/arm-linux-gnueabi/glib-2.0/include -pthread -I/usr/include/libmount -I/usr/include/blkid -I/usr/include/gio-unix-2.0 -I/usr/include/glib-2.0 -I/usr/lib/arm-linux-gnueabi/glib-2.0/include -pthread -I/usr/include/libmount -I/usr/include/blkid -g -O2 -Werror=implicit-function- declaration -ffile-prefix-map=/<>=. -fstack-protector-strong -fstack-clash-protection -Wformat -Werror=format-security -O0 -Wl,-z,relro -Wl,-z,now -Wl,-O1 -o Xvnc Xvnc-Xvnc.o Xvnc-x-authority.o Xvnc-x-common.o Xvnc- x-server.o Xvnc-status.o -lgobject-2.0 -lglib-2.0 -lglib-2.0 -lgio-2.0 -lgobject-2.0 -lglib-2.0 -lgio-2.0 -lgobject-2.0 -lglib-2.0 /tmp/ccCHYR2t.s: Assembler messages: /tmp/ccCHYR2t.s:2779: Error: symbol `open64' is already defined /tmp/ccCHYR2t.s:3181: Error: symbol `creat64' is already defined /tmp/ccCHYR2t.s:3508: Error: symbol `__stat64_time64' is already defined -- System Information: Debian Release: trixie/sid APT prefers unstable-debug APT policy: (500, 'unstable-debug'), (500, 'testing-debug'), (500, 'unstable'), (500, 'testing'), (101, 'experimental') Architecture: amd64 (x86_64) Foreign Architectures: i386 Kernel: Linux 6.7.9-amd64 (SMP w/4 CPU threads; PREEMPT) Kernel taint flags: TAINT_PROPRIETARY_MODULE, TAINT_OOT_MODULE, TAINT_UNSIGNED_MODULE Locale: LANG=ru_RU.UTF-8, LC_CTYPE=ru_RU.UTF-8 (charmap=UTF-8), LANGUAGE not set Shell: /bin/sh linked to /usr/bin/dash Init: systemd (via /run/systemd/system) LSM: AppArmor: enabled
