Hi Juergen,
* Juergen Beisert j...@pengutronix.de [2012-03-20 20:28]:
Bernhard Walle wrote:
* Juergen Beisert j...@pengutronix.de [2012-03-20 10:09]:
Just delay the evaluation of the expression to the point of time when it
is really required.
diff --git a/rules/flashbench.make b
-by: Bernhard Walle bernh...@bwalle.de
---
rules/flashbench.make |2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/rules/flashbench.make b/rules/flashbench.make
index fcb669e..0ef52d1 100644
--- a/rules/flashbench.make
+++ b/rules/flashbench.make
@@ -31,7 +31,7 @@ FLASHBENCH_LICENSE
Hi Andreas,
* Andreas Bießmann biessm...@corscience.de [2012-03-20 09:37]:
this is annoying, it will require gnu binutils which is not available on
my mac by default.
Couldn't we use some approach like
http://thread.gmane.org/gmane.comp.embedded.ptxdist.devel/8533
(pachting ipkg to use
Usually there is one, but don't print an error if there is no one.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
bin/ptxdist |4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/bin/ptxdist b/bin/ptxdist
index 440a7aa..37ab098 100755
--- a/bin/ptxdist
+++ b/bin
.
This patch attempts to fix the problem.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
rules/post/function-targetinfo.make |3 +++
rules/post/function-touch.make |3 +++
2 files changed, 6 insertions(+)
diff --git a/rules/post/function-targetinfo.make
b/rules/post/function
From: Bernhard Walle wa...@corscience.de
This patch adds the flashbench tool from Linaro Project
git://git.linaro.org/people/arnd/flashbench.git. See also
http://lwn.net/Articles/428584/. Because there is no tarball, I
exported the current git repo and put it on my server.
The default maximum
From: Bernhard Walle wa...@corscience.de
This patch adds the flashbench tool from Linaro Project
git://git.linaro.org/people/arnd/flashbench.git. See also
http://lwn.net/Articles/428584/. Because there is no tarball, I
exported the current git repo and put it on my server.
The default maximum
* Michael Olbrich m.olbr...@pengutronix.de [2012-03-13 11:14]:
Not the 'local', the whole function. Getting all corner-cases right for
something like this is not trivial.
Ah, ok. Indeed. That's why I started using something from the
Internet[tm] and that's also why I only replaced that
-Makefile-Use-LZOLDFLAGS-when-building-mkfs.ubifs.patch
@@ -0,0 +1,24 @@
+From: Bernhard Walle bernh...@bwalle.de
+Date: Sun, 11 Mar 2012 18:38:20 +0100
+Subject: [PATCH] Makefile: Use LZOLDFLAGS when building mkfs.ubifs
+
+The LZOLDFLAGS variable is used by jffs2 but not by mkfs.ubifs
-by: Bernhard Walle bernh...@bwalle.de
---
.../0001-make-liblzo-optional.patch| 107
.../0002-make-mkubifs-optional.patch | 26 -
patches/mtd-utils-1.4.6/series |2 -
.../mtd-utils-1.4.9/0001-make-ubifs-optional.patch | 25
Some tools that mtd-utils provide can be also provided by busybox. Add
the dependencies here as done in other configuration files to prevent
the user from misconfiguration.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
rules/mtd-utils.in | 28
1 file changed
Hi Michael,
thanks again for the review.
* Michael Olbrich m.olbr...@pengutronix.de [2012-02-14 23:05]:
+# bash implementation of realpath / readlink -f
+# arg1 - filename
+realpath()
+{
+ local fname oldfname
+
hmmm, scary.
What's wrong about 'local'?
Another idea: Is the
We check for a xargs binary that supports the '-r' flag, so put the
usual symlink into the bin directory once we have found a proper
implementation.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
Makefile.in |1 +
1 file changed, 1 insertion(+)
diff --git a/Makefile.in b/Makefile.in
http://mailman.archlinux.org/pipermail/pacman-dev/2009-February/008130.html.
Tested on Mac OS 10.7.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
Makefile.in |1 +
bin/ptxdist | 22 +-
configure.ac |1 +
3 files changed, 23 insertions(+), 1 deletion
Am 24.02.12 18:20, schrieb Robert Palmer:
I've searched the archives (unfortunately patch is VERY common)
and I read the How to become a PTXdist Guru from the pengutronix
web site. In section 4.1.2 patch series, it gives an overview of
patching and it then list the search order, giving
from Apple is sufficient, but at least
/usr/lib/libncuses.dylib is there. Apart from building kconfig, ptxdist
uses its own host-ncurses, though.
Bernhard Walle started a generic arm bsp [10] (targeted on beagle-board
and clones). He uses ptxdist with OS X too, therefore this is a good
place to go
We need that because scripts/lib/ptxd_make_xpkg_pkg.sh uses the -m
parameter of mknod which doesn't exist in BSD mknod.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
Makefile.in |1 +
configure.ac |1 +
2 files changed, 2 insertions(+), 0 deletions(-)
diff --git a/Makefile.in b
for
GNU tar.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
Makefile.in |2 +-
configure.ac | 41 +++--
2 files changed, 20 insertions(+), 23 deletions(-)
diff --git a/Makefile.in b/Makefile.in
index 0c05294..b976a3f 100644
--- a/Makefile.in
+++ b
* Michael Olbrich m.olbr...@pengutronix.de [2012-02-14 09:11]:
On Tue, Feb 07, 2012 at 06:00:06PM +0100, bernh...@bwalle.de wrote:
From: Bernhard Walle wa...@corscience.de
I don't know why the name of the tarball is completely different and I
also don't know why the tarball doesn't
http://mailman.archlinux.org/pipermail/pacman-dev/2009-February/008130.html.
Tested on Mac OS 10.7.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
Makefile.in |1 +
bin/ptxdist | 25 -
configure.ac |1 +
3 files changed, 26 insertions(+), 1 deletions
is useful for systems without gettext in the default
installation like Darwin.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
include/libintl.h |8
1 files changed, 8 insertions(+), 0 deletions(-)
diff --git a/include/libintl.h b/include/libintl.h
index 096aa8c..a8d33dd 100644
an additional u-boot header) on a SD card.
Because of the file name conflict, U_BOOT_INSTALL_MLO is only selectable
when x-loader is not built.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
platforms/u-boot.in | 19 +++
rules/u-boot.make |6 ++
2 files changed, 25
The patches are obsolete as they are mainline. Verified on a NSLU2
(xscale-arm) little endian board/configuration.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
.../0001-revert-toolchain-problem.patch| 313
...uild-Pass-no-undefined-as-a-linker
* Michael Olbrich m.olbr...@pengutronix.de [2012-02-07 15:46]:
On Sun, Jan 22, 2012 at 12:29:48AM +0100, Bernhard Walle wrote:
This patch is by far the most ugly part of all of the Darwin patches.
At least the md5sum of a locale-archive when generating a de_DE.UTF-8
locale is identical
The build system of zlib uses 'uname' to determine the operating system
for which it compiles zlib. However, on non-Linux hosts, this leads to
wrong results.
Fixes Darwin (tested on Mac OS 10.7 Lion), doesn't break Linux (verified
with current Arch Linux on x86_64).
Signed-off-by: Bernhard Walle
-by: Bernhard Walle bernh...@bwalle.de
---
rules/other/Toplevel.make |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/rules/other/Toplevel.make b/rules/other/Toplevel.make
index 35c2c27..85ad515 100644
--- a/rules/other/Toplevel.make
+++ b/rules/other/Toplevel.make
@@ -39,7
Hi,
* Tim Sander ti...@vlsi.informatik.tu-darmstadt.de [2012-02-01 00:53]:
Hi
Does anybody by chance have a beaglebone ptxdist whe (s)he likes to share?
% hg clone https://bitbucket.org/bwalle/ptxdist-arm-boards -r 26030b06fd4a
However, the kernel configuration is not tested to have DVI
Hi Michael,
* Michael Olbrich m.olbr...@pengutronix.de [2012-01-30 15:06]:
This is not correct. It will sort all post files and does not honor the
order of PTXDIST_PATH_POSTRULES. This should probably be something like
this (untested):
$(foreach dir, $(call reverse,$(subst
-by: Bernhard Walle bernh...@bwalle.de
---
rules/other/Toplevel.make |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/rules/other/Toplevel.make b/rules/other/Toplevel.make
index 35c2c27..0cb7e82 100644
--- a/rules/other/Toplevel.make
+++ b/rules/other/Toplevel.make
@@ -95,7
Hi Michael,
Am 24.01.12 10:19, schrieb Michael Olbrich:
- Is sysctl alwyas in $PATH? On Linux it's in sbin.
I think so, but it's in /usr/sbin, so I can use /usr/sbin/syctl to be
more safe.
- the value should be quoted.
ok.
- Does this _always_ and _only_ work on Mac OS X?
Good point. I
Tested on Mac OS 10.7 (Lion).
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
v2: Use /usr/sbin/sysctl instead of relying on the $PATH and don't check
for the operating system but just try it out.
bin/ptxdist |4 +++-
1 files changed, 3 insertions(+), 1 deletions(-)
diff --git
This is the second version of the patches. The third patch is not needed any
more because host-elf-h-compat is now built as base package rather than as
dependency. All other remarks of the review have been integrated in the patch
series.
The package is maintained at
not on Linux (per uname -s output). This makes
maintaining the dependencies easier than requiring it in any packages
that need elf.h.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
rules/host-elf-h-compat.in |5 +
rules/host-elf-h-compat.make | 36
2
://article.gmane.org/gmane.linux.kbuild.devel/5210 for a
discussion about HOST_EXTRACFLAGS.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
rules/kernel.make |3 +--
1 files changed, 1 insertions(+), 2 deletions(-)
diff --git a/rules/kernel.make b/rules/kernel.make
index a445182..336b76d 100644
* Andreas Bießmann biessm...@corscience.de [2012-01-23 13:40]:
shouldn't we build two different packages (u-boot-tool-mkimage,
u-boot-tool-env) here?
I would add that tools as part of the u-boot package.
Mixing a host tool with target tools don't sound like a good idea.
Regards,
Bernhard
* Andreas Bießmann biessm...@corscience.de [2012-01-23 14:16]:
no, you got me wrong here. We have a host-u-boot-tools package which
have the mkimage build for the host (currently only that tool, but maybe
sometimes mkenvimage too?). Here we have the target package build for
the specified
* Andreas Bießmann andr...@biessmann.de [2012-01-23 22:29]:
On 22.01.12 00:29, Bernhard Walle wrote:
This patch is by far the most ugly part of all of the Darwin patches.
At least the md5sum of a locale-archive when generating a de_DE.UTF-8
locale is identical...
A first review, please
doesn't have any file type in vim).
However, with this setting vim automatically closes all folds, including
the first '{{{' which only is for Emacs. That's why I add fdl=1, which
results in vim opening the file visible one fold per ptxdist run.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
doesn't have any file type in vim).
However, with this setting vim automatically closes all folds, including
the first '{{{' which only is for Emacs. That's why I add fdl=1, which
results in vim opening the file visible one fold per ptxdist run.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
v2
This patch series is an attempt to solve the problem that Darwin has no
/usr/include/elf.h. I'm sure that other non-Linux but POSIX-compatible operating
systems have no /usr/include/elf.h, so it should not only increase portability
for Mac OS but also for *BSD.
I think there are various ways to
The kernel already depends on it, but module-init-tools may be built
earlier and if we don't add it here to the dependency chain, the build
order could be module-init-tools - host-elf-h-compat - kernel which
would not work on systems that have no /usr/include/elf.h.
Signed-off-by: Bernhard Walle
Add a dependency to the host tool and set the include flags properly.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
platforms/hosttools.in |1 +
platforms/kernel.in|1 +
rules/kernel.make |1 +
3 files changed, 3 insertions(+), 0 deletions(-)
diff --git a/platforms
, it does
nothing on make install. If not, it installs its own copy of elf.h in
make install to the include directory set in autoconf.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
rules/host-elf-h-compat.in |4
rules/host-elf-h-compat.make | 32
The build system of sitecopy uses 'uname' to determine the operating
system for which it compiles sitecopy. However, on non-Linux hosts, this
leads to wrong results.
Fixes Darwin (tested on Mac OS 10.7 Lion).
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
v3: Same as v1
* Andreas Bießmann andr...@biessmann.de [2012-01-22 15:49]:
+config ROOTFS_FW_ENV_CONFIG
+ bool
+ prompt install /etc/fw_env.config
+ default y
Shouldn't this default to n if there's no generic fallback file
(in the generic/ directory)?
Regards,
Bernhard
--
ptxdist mailing
Tested on 10.7 (Lion).
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
bin/ptxdist |2 ++
1 files changed, 2 insertions(+), 0 deletions(-)
diff --git a/bin/ptxdist b/bin/ptxdist
index 74a52ad..4edbac3 100755
--- a/bin/ptxdist
+++ b/bin/ptxdist
@@ -2093,6 +2093,8 @@ setup_config
Am 21.01.12 19:22, schrieb Michael Olbrich:
I don't know fw_printenv. Please elaborate.
This is a tool to print the u-boot environment in Linux. fw_setenv can
be used to change the u-boot environment from Linux.
Regards,
Bernhard
--
ptxdist mailing list
ptxdist@pengutronix.de
Am 19.01.12 16:02, schrieb Andreas Bießmann:
On OS X (and other maybe other systems) the default provided ar is not a
GNU version and does not understand all the required switches.
That should also fix the problem that ar on Darwin (regardless whether
it's the GNU variant built by Mac Ports or
The build system of sitecopy uses 'uname' to determine the operating
system for which it compiles sitecopy. However, on non-Linux hosts, this
leads to wrong results.
Fixes Darwin (tested on Mac OS 10.7 Lion).
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
patches/sitecopy-0.16.6/autoconf
The build system of zlib uses 'uname' to determine the operating system
for which it compiles zlib. However, on non-Linux hosts, this leads to
wrong results.
Fixes Darwin (tested on Mac OS 10.7 Lion), doesn't break Linux (verified
with current Arch Linux on x86_64).
Signed-off-by: Bernhard Walle
would also work, but this solution is
cleaner.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
.../0003-darwin_directio_fix.patch | 46
patches/e2fsprogs-1.41.14/series |1 +
2 files changed, 47 insertions(+), 0 deletions(-)
create
Am 19.01.12 20:55, schrieb Michael Olbrich:
On Mon, Jan 16, 2012 at 09:30:25PM +0100, Bernhard Walle wrote:
The extension of shared libraries on Darwin is .dylib, not .so. While
the fakeroot script itself (which is part of the fakeroot distribution)
can handle Darwin, ptxdist calls 'fakeroot
The extension of shared libraries on Darwin is .dylib, not .so.
Because fakeroot doesn't need specification of the pre-load library
but finds the library automatically, remove calling fakeroot with -l.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
rules/pre/Rules.make |3 ++-
1 files
would also work, but this solution is
cleaner.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
.../0003-Fix-compilation-on-Darwin.patch | 44
patches/e2fsprogs-1.41.14/series |3 +-
2 files changed, 46 insertions(+), 1 deletions(-)
create
Hi Jürgen,
Am 18.01.2012 11:05, schrieb Juergen Beisert:
Can you send me the config.log file of the sitecopy package, after
*preparing* (prepare stage) it? Please run it one times without your patch
shown above.
I have no other system than Linux here, to check what the autotools really
Am 18.01.12 11:23, schrieb Juergen Beisert:
Hi Bernhard,
Bernhard Walle wrote:
Am 18.01.2012 11:05, schrieb Juergen Beisert:
Can you send me the config.log file of the sitecopy package, after
*preparing* (prepare stage) it? Please run it one times without your
patch shown above.
I have
I added this option in rootfs and not in module-init-tools because there
are various implementations of modprobe that are using that file
(module-init-tools, busybox and possibly kmod in future).
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
rules/rootfs.in | 10 ++
rules
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
rules/sqlite.make |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/rules/sqlite.make b/rules/sqlite.make
index 9491c4e..451c3b3 100644
--- a/rules/sqlite.make
+++ b/rules/sqlite.make
@@ -20,8 +20,8 @@ PACKAGES
-by: Bernhard Walle bernh...@bwalle.de
---
rules/sitecopy.make |3 ++-
1 files changed, 2 insertions(+), 1 deletions(-)
diff --git a/rules/sitecopy.make b/rules/sitecopy.make
index 44a7b8a..c5af415 100644
--- a/rules/sitecopy.make
+++ b/rules/sitecopy.make
@@ -40,7 +40,8 @@ SITECOPY_AUTOCONF
.
This patch adds a simple check, explicitly checking for Darwin in the
'uname -s' output. Verified to fix the build on Mac OS 10.7 (Lion).
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
rules/pre/Rules.make |9 -
1 files changed, 8 insertions(+), 1 deletions(-)
diff --git a/rules/pre
:
% grep -r tzsetwall /usr/include/
/usr/include/time.h:void tzsetwall(void);
So this patch doesn't break Linux.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
rules/host-tz-database.make |2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/rules/host-tz
The build system of zlib uses 'uname' to determine the operating system
for which it compiles zlib. However, on non-Linux hosts, this leads to
wrong results.
Fixes Darwin (tested on Mac OS 10.7 Lion), doesn't break Linux (verified
with current Arch Linux on x86_64).
Signed-off-by: Bernhard Walle
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
rules/i2c-tools.make |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/rules/i2c-tools.make b/rules/i2c-tools.make
index 6ee8582..a8d2d00 100644
--- a/rules/i2c-tools.make
+++ b/rules/i2c-tools.make
@@ -16,8 +16,8
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
...-only-look-for-Python-with-version-number.patch |0
patches/{cmake-2.8.6 = cmake-2.8.7}/series|0
rules/host-cmake.make |4 ++--
3 files changed, 2 insertions(+), 2 deletions(-)
rename
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
rules/strace.make |6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/rules/strace.make b/rules/strace.make
index 3598c9c..2ac89e4 100644
--- a/rules/strace.make
+++ b/rules/strace.make
@@ -18,10 +18,10 @@ PACKAGES
* Juergen Beisert j...@pengutronix.de [2011-11-08 15:29]:
bernh...@bwalle.de wrote:
[...]
diff --git a/rules/alsa-lib.make b/rules/alsa-lib.make
index 929ab4e..bfec83f 100644
--- a/rules/alsa-lib.make
+++ b/rules/alsa-lib.make
@@ -207,6 +207,12 @@ endif
Hi,
* Erwin Rol mailingli...@erwinrol.com [2011-11-03 23:48]:
could someone explain the idea behind the following ptxdist options ?
Build Type (development)
[ ] use pre-built production release
[ ] built on top of another platform
Don't know about that options, but I'm also
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
rules/host-tz-database.make |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/rules/host-tz-database.make b/rules/host-tz-database.make
index 180e2bd..453c120 100644
--- a/rules/host-tz-database.make
+++ b/rules/host
From: Bernhard Walle wa...@corscience.de
Since we have host-flex, it makes also sense to get independent of
the bison installation of the host.
Signed-off-by: Bernhard Walle wa...@corscience.de
---
rules/host-bison.in |8
rules/host-bison.make | 32
From: Bernhard Walle wa...@corscience.de
This commit contains also two small whitespace fixes.
Signed-off-by: Bernhard Walle wa...@corscience.de
---
rules/bash.in |7 ---
1 files changed, 4 insertions(+), 3 deletions(-)
diff --git a/rules/bash.in b/rules/bash.in
index 7ef65a9..fd5a9a6
See http://www.sqlite.org/releaselog/3_7_9.html for a full changelog.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
rules/sqlite.make |4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/rules/sqlite.make b/rules/sqlite.make
index a86e394..9491c4e 100644
--- a/rules
* Robert Schwebel r.schwe...@pengutronix.de [2011-11-01 17:28]:
On Tue, Nov 01, 2011 at 02:07:59AM +0100, Ludovic BOUE wrote:
My program returns empty string from iconv function.
I think iconv is not well tested (if at all) in ptxdist, so any help in
that area is welcome!
Basically it
* Jon Ringle j...@ringle.org [2011-10-19 21:55]:
What do you think about the bash implementation found here:
http://lists.us.dell.com/pipermail/dkms-devel/2004-July/000142.html
If it is correct and works, why not.
Regards,
Bernhard
--
ptxdist mailing list
ptxdist@pengutronix.de
Am 19.10.2011 19:46, schrieb Robert Schwebel:
On Tue, Oct 18, 2011 at 04:42:46AM -0400, Jon Ringle wrote:
After more testing, I found that there were other situations where the version
comparison does not work well.
I ended up replacing the implementation of ptxd_ipkg_rev_smaller() with
* Andreas Bießmann andr...@biessmann.de [2011-10-17 09:06]:
It's not really nice, but I think it's better than patching lots of
tools. And having scripts as executable programs is IMO still better
than having to call the interpreter manually (which is necessary with
your $PTXDIST_SED
Hi,
* Michael Olbrich m.olbr...@pengutronix.de [2011-10-14 16:04]:
Hmmm, I just had another idea: How about fixing the environment in ptxdist:
mkdir TMP_BIN/
ln -s @PYTHON@ TMP_BIN/python
ln -s @SED@ TMP_BIN/sed
[...]
export PATH=TMP_BIN:$PATH
[...]
just do this as early as possible
Hi Michael,
* Michael Olbrich m.olbr...@pengutronix.de [2011-10-10 14:53]:
I've just released PTXdist 2011.10.1. Just one patch this time, but an
annoying issue, the Qt 4.7.4 md5sum changed.
could you also provide a tarball at
http://www.pengutronix.de/software/ptxdist/download/?
Regards,
Am 20.09.11 05:37, schrieb Jon Ringle:
diff --git a/rules/udev.make b/rules/udev.make
index ac2f0d4..b57bd01 100644
--- a/rules/udev.make
+++ b/rules/udev.make
@@ -47,6 +47,11 @@ $(UDEV_SOURCE):
#
# autoconf
+
+UDEV_CONF_ENV := \
+ $(CROSS_ENV) \
+
Hi,
thanks for the review. I'll post the updated patch.
Am 29.08.11 12:04, schrieb Michael Olbrich:
--- /dev/null
+++ b/rules/pre/bison.make
@@ -0,0 +1,5 @@
+# -*-makefile-*-
+
+ifdef PTXCONF_HOST_BISON
+export BISON_PKGDATADIR=$(PTXCONF_SYSROOT_HOST)/share/bison
+endif
What is
From: Bernhard Walle wa...@corscience.de
Since we have host-flex, it makes also sense to get independent of
the bison installation of the host.
Signed-off-by: Bernhard Walle wa...@corscience.de
---
rules/host-bison.in |8
rules/host-bison.make | 32
Am 16.08.2011 15:56, schrieb Michael Olbrich:
On Mon, Aug 15, 2011 at 07:34:58PM +0200, bernh...@bwalle.de wrote:
From: Bernhard Walle wa...@corscience.de
When a toolchain without kernel headers (such as the Lite version from
Codesourcery) or with very old kernel header is used, udev doesn't
Hi,
if I run ./fixup_ptxconfigs.sh, I get an error message in line 280 that
./p doesn't exist. What is ./p?
Regards,
Bernhard
--
ptxdist mailing list
ptxdist@pengutronix.de
* Marc Kleine-Budde m...@pengutronix.de [2011-08-15 12:25]:
On 08/15/2011 12:20 PM, Juergen Beisert wrote:
Marc Kleine-Budde wrote:
On 08/14/2011 12:51 PM, Bernhard Walle wrote:
The toolchain has been tested with a so-called Bifferboard
(http://bifferos.co.uk/) which has a S3282 BGA CPU
Am 15.08.2011 12:31, schrieb Marc Kleine-Budde:
You have to build the compiler with no fpu support - and probably the
glibc, too.
In theory it's clear, but it didn't work. I'll check which args I did
use and will post them, once I'm at home...
Regards,
Bernhard
--
ptxdist mailing list
While valgrind shouldn't check the Linux kernel version of the host, it
does and this commit fixes the build if the host is running a 3.0
kernel.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
.../0002-configure-fix-build-with-linux-3.0.patch | 47
patches/valgrind
While it's a good idea to avoid running autoreconf just because two
letters that have changed, it makes sense to add the change in the
configure.in file, so that autoreconf doesn't break the build.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
...ure-allow-all-arm-don-t-restrict-to-armv7
Hi,
Am 01.08.2011 08:50, schrieb Raz:
Hello
I am trying to create ext3 image and I came across the bellow discussion.
See that commit
http://git.pengutronix.de/?p=ptxdist.git;a=commit;h=2e2bf6b952c86f4ee25c61876b80bd290e8389b9.
It's not yet in a released version of ptxdist, so you need to use
* Josef Holzmayr holzm...@rsi-elektrotechnik.de [2011-07-15 09:39]:
How can I determine how much RAM is
available for my use (user application).
Type man free into a terminal application on your development
machine and use the knowledge gained to apply the command to your
target. :-)
cat
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
platforms/image_ext2.in|8
rules/post/image_ext2.make | 13 +
2 files changed, 21 insertions(+), 0 deletions(-)
diff --git a/platforms/image_ext2.in b/platforms/image_ext2.in
index b28c5e3..c26efd0 100644
This fixes commit 51aaa12b07b03b8e60909539a3dabf6594f462f2. Without that
fix, the CMAKE_SYSTEM_PROCESSOR entry in the toolchain file was actually
empty.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
rules/post/ptxd_make_cmake_toolchain.make |1 +
1 files changed, 1 insertions(+), 0
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
platforms/image_ext2.in|8
rules/post/image_ext2.make | 12
2 files changed, 20 insertions(+), 0 deletions(-)
diff --git a/platforms/image_ext2.in b/platforms/image_ext2.in
index b28c5e3..c26efd0 100644
Also install Gnuplot runtime files which are needed for some formats
like Postscript or HTML5/Javascript when the configuration option is
selected.
Disable lua regardless whether lua is available or not.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
patches/gnuplot-4.2.6/series
* Alexander Dahl p...@lespocky.de [2011-06-23 18:10]:
Hei hei,
On Thu, 23 Jun 2011 17:48:02 +0200, Bernhard Walle wrote:
Set the file type to kconfig. The *.make files have a modeline for vim,
so why not also for the Kconfig files.
% grep vim: rules/*.in
rules/argtable2.in:# vim
* Alexander Dahl p...@lespocky.de [2011-06-23 19:42]:
Hello,
On 23.06.2011 18:37, Bernhard Walle wrote:
Well, IMO setting the filetype is better than just the syntax, because
then you can add
autocmd FileType kconfig set ts=8 sw=8 noet
autocmd FileType makeset ts=8 sw=8
* Michael Olbrich m.olbr...@pengutronix.de [2011-06-21 12:25]:
Geht auf !x86 nicht:
./doc2gih ./gnuplot.doc gnuplot.gih
./doc2gih: 1: Syntax error: word unexpected (expecting ))
doc2gih wird fuers target compiliert.
GNUPLOT_MAKE_OPT := -C src/
GNUPLOT_INSTALL_OPT :=
Also install Gnuplot runtime files which are needed for some formats
like Postscript or HTML5/Javascript.
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
patches/gnuplot-4.2.6/series |1 -
.../{gnuplot-4.2.6 = gnuplot-4.4.3}/autogen.sh|0
.../gnuplot-4.4.3
Hi,
* Michael Olbrich m.olbr...@pengutronix.de [2011-06-18 10:19]:
patchin: quilt: apply 'ptxdist-git/patches/lpc21isp-1.79/series'
Applying patch lpc21isp-fix-arm.diff
patching file lpc21isp.h
Hunk #1 FAILED at 37.
1 out of 1 hunk FAILED -- rejects in file lpc21isp.h
Patch
* Bernhard Walle bernh...@bwalle.de [2011-06-20 09:57]:
* Michael Olbrich m.olbr...@pengutronix.de [2011-06-18 10:19]:
patchin: quilt: apply 'ptxdist-git/patches/lpc21isp-1.79/series'
Applying patch lpc21isp-fix-arm.diff
patching file lpc21isp.h
Hunk #1 FAILED at 37.
1 out of 1 hunk
* Bernhard Walle bernh...@bwalle.de [2011-06-20 10:46]:
* Bernhard Walle bernh...@bwalle.de [2011-06-20 09:57]:
* Michael Olbrich m.olbr...@pengutronix.de [2011-06-18 10:19]:
patchin: quilt: apply 'ptxdist-git/patches/lpc21isp-1.79/series'
Applying patch lpc21isp-fix-arm.diff
Signed-off-by: Bernhard Walle bernh...@bwalle.de
---
rules/gnuplot.make |9 +
1 files changed, 5 insertions(+), 4 deletions(-)
diff --git a/rules/gnuplot.make b/rules/gnuplot.make
index 4de83b0..d8fabf6 100644
--- a/rules/gnuplot.make
+++ b/rules/gnuplot.make
@@ -16,8 +16,8
201 - 300 of 313 matches
Mail list logo