| Mon, 15 Jun 2009 01:11:26 -0500 |
Rob Landley |
Teach ccwrap to handle --print-prog-name=ld and such.
|
| Mon, 15 Jun 2009 01:09:30 -0500 |
Rob Landley |
The powerpc stuff doesn't use this anymore, so delete it.
|
| Thu, 11 Jun 2009 19:19:34 -0500 |
Rob Landley |
Missed a script that needs to call read_arch_dir now.
|
| Thu, 11 Jun 2009 19:04:51 -0500 |
Rob Landley |
Change a comment to silence a warning.
|
| Thu, 11 Jun 2009 05:42:51 -0500 |
Rob Landley |
Refactor so include.sh mostly just sets environment variables. Move read_arch_dir to function and call it explicitly (no more need for $NO_ARCH). Make blank_tempdir a function, called explicitly by stages when needed, with some sanity checks. Insert prerequisite tests to later stages so they can detect failure early and provide an explicit erro rmessage, and have those tests happen before blanking $WORK dir, to preserve debugging info. Make buildall.sh depend on prerequisite tests rather than trying to avoid calling later stages (and thus do flow control from asynchronous context). Add FAIL_QUIET option so buildall.sh doesn't spam the log with the new prerequisite error messages.
|
| Thu, 11 Jun 2009 00:26:15 -0500 |
Rob Landley |
Commit 729 broke canadian cross to fix ppc440 build, so try again to hopefully fix _both_ issues.
|
| Sun, 31 May 2009 02:36:22 -0500 |
Rob Landley |
Add a 60 second timeout to smoketest. (Some qemu failure modes involve hanging.)
|
| Sun, 31 May 2009 02:35:41 -0500 |
Rob Landley |
We already added an /etc/resolv.conf with the qemu default value, so don't have init.sh overwrite it (which outputs an error message on read only filesystems).
|
| Sat, 30 May 2009 18:21:03 -0500 |
Rob Landley |
Quick and dirty script to run smoketest.sh on all targets, to see if they can compile and run "hello world".
|
| Fri, 29 May 2009 14:17:36 -0500 |
Rob Landley |
Move more stuff towards using "cc" instead of "gcc".
|
| Fri, 29 May 2009 14:14:13 -0500 |
Rob Landley |
For armv4eb, use the armv4l config. (Linux kernel .config doesn't specify endianness, and I didn't notice when they fell out of sync...)
|
| Fri, 29 May 2009 14:12:33 -0500 |
Rob Landley |
Use $ARCH-cc instead of $ARCH-gcc where possible. (pcc, llvm, tcc, qcc...)
|
| Fri, 29 May 2009 14:11:39 -0500 |
Rob Landley |
Build statically linked native compiler tarballs (extract 'em on any native system and build with them), and fix several unrelated bugs.
|
| Mon, 18 May 2009 16:40:09 -0500 |
Rob Landley |
Squashfs needs an /etc/resolv.conf, not just /usr/etc/resolv.conf
|
| Mon, 18 May 2009 15:08:18 -0500 |
Rob Landley |
Fix obvious typo.
|
| Mon, 18 May 2009 15:07:38 -0500 |
Rob Landley |
Patch no longer needed/applies with busybox 1.14.
|
| Mon, 18 May 2009 15:06:41 -0500 |
Rob Landley |
Root directory isn't writable with squashfs, use /tmp instead.
|
| Mon, 18 May 2009 15:05:40 -0500 |
Rob Landley |
Documentation tweak.
|
| Mon, 18 May 2009 15:05:02 -0500 |
Rob Landley |
Upgrade to busybox 1.14.0.
|
| Sun, 17 May 2009 15:14:46 -0500 |
Rob Landley |
Fix the canadian cross stuff for arches that override CROSS_TARGET (ala ppc440)
|
| Mon, 11 May 2009 01:06:39 -0500 |
Rob Landley |
Teach squashfs to create /dev/console and change default image time to squashfs.
|
| Mon, 11 May 2009 01:05:52 -0500 |
Rob Landley |
Cosmetic improvement, have === bars show stage name.
|
| Thu, 07 May 2009 19:09:56 -0500 |
Rob Landley |
Centrialize BUILD_STATIC and make it apply to more of root-filesystem.sh
|
| Tue, 28 Apr 2009 22:38:38 -0500 |
Rob Landley |
No longer need this, ./buildall.sh does it now.
|
| Tue, 28 Apr 2009 04:08:11 -0500 |
Rob Landley |
Make MANIFEST generation less brittle.
|
| Mon, 27 Apr 2009 04:30:52 -0500 |
Rob Landley |
Convert i686 target from virtual ne2k to virtual e1000 gigabit adapter.
|
| Sun, 26 Apr 2009 19:20:39 -0500 |
Rob Landley |
Fix thinkos.
|
| Sun, 26 Apr 2009 15:30:26 -0500 |
Rob Landley |
Do canadian cross static toolchains properly, via buildall.sh.
|
| Sun, 26 Apr 2009 14:19:21 -0500 |
Rob Landley |
Make "wait4background" with no arguments wait for all background tasks.
|
| Fri, 24 Apr 2009 22:33:31 -0500 |
Rob Landley |
Have system-image.sh do its own setsid variant so when it does dienow() it can kill all its own backgrounded processes, but won't take down the parent shell that ran it.
|
