http://blog.gmane.org/gmane.comp.t2.devel hourly 1 1901-01-01T00:00+00:00 http://gmane.org/img/gmane-25t.png http://gmane.org http://permalink.gmane.org/gmane.comp.t2.devel/5247 <pre>My build host is the latest 8.0-stable image i downloaded from the T2 binary download, all packages installed. After failing to build our custom target(Atronix), I tried the Generic Minimal target. I had to Emerge-pkg python to get past the glib in stage 0. Now I have issues with e2fsprogs in stage 1, and I can't figure out what the error actually is...it seem to build and install, the saying due to previous errors, log file was not created. I have attached the ERROR-LOG from the src directory and the log file from var/adm/logs on the build tree. I hope that someone will be able to answer this as i have noticed very little activity on this list lately and start wondering if t2 SDE is now abandon ware and no longer actively used by anyone. If anyone have built t2-trunk in it's current configuration on a t2-8.0 system and know whats missing/fails, i would appreciate any help to resolve this. Mikael </pre> Mikael Ostensson 2012-05-22T19:50:38 http://permalink.gmane.org/gmane.comp.t2.devel/5246 <pre>After reading about about release .26 of pkgconfig is seem to conflict with alot of build anvironments. Have this been successfully used with 9.0-trunk? I have restarted my build after downgrading to release .25...will see how it goes. Thanks Mikael On Thu, May 17, 2012 at 12:25 PM, Mikael Ostensson &lt;mikaelo&lt; at &gt;gmail.com&gt; wrote: </pre> Mikael Ostensson 2012-05-17T17:58:00 http://permalink.gmane.org/gmane.comp.t2.devel/5245 <pre>Ok found out what the problem was: pkgconfig had not been built. However that package in turn had been commented out due to failure to find glib.h etc...in stage 0 After creating a few symlinks I passed stage 0, but now it fails in stage 1 with: !&gt; configure: error: pkg-config and glib-2.0 not found, please set GLIB_CFLAGS and GLIB_LIBS to the correct values. Full log attached. On Thu, May 17, 2012 at 11:12 AM, Mikael Ostensson &lt;mikaelo&lt; at &gt;gmail.com&gt; wrote: </pre> Mikael Ostensson 2012-05-17T16:25:17 http://permalink.gmane.org/gmane.comp.t2.devel/5244 <pre>BTW the file was generated by autoconf 2.68 I have attached the generated Configure file. On Thu, May 17, 2012 at 10:05 AM, Mikael Ostensson &lt;mikaelo&lt; at &gt;gmail.com&gt; wrote: </pre> Mikael Ostensson 2012-05-17T15:12:39 http://permalink.gmane.org/gmane.comp.t2.devel/5243 <pre>forgot to attach log. On Thu, May 17, 2012 at 10:04 AM, Mikael Ostensson &lt;mikaelo&lt; at &gt;gmail.com&gt; wrote: </pre> Mikael Ostensson 2012-05-17T14:05:33 http://permalink.gmane.org/gmane.comp.t2.devel/5242 <pre> I just started to attempt to build with a custom target (Rene built initial version for us) Host system: Ubuntu 12.04 Package: base/util-linux [2.18 9.0-trunk] Attached Error.log Message during build: checking whether to build static libraries... yes ./configure: line 11650: PKG_PROG_PKG_CONFIG: command not found ./configure: line 11673: syntax error near unexpected token `gtk-doc' ./configure: line 11673: ` PKG_CHECK_EXISTS(gtk-doc &gt;= 1.10,,' make: *** [config.status] Error 2 Due to previous errors, no 1-util-linux.log file! (Try enabling xtrace in the config to track an error inside the build system.) Am I missing something on the host system? This error was not straight forward to me to resolve, and I need some help in this. Thanks Mikael </pre> Mikael Ostensson 2012-05-17T14:04:44 http://permalink.gmane.org/gmane.comp.t2.devel/5241 <pre> Gerardo Di Iorio wrote: 404... abandoned prematurely? I saw that something exists in T2 trunk, has anyone already tried? I'm planning to bring a project to ARM, beaglebone and raspberrypi will probably be the targets on which the first efforts will focus. &lt; at &gt;gerardo: accidentally sent that message direct to you, sorry for that. </pre> giovanni.v 2012-04-23T08:00:11 http://permalink.gmane.org/gmane.comp.t2.devel/5240 <pre> hello everybody. finally with bit of effort and lots' of machine cycles i'm able to cross build mips distro from t2 trunk! analyzer tells it takes 30.05min to build cross/toolchain in stage-0 and roughly 3.6hrs to build the stage-1 (gnome2 base) in a clean build. now i'd like to try more brave tricks. is there a easy (crazy) way to build either(or both RPMs and Debian packages out of these tar.bz2's? next obvious thing to-do i believe will be to wire a comprehensive installer (anaconda?) to complete the job. any help, tips? indeed, its great job folks!!! thanks. girish.</pre> girish gulawani 2012-04-22T17:31:12 http://permalink.gmane.org/gmane.comp.t2.devel/5239 <pre>Hi, i try to build an native x64 config. I have error on build package disktype and all package cmpilated with dietlibc the error is: /root/t2/build/xfce-9.0-trunk-generic-x86-64-/usr/dietlibc/lib-x86_64/libc.a(vsn printf.o): In function `vsnprintf': vsnprintf.c:(.text+0xc4): warning: warning: the printf functions add several kil obytes of bloat. /root/t2/build/xfce-9.0-trunk-generic-x86-64-/usr/dietlibc/lib-x86_64/libc.a(std err.o): In function `__fflush_stderr': stderr.c:(.text+0x8): warning: warning: your code uses stdio (7+k bloat). /root/t2/build/xfce-9.0-trunk-generic-x86-64-/TOOLCHAIN/tools.cross/crosscc/../l ib/gcc/x86_64-t2-linux-gnu/4.7.0/../../../../x86_64-t2-linux-gnu/bin/ld: error i n /root/t2/build/xfce-9.0-trunk-generic-x86-64-/usr/dietlibc/lib-x86_64/dyn_stop .o(.eh_frame); no .eh_frame_hdr table will be created. lib.o: In function `format_blocky_size': lib.c:(.text+0x40b): undefined reference to `stpcpy' collect2: error: ld returned 1 exit status make: *** [disktype] Error 1 Due to previous errors, no 1-disktype.log file! regards Gerardo Di Iorio </pre> Gerardo Di Iorio 2012-04-11T21:47:27 http://permalink.gmane.org/gmane.comp.t2.devel/5238 <pre>Hey folks, for newcomers, and people who like to see new sources, educate and learn, I have a Janitor task. As usually updating to a new, stricter, compiler, the update to GCC-4.7 (http://t2-project.org/packages/gcc.html) brought a number of new build errors. Obviously, as expected, ... One thing, they constantly clean header, includes, not only for standard conformance, but also speedup builds. So some very easy things to fix, are missing includes and thus undeclared identifiers: grep "error: .*not declared" build/reference-9.0-trunk-reference-x86-i686-linux/var/adm/logs/*.err Of course there are many more, usually C++, errors due more standard compliance. PS: If you have no time, or basic C experience yourself, maybe spread the word to pupils, students near you! Happy easter and patches welcome! René </pre> René Rebe 2012-04-04T15:26:55 http://permalink.gmane.org/gmane.comp.t2.devel/5237 <pre> Hello Jan. Took me a while to get back to this project.   Glad to report that these hints did the trick! The mips64 target now builds proper without Dietlib. Also i had to change the pkgsel to remove mandatory dietlib inclusion (generic/pkgsel/00-minimal.in). This i suppose can be moved on to either target/architecture specific pkgsel. Thanks a bunch! Best. Girish. </pre> girish gulawani 2012-03-25T05:27:39 http://permalink.gmane.org/gmane.comp.t2.devel/5236 <pre> I have made quite alot of progress on sparc32 T2 Linux. Openssl needs sparc v7 or v8 to have -m32 not -mv8 in the sparc patch as it isn't recognised by the compiler -m32 should be correct although probably redundant. Also the machine_arch for the toolchain should be sparc not sparcv8 as that causes gcc to fail. Optimization for sparcv8 should still work so I don't see any issue with that. So the tuple ends up being sparc-t2-linux-gnu instead of sparcv8-t2-linux-gnu which is a problem I rand into 3 years ago I believe I was rather glad to finally figure it out. I'm currently working on apr-utils. I had to update the apr-utils cross-bdb.patch for updated bdb versions as it was attempting to -ldb-4.4 didn't work when -ldb-4.8 should have been selected.. I I just extended that patch accordingly and that part seemed ok. I'll send it in sometime this weekend. Now it fails to find /usr/share/build-1/libtool I believe it may not be getting redirected to the chroot as my system doesn't have that file in that directory. It is in the build directly though so I may have to jump on IRC late one night when I can get in contact to try and figure out how to fix it. -Chase Rayfield </pre> Chase Rayfield 2012-03-24T16:27:18 http://permalink.gmane.org/gmane.comp.t2.devel/5235 <pre>Hi all, On 14.03.2012, at 22:20, René Rebe wrote: I even "forgot" we already removed the "very old" linux24 packages :-) So now all this linux26 stuff should have been renamed to just "linux": Committed revision 40366. Committed revision 40367. Enjoy! Some minor missing bits and pieces are to be expected, patches welcome! René </pre> René Rebe 2012-03-20T11:03:18 http://permalink.gmane.org/gmane.comp.t2.devel/5234 <pre> At the bottom is my first attempt at a patch. It adds all the syscalls that were not already alocated and updates the ones required to get the build to continue past udev linking and embutils.I refereed to the Linux-3.3 source for the updated syscalls. I think it would probably be better to go ahead and change *all* the syscalls so that they are in sync with Linux 3.3 but I would like to get some input from the list on that. I am now to this point in compiling sparc32 on t2-9.0 looks like libffi hasn't been built yet how would I go about doing that? I checked and T2 definitly has libffi &gt; 3.0 so it just hasn't been built yet. checking for LIBFFI... no configure: error: Package requirements (libffi &gt;= 3.0.0) were not met: No package 'libffi' found Consider adjusting the PKG_CONFIG_PATH environment variable if you installed software in a non-standard prefix. Alternatively, you may set the environment variables LIBFFI_CFLAGS and LIBFFI_LIBS to avoid the need to call pkg-config. See the pkg-config man page for more details. Due to previous errors, no 1-glib.log file! (Try enabling xtrace in the config to track an error inside the build system.) --- BUILD ERROR --- Creating file list and doing final adaptions ... Searching for orphaned files ... Found 15 files for this package. Found 0 orphaned files for this package. Clear (old) md5sums ... Creating md5sum files ... done. Creating package description ... Making post-install adaptions. == 03/19/12 13:37:05 =[1]=&gt; Aborted building package glib. -&gt; Unmounting loop mounts ... sparc32_syscalls-1.patch diff -rupN dietlibc-0.32/sparc/syscalls.h dietlibc-0.32-new//sparc/syscalls.h --- dietlibc-0.32/sparc/syscalls.h 2008-09-30 16:02:35.000000000 -0500 +++ dietlibc-0.32-new//sparc/syscalls.h 2012-03-19 12:04:29.000000000 -0500 &lt; at &gt;&lt; at &gt; -106,11 +106,11 &lt; at &gt;&lt; at &gt; #define __NR_rt_sigtimedwait 105 /* Linux Specific */ #define __NR_rt_sigqueueinfo 106 /* Linux Specific */ #define __NR_rt_sigsuspend 107 /* Linux Specific */ -#define __NR_setresuid32 108 /* Linux Specific, sigvec under SunOS */ -#define __NR_getresuid32 109 /* Linux Specific, sigblock under SunOS */ -#define __NR_setresgid32 110 /* Linux Specific, sigsetmask under SunOS */ -#define __NR_getresgid32 111 /* Linux Specific, sigpause under SunOS */ -#define __NR_setregid32 112 /* Linux sparc32, sigstack under SunOS */ +#define __NR_setresuid 108 /* Linux Specific, sigvec under SunOS */ +#define __NR_getresuid 109 /* Linux Specific, sigblock under SunOS */ +#define __NR_setresgid 110 /* Linux Specific, sigsetmask under SunOS */ +#define __NR_getresgid 111 /* Linux Specific, sigpause under SunOS */ +#define __NR_setregid 112 /* Linux sparc32, sigstack under SunOS */ #define __NR_recvmsg 113 /* Common */ #define __NR_sendmsg 114 /* Common */ #define __NR_getgroups32 115 /* Linux sparc32, vtrace under SunOS */ &lt; at &gt;&lt; at &gt; -149,12 +149,12 &lt; at &gt;&lt; at &gt; #define __NR_pciconfig_read 148 /* ENOSYS under SunOS */ #define __NR_pciconfig_write 149 /* ENOSYS under SunOS */ #define __NR_getsockname 150 /* Common */ -/* #define __NR_getmsg 151 SunOS Specific */ -/* #define __NR_putmsg 152 SunOS Specific */ +#define __NR_inotify_init 151 /* Linux specific */ +#define __NR_inotify_add_watch 152 /* Linux specific */ #define __NR_poll 153 /* Common */ #define __NR_getdents64 154 /* Linux specific */ #define __NR_fcntl64 155 /* Linux sparc32 Specific */ -/* #define __NR_getdirentires 156 SunOS Specific */ +#define __NR_inotify_rm_watch 156 /* */ #define __NR_statfs 157 /* Common */ #define __NR_fstatfs 158 /* Common */ #define __NR_umount 159 /* Common */ &lt; at &gt;&lt; at &gt; -315,6 +315,29 &lt; at &gt;&lt; at &gt; #define __NR_fallocate 314 #define __NR_timerfd_settime 315 #define __NR_timerfd_gettime 316 +#define __NR_signalfd4 317 +#define __NR_eventfd2 318 +#define __NR_poll_create1 319 +#define __NR_dup3 320 +#define __NR_pipe2 321 +#define __NR_inotify_init1 322 +#define __NR_accept4 323 +#define __NR_preadv 324 +#define __NR_pwritev 325 +#define __NR_rt_tgsigqueueinfo 326 +#define __NR_perf_event_open 327 +#define __NR_recvmmsg 328 +#define __NR_fanotify_init 329 +#define __NR_fanotify_mark 330 +#define __NR_prlimit64 331 +#define __NR_name_to_handle_at 332 +#define __NR_open_by_handle_at 333 +#define __NR_clock_adjtime 334 +#define __NR_syncfs 335 +#define __NR_sendmmsg 336 +#define __NR_setns 337 +#define __NR_process_vm_readv 338 +#define __NR_process_vm_writev 339 #define syscall_weak(name,wsym,sym) \ .text; \ </pre> chase rayfield 2012-03-20T01:57:02 http://permalink.gmane.org/gmane.comp.t2.devel/5233 <pre>Hey, Sorry: Committed revision 40362. Committed revision 40363. No idea why it built initially for me, ... René On 14.03.2012, at 11:47, René Rebe wrote: </pre> René Rebe 2012-03-19T15:51:38 http://permalink.gmane.org/gmane.comp.t2.devel/5232 <pre>So I had a go at building Sparc32 again with T2 the build fails at udev in stage 1. From what I understand the header that defines this since we are using dietlibc right? /build/sparc-9.0-trunk-generic-sparc-v7-linux/usr/dietlibc/include/sys/inotify.h I also tried declaring the functions extern but that didn't work and failed with the same error. From what I understand the kernel is supposed to provide that but it isn't built yet but the headers should be there.. so. Any ideas? Thanks, Chase Rayfield  CCLD   udev/udevd /root/haiku/t2-trunk/build/sparc-9.0-trunk-generic-sparc-v7-linux/usr/dietlibc/lib-sparc/libc.a(sprintf.o): In function `sprintf': sprintf.c:(.text+0x20): warning: warning: Avoid *sprintf; use *snprintf. It is more secure. libudev/.libs/libudev-private.a(libudev-queue-private.o): In function `rebuild_queue_file': libudev-queue-private.c:(.text+0x434): warning: warning: your code uses stdio (7+k bloat). /root/haiku/t2-trunk/build/sparc-9.0-trunk-generic-sparc-v7-linux/usr/dietlibc/lib-sparc/libc.a(vsscanf.o): In function `vsscanf': vsscanf.c:(.text+0x88): warning: warning: the scanf functions add several kilobytes of bloat. /root/haiku/t2-trunk/build/sparc-9.0-trunk-generic-sparc-v7-linux/usr/dietlibc/lib-sparc/libc.a(vsnprintf.o): In function `vsnprintf': vsnprintf.c:(.text+0xb8): warning: warning: the printf functions add several kilobytes of bloat. udev/udev-watch.o: In function `udev_watch_begin': udev-watch.c:(.text+0xb0): undefined reference to `inotify_add_watch' udev-watch.c:(.text+0x16c): undefined reference to `inotify_add_watch' udev/udev-watch.o: In function `udev_watch_end': udev-watch.c:(.text+0x4a4): undefined reference to `inotify_rm_watch' udev/udevd.o: In function `main': udevd.c:(.text.startup+0xb2c): undefined reference to `inotify_add_watch' udevd.c:(.text.startup+0x1c0c): undefined reference to `inotify_add_watch' udevd.c:(.text.startup+0x1c20): undefined reference to `inotify_add_watch' udevd.c:(.text.startup+0x1c8c): undefined reference to `inotify_add_watch' /root/haiku/t2-trunk/build/sparc-9.0-trunk-generic-sparc-v7-linux/usr/dietlibc/lib-sparc/libc.a(inotify_init1.o): In function `inotify_init1': (.text+0x4): undefined reference to `__NR_inotify_init1' /root/haiku/t2-trunk/build/sparc-9.0-trunk-generic-sparc-v7-linux/usr/dietlibc/lib-sparc/libc.a(inotify_init1.o): In function `inotify_init1': (.text+0x4): relocation truncated to fit: R_SPARC_13 against undefined symbol `__NR_inotify_init1' collect2: ld returned 1 exit status make[2]: *** [udev/udevd] Error 1 make[1]: *** [all-recursive] Error 1 make: *** [all] Error 2 Due to previous errors, no 1-udev.log file! (Try enabling xtrace in the config to track an error inside the build system.) --- BUILD ERROR --- Creating file list and doing final adaptions ... removed `/root/haiku/t2-trunk/build/sparc-9.0-trunk-generic-sparc-v7-linux/lib/libudev.la' Searching for orphaned files ... Found 60 files for this package. Found 11 orphaned files for this package. Clear (old) md5sums ... Creating md5sum files ... done. Creating package description ... Making post-install adaptions.</pre> chase rayfield 2012-03-16T17:40:54 http://permalink.gmane.org/gmane.comp.t2.devel/5231 <pre> On 3/15/2012 12:37 AM, girish gulawani wrote: I did not find a clean way to do it, this method is rather klugey, but it worked for experimental purposes: First I tried removing the DIETLIBC specifications in the package .desc files, they are in the [F] section. but that did not seem to work, perhaps I did not totally clean &amp; reconfigure after that. but what did work was going into the config dir for my target and editing the config file. I changed all SDECFG_DIETLIBC_&lt;package-name&gt;=1 to =0. </pre> Jan Rovins 2012-03-15T16:11:23 http://permalink.gmane.org/gmane.comp.t2.devel/5230 <pre>   Hello René. Thanks for the following up! OK. Let me try that out and see if I get through. I'm somehow skeptical it won't work without hacks :-( Yes. The newlib is being used for the bare metal kind platforms. It is mips64, to be precise Cavium's Octeon series processor, based platform. Best regards, Girish. </pre> girish gulawani 2012-03-15T08:48:56 http://permalink.gmane.org/gmane.comp.t2.devel/5229 <pre>Hey all, On 15.03.2012, at 05:37, girish gulawani wrote: Mips64 would be more tested, if I had a mips64 machine that is fun, my Octane is just too loud to keep running in the office (or at home), ... But I do not need any silicon, sending patches along is generally nice (tm) was well :-) There are even architectures without dietlibc support, superh, and whatnot, ... you only need to touch the initrd stuff al ittle, as it relies on statically compiled binaries for the initrd, ... Wasn't newlib for freestanding stuff without OS kernel? What architecture are you targeting? Greetings, </pre> René Rebe 2012-03-15T08:41:37 http://permalink.gmane.org/gmane.comp.t2.devel/5228 <pre> Hello Jan. Thanks very much for the detailed information on every point!!!      undo this for most of them, but the embedded-utils needed for the      initrd  was impossible to split from dietlib without a major     re-write. If you don't mind can you share some ways' to skip Dietlib? Because in my case the initrd is not exactly a problem (for the busybox is working), so the embed-utils part can be skipped over. I'm trying to check the Dietlib port. Hope i don't have to give up midway and run behind making it full-fat Glibc/Newlib :( Wish me luck! Thanks again. Best regards, Girish. </pre> girish gulawani 2012-03-15T04:37:58 http://permalink.gmane.org/gmane.comp.t2.devel/5227 <pre>Hi Girish, I am using T2 on mips64, but it has been many months since I have built anything. I was using a fork of the 8.0 built for just the n32 abi, the multilib on mips64 was not quite working at the time. I tried a few experimental builds with 9.0 trunk a while back, some worked better than others, since the GCC &amp; Glibc versions kept changing, and mips64 seems to be always broken on the bleeding edge. If the latest trunk does not work on mips64, try going back one or 2 revisions to get something that works. No, but I have not built anything recently. Yes I remember running into this on 9.0. there was an older version of perl that built, but I was having trouble forward porting the patches from the older version. The new versions of perl would not build on mips64. I think that I left my build on the older version that would build on mips. I was working with a generic minimal build, so those packages were never tried. Dietlib was definitely not working on mips64 for 8.0, Rene did a lot of porting and it should be functional for 9.0 by now, but I have not been using it, and do not know how much it has been tested. Dietlib was hard coded into many of the projects, I found a way to undo this for most of them, but the embedded-utils needed for the initrd was impossible to split from dietlib without a major re-write. I do not have access to the T2 build machines that I was using back then, so I cant give you any specifics right now. I can get those machines back on line in a few days, and share any patches that I was using. Jan </pre> Jan Rovins 2012-03-15T04:08:16 Search the mailing list at Gmane query http://search.gmane.org/?group=$group=gmane.comp.t2.devel