Skip to contentSkip to conversation

Pull request for branch yem-host-image-tools

GIT PULL New Yann E. MORIN 5 participants 20 messages Mar 7, 2013 – Mar 15, 2013

lore
Share & tools
lore
#0 lore raw

Hello All!

This series adds a few host packages usefull for manipulating the generated filesystem images.

It is of much interest now that the post-image series from Thomas has been applied.

I plan on submitting sample post-{build,image} scripts demonstrating use of those packages in the (hopefuly near) future (my current scripts are working great, but need some cleanup first).

Changes v3 -> v4:
- rebased on master

Changes v2 -> v3:
- rebased onto the 'next' branch for Peter

Changes v1 -> v2:
- use $(HOST_CONFIGURE_OPTS) instead of hand-crafted CC et al. (Thomas)
- use $(BR2_GNU_MIRROR) (Thomas)
- strip superfluous spaces (Thomas)

The following changes since commit 203c30796e232faaf3a7de19090657bf3e7db16a:

rpi-userland: provides OpenGL ES, EGL and OpenVG (2013-03-06 22:18:39 +0100)

are available in the git repository at:
git://gitorious.org/buildroot/buildroot.git yem-host-image-tools

Yann E. MORIN (6):
package/e2fsprogs: add host-package selection package/dosfstools: add host-package selection package/libconfuse: add host variant package/genimage: new host-only package package/genpart: new host-only package package/mtools: new host-package

 package/Config.in.host                        |    5 +++++
 package/dosfstools/Config.in.host             |    6 ++++++
 package/dosfstools/dosfstools.mk              |    9 +++++++++
 package/e2fsprogs/Config.in.host              |    6 ++++++
 package/genimage/Config.in.host               |   12 ++++++++++++
 package/genimage/genimage.mk                  |   12 ++++++++++++
 package/genpart/Config.in.host                |    8 ++++++++
 package/genpart/genpart-fix-return-code.patch |   25 +++++++++++++++++++++++++
 package/genpart/genpart.mk                    |   11 +++++++++++
 package/libconfuse/libconfuse.mk              |    1 +
 package/mtools/Config.in.host                 |    9 +++++++++
 package/mtools/mtools.mk                      |   15 +++++++++++++++
 12 files changed, 119 insertions(+), 0 deletions(-)
 create mode 100644 package/dosfstools/Config.in.host
 create mode 100644 package/e2fsprogs/Config.in.host
 create mode 100644 package/genimage/Config.in.host
 create mode 100644 package/genimage/genimage.mk
 create mode 100644 package/genpart/Config.in.host
 create mode 100644 package/genpart/genpart-fix-return-code.patch
 create mode 100644 package/genpart/genpart.mk
 create mode 100644 package/mtools/Config.in.host
 create mode 100644 package/mtools/mtools.mk

Regards,
Yann E. MORIN

Signature · 6 lines
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'
#1 lore raw

PATCH 1/6 package/e2fsprogs: add host-package selection

Add the e2fsprogs as a host-package selection in the menuconfig.

---
 package/Config.in.host           |    1 +
 package/e2fsprogs/Config.in.host |    6 ++++++
 2 files changed, 7 insertions(+), 0 deletions(-)
 create mode 100644 package/e2fsprogs/Config.in.host
package/Config.in.host +1 -0
index 79050f2..8813f5b 100644
@@ -1,6 +1,7 @@
 menu "Host utilities"
 
 source "package/dfu-util/Config.in.host"
+source "package/e2fsprogs/Config.in.host"
 source "package/lpc3250loader/Config.in.host"
 source "package/omap-u-boot-utils/Config.in.host"
 source "package/openocd/Config.in.host"
package/e2fsprogs/Config.in.host new file +6 -0
new file mode 100644
index 0000000..ea6a0bd
@@ -0,0 +1,6 @@
+config BR2_PACKAGE_HOST_E2FSPROGS
+	bool "host e2fsprogs"
+	help
+	  The EXT2 file system utilities.
+	  
+	  http://e2fsprogs.sf.net
Signature · 1 line
1.7.2.5
#2 lore raw

PATCH 2/6 package/dosfstools: add host-package selection

In reply to Yann E. MORIN

Add dosfstools as a host-package selection in the menuconfig.

Cc Thomas Petazzoni <[email protected]>

---
v1 -> v2
- use $(HOST_CONFIGURE_OPTS) (Thomas)

---
 package/Config.in.host            |    1 +
 package/dosfstools/Config.in.host |    6 ++++++
 package/dosfstools/dosfstools.mk  |    9 +++++++++
 3 files changed, 16 insertions(+), 0 deletions(-)
 create mode 100644 package/dosfstools/Config.in.host
package/Config.in.host +1 -0
index 8813f5b..43f292c 100644
@@ -1,6 +1,7 @@
 menu "Host utilities"
 
 source "package/dfu-util/Config.in.host"
+source "package/dosfstools/Config.in.host"
 source "package/e2fsprogs/Config.in.host"
 source "package/lpc3250loader/Config.in.host"
 source "package/omap-u-boot-utils/Config.in.host"
package/dosfstools/Config.in.host new file +6 -0
new file mode 100644
index 0000000..9f9ceab
@@ -0,0 +1,6 @@
+config BR2_PACKAGE_HOST_DOSFSTOOLS
+	bool "host dosfstools"
+	help
+	  Tools for creating and checking DOS FAT filesystems.
+	  
+	  http://www.daniel-baumann.ch/software/dosfstools/
package/dosfstools/dosfstools.mk +9 -0
index 3a0856a..332f2ee 100644
@@ -43,4 +43,13 @@ define DOSFSTOOLS_CLEAN_CMDS
 	-$(MAKE) -C $(@D) clean
 endef
 
+define HOST_DOSFSTOOLS_BUILD_CMDS
+	$(MAKE) $(HOST_CONFIGURE_OPTS) -C $(@D)
+endef
+
+define HOST_DOSFSTOOLS_INSTALL_CMDS
+	$(MAKE) -C $(@D) $(HOST_CONFIGURE_OPTS) PREFIX=$(HOST_DIR)/usr install
+endef
+
 $(eval $(generic-package))
+$(eval $(host-generic-package))
Signature · 1 line
1.7.2.5
#3 lore raw

PATCH 3/6 package/libconfuse: add host variant

In reply to Yann E. MORIN

Needed later by genimage.

---
 package/libconfuse/libconfuse.mk |    1 +
 1 files changed, 1 insertions(+), 0 deletions(-)
package/libconfuse/libconfuse.mk +1 -0
index 2fa8241..d88a209 100644
@@ -10,3 +10,4 @@ LIBCONFUSE_INSTALL_STAGING = YES
 LIBCONFUSE_CONF_OPT = --disable-rpath
 
 $(eval $(autotools-package))
+$(eval $(host-autotools-package))
Signature · 1 line
1.7.2.5
#4 lore raw

PATCH 4/6 package/genimage: new host-only package

In reply to Yann E. MORIN
---
 package/Config.in.host          |    1 +
 package/genimage/Config.in.host |   12 ++++++++++++
 package/genimage/genimage.mk    |   12 ++++++++++++
 3 files changed, 25 insertions(+), 0 deletions(-)
 create mode 100644 package/genimage/Config.in.host
 create mode 100644 package/genimage/genimage.mk
package/Config.in.host +1 -0
index 43f292c..c456d5d 100644
@@ -3,6 +3,7 @@ menu "Host utilities"
 source "package/dfu-util/Config.in.host"
 source "package/dosfstools/Config.in.host"
 source "package/e2fsprogs/Config.in.host"
+source "package/genimage/Config.in.host"
 source "package/lpc3250loader/Config.in.host"
 source "package/omap-u-boot-utils/Config.in.host"
 source "package/openocd/Config.in.host"
package/genimage/Config.in.host new file +12 -0
new file mode 100644
index 0000000..cfa267c
@@ -0,0 +1,12 @@
+config BR2_PACKAGE_HOST_GENIMAGE
+	bool "host genimage"
+	help
+	  genimage is a tool to generate multiple filesystem and flash images
+	  from a given root filesystem tree. genimage is intended to be run
+	  in a fakeroot environment.
+	  
+	  Configuration is done is a config file parsed by libconfuse. Options
+	  like the path to tools can be given in environment variables, the config
+	  file or commandline switches.
+	  
+	  http://www.pengutronix.de/software/genimage/
package/genimage/genimage.mk new file +12 -0
new file mode 100644
index 0000000..8fe9bf7
@@ -0,0 +1,12 @@
+#############################################################
+#
+# genimage
+#
+##############################################################
+
+GENIMAGE_VERSION        = 4
+GENIMAGE_SOURCE         = genimage-$(GENIMAGE_VERSION).tar.xz
+GENIMAGE_SITE           = http://www.pengutronix.de/software/genimage/download/
+GENIMAGE_DEPENDENCIES   = libconfuse
+
+$(eval $(host-autotools-package))
Signature · 1 line
1.7.2.5
#5 lore raw

PATCH 4/6 package/genimage: new host-only package

Dear Yann E. MORIN,

package/genimage/genimage.mk line 10

Quoted 4 lines · Yann E. MORIN
package/genimage/genimage.mk new file +4 -0
+GENIMAGE_VERSION        = 4
+GENIMAGE_SOURCE         = genimage-$(GENIMAGE_VERSION).tar.xz
+GENIMAGE_SITE           = http://www.pengutronix.de/software/genimage/download/
+GENIMAGE_DEPENDENCIES   = libconfuse

Even though it is a host package, we could add:

GENIMAGE_LICENSE = GPLv2

Unfortunately, there is no license file.

Thomas

Signature · 4 lines
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com
#6 lore raw

PATCH 4/6 package/genimage: new host-only package

Thomas, All,

Quoted 11 lines · On Sunday 10 March 2013 Thomas Petazzoni wrote:

package/genimage/genimage.mk line 10

Quoted 4 lines · Yann E. MORIN
package/genimage/genimage.mk new file +4 -0
+GENIMAGE_VERSION        = 4
+GENIMAGE_SOURCE         = genimage-$(GENIMAGE_VERSION).tar.xz
+GENIMAGE_SITE           = http://www.pengutronix.de/software/genimage/download/
+GENIMAGE_DEPENDENCIES   = libconfuse

Even though it is a host package, we could add:

GENIMAGE_LICENSE = GPLv2

Unfortunately, there is no license file.

I am a bit uneasy at adding licensing information for packages that are purely host packages.

As you said in your other mail, this might confuse a user reviewing the licensing terms.
But I will add it.

Thank you!

Regards,
Yann E. MORIN.

Signature · 6 lines
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'
#7 lore raw

PATCH 4/6 package/genimage: new host-only package

Dear Yann E. MORIN,

Quoted 5 lines · Yann E. MORIN

I am a bit uneasy at adding licensing information for packages that are purely host packages.

As you said in your other mail, this might confuse a user reviewing the licensing terms.

I think this is not a good reason to not add the license. We already have it in a number of packages that have a host variant. It should be something for the licensing infrastructure to handle.

Quoted 1 line

But I will add it.

Thanks!

Thomas

Signature · 4 lines
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com
#8 lore raw

PATCH 5/6 package/genpart: new host-only package

In reply to Yann E. MORIN
---
 package/Config.in.host                        |    1 +
 package/genpart/Config.in.host                |    8 ++++++++
 package/genpart/genpart-fix-return-code.patch |   25 +++++++++++++++++++++++++
 package/genpart/genpart.mk                    |   11 +++++++++++
 4 files changed, 45 insertions(+), 0 deletions(-)
 create mode 100644 package/genpart/Config.in.host
 create mode 100644 package/genpart/genpart-fix-return-code.patch
 create mode 100644 package/genpart/genpart.mk
package/Config.in.host +1 -0
index c456d5d..e97f6e8 100644
@@ -4,6 +4,7 @@ source "package/dfu-util/Config.in.host"
 source "package/dosfstools/Config.in.host"
 source "package/e2fsprogs/Config.in.host"
 source "package/genimage/Config.in.host"
+source "package/genpart/Config.in.host"
 source "package/lpc3250loader/Config.in.host"
 source "package/omap-u-boot-utils/Config.in.host"
 source "package/openocd/Config.in.host"
package/genpart/Config.in.host new file +8 -0
new file mode 100644
index 0000000..f986504
@@ -0,0 +1,8 @@
+config BR2_PACKAGE_HOST_GENPART
+	bool "host genpart"
+	help
+	  Generate a 16 byte partition table entry defined by command line
+	  arguments and dump it to stdout. No CHS magic is done, only lba
+	  entries are filled out.
+	  
+	  http://www.pengutronix.de/software/genpart/index_en.html
package/genpart/genpart-fix-return-code.patch new file +25 -0
new file mode 100644
index 0000000..14ff165
@@ -0,0 +1,25 @@
+genpart: return 0 if partition writen OK, with no do_magic
+
+If do_magic is false, we don't won't write the 0xaa55 magic.
+However, if we succesfully wrote the partition entry, we still
+want to return 0 to indicate success.
+
+Signed-off-by: "Yann E. MORIN" <[email protected]>
+Cc: Robert Schwebel <[email protected]>
+
+--
+Patch sent upstream, but no mailing list.
+Status: pending...
+
+diff -durN host-genpart-1.0.2.orig/src/genpart.c host-genpart-1.0.2/src/genpart.c
+--- host-genpart-1.0.2.orig/src/genpart.c	2009-12-20 21:54:56.000000000 +0100
++++ host-genpart-1.0.2/src/genpart.c	2013-01-22 23:13:24.109752579 +0100
+@@ -92,6 +92,8 @@
+ 		if (do_magic) {
+ 			if (fwrite(&magic, 2, 1, stdout) > 0)
+ 				rc=0;
++		} else {
++			rc=0;
+ 		}
+ 	}
+ 	return rc;
package/genpart/genpart.mk new file +11 -0
new file mode 100644
index 0000000..735333a
@@ -0,0 +1,11 @@
+#############################################################
+#
+# genpart
+#
+#############################################################
+
+GENPART_VERSION       = 1.0.2
+GENPART_SOURCE        = genpart-$(GENPART_VERSION).tar.bz2
+GENPART_SITE          = http://www.pengutronix.de/software/genpart/download/
+
+$(eval $(host-autotools-package))
Signature · 1 line
1.7.2.5
#9 lore raw

PATCH 5/6 package/genpart: new host-only package

Dear Yann E. MORIN,

package/genpart/genpart.mk line 9

Quoted 12 lines · Yann E. MORIN
package/genpart/genpart.mk new file +9 -0
@@ -0,0 +1,11 @@
+#############################################################
+#
+# genpart
+#
+#############################################################
+
+GENPART_VERSION       = 1.0.2
+GENPART_SOURCE        = genpart-$(GENPART_VERSION).tar.bz2
+GENPART_SITE          = http://www.pengutronix.de/software/genpart/download/

There's no license at all mentioned in the source code. It's true the code is pretty trivial, but anyway. Maybe you could add:

GENPART_LICENSE = Currently being decided by authors

And contact the Pengutronix people to get your patches merged and ask them to put a proper license in there.

Best regards,

Thomas

Signature · 4 lines
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com
#10 lore raw

PATCH 5/6 package/genpart: new host-only package

Thomas, All,

Quoted 18 lines · On Sunday 10 March 2013 Thomas Petazzoni wrote:

package/genpart/genpart.mk line 9

Quoted 12 lines · Yann E. MORIN
package/genpart/genpart.mk new file +9 -0
@@ -0,0 +1,11 @@
+#############################################################
+#
+# genpart
+#
+#############################################################
+
+GENPART_VERSION       = 1.0.2
+GENPART_SOURCE        = genpart-$(GENPART_VERSION).tar.bz2
+GENPART_SITE          = http://www.pengutronix.de/software/genpart/download/

There's no license at all mentioned in the source code. It's true the code is pretty trivial, but anyway. Maybe you could add:

Quoted 4 lines

GENPART_LICENSE = Currently being decided by authors

And contact the Pengutronix people to get your patches merged and ask them to put a proper license in there.

Path has already been accepted upstream, but they said they will probably never ever make another release of genpart, which in their opinion has been superseded by genimage.

I'll make sure to contact them to clarify the license.

Thank you!

Regards,
Yann E. MORIN.

Signature · 6 lines
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'
#11 lore raw

PATCH 6/6 package/mtools: new host-package

In reply to Yann E. MORIN
Cc Thomas Petazzoni <[email protected]>

---
v1 -> v2
- remove extraneous spaces, use $(BR2_GNU_MIRROR) (Thomas)

---
 package/Config.in.host        |    1 +
 package/mtools/Config.in.host |    9 +++++++++
 package/mtools/mtools.mk      |   15 +++++++++++++++
 3 files changed, 25 insertions(+), 0 deletions(-)
 create mode 100644 package/mtools/Config.in.host
 create mode 100644 package/mtools/mtools.mk
package/Config.in.host +1 -0
index e97f6e8..7c0dd20 100644
@@ -6,6 +6,7 @@ source "package/e2fsprogs/Config.in.host"
 source "package/genimage/Config.in.host"
 source "package/genpart/Config.in.host"
 source "package/lpc3250loader/Config.in.host"
+source "package/mtools/Config.in.host"
 source "package/omap-u-boot-utils/Config.in.host"
 source "package/openocd/Config.in.host"
 source "package/sam-ba/Config.in.host"
package/mtools/Config.in.host new file +9 -0
new file mode 100644
index 0000000..d534b5e
@@ -0,0 +1,9 @@
+config BR2_PACKAGE_HOST_MTOOLS
+	bool "host mtools"
+	help
+	  Mtools is a collection of utilities to access MS-DOS disks from
+	  Unix without mounting them. It supports Win'95 style long file
+	  names, OS/2 Xdf disks and 2m disks (store up to 1992k on a high
+	  density 3 1/2 disk).
+	  
+	  http://www.tux.org/pub/tux/mtools/
package/mtools/mtools.mk new file +15 -0
new file mode 100644
index 0000000..1542799
@@ -0,0 +1,15 @@
+#############################################################
+#
+# dosfstools
+#
+#############################################################
+
+MTOOLS_VERSION  = 4.0.18
+MTOOLS_SOURCE   = mtools-$(MTOOLS_VERSION).tar.bz2
+MTOOLS_SITE     = $(BR2_GNU_MIRROR)/mtools/
+
+MTOOLS_CONF_OPT = --enable-xdf        \
+                  --enable-vold       \
+                  --enable-new-vold
+
+$(eval $(host-autotools-package))
Signature · 1 line
1.7.2.5
#12 lore raw

PATCH 6/6 package/mtools: new host-package

Yann, Luca,

Cc'ing Luca, there is a licensing infrastructure question below.

package/mtools/mtools.mk line 9

Quoted 3 lines · Yann E. MORIN
package/mtools/mtools.mk new file +3 -0
+MTOOLS_VERSION  = 4.0.18
+MTOOLS_SOURCE   = mtools-$(MTOOLS_VERSION).tar.bz2
+MTOOLS_SITE     = $(BR2_GNU_MIRROR)/mtools/

MTOOLS_LICENSE = GPLv3+
MTOOLS_LICENSE_FILES = COPYING

However, this makes me realize that we don't distinguish host packages from target packages in the manifest.csv file. For example, once I had the licensing informations for this mtools file, I have the following line in my manifest.csv:

"mtools","4.0.18","GPLv3+","COPYING","mtools-4.0.18.tar.bz2"

And then, someone might think "damn, I have some GPLv3 software in my system, I now need to comply with the additional special requirements of GPLv3", while in fact it's not the case (as far as I know) because mtools is not distributed on the embedded device. Maybe we need to explicit which packages are on the target, which packages are on the host, and which packages are on both, no?

Best regards,

Thomas

Signature · 4 lines
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com
#13 lore raw

PATCH 6/6 package/mtools: new host-package

Thomas, Luca, All,

Quoted 8 lines · On Sunday 10 March 2013 Thomas Petazzoni wrote:

package/mtools/mtools.mk line 9

Quoted 3 lines · Yann E. MORIN
package/mtools/mtools.mk new file +3 -0
+MTOOLS_VERSION  = 4.0.18
+MTOOLS_SOURCE   = mtools-$(MTOOLS_VERSION).tar.bz2
+MTOOLS_SITE     = $(BR2_GNU_MIRROR)/mtools/

MTOOLS_LICENSE = GPLv3+
MTOOLS_LICENSE_FILES = COPYING

Done.

Quoted 5 lines

However, this makes me realize that we don't distinguish host packages from target packages in the manifest.csv file. For example, once I had the licensing informations for this mtools file, I have the following line in my manifest.csv:

Quoted 8 lines

"mtools","4.0.18","GPLv3+","COPYING","mtools-4.0.18.tar.bz2"

And then, someone might think "damn, I have some GPLv3 software in my system, I now need to comply with the additional special requirements of GPLv3", while in fact it's not the case (as far as I know) because mtools is not distributed on the embedded device. Maybe we need to explicit which packages are on the target, which packages are on the host, and which packages are on both, no?

Yes, fully agreed; this would be very puzzling.

I would suggest that we not install licensing information about host packages, unless they are also used on the target.

Regards,
Yann E. MORIN.

Signature · 6 lines
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'
#14 lore raw

PATCH 6/6 package/mtools: new host-package

Dear Yann E. MORIN,

Quoted 2 lines · Yann E. MORIN

I would suggest that we not install licensing information about host packages, unless they are also used on the target.

That would make sense, IMO.

Thomas

Signature · 4 lines
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com
#15 lore raw

PATCH 6/6 package/mtools: new host-package

In reply to Yann E. MORIN
Quoted 28 lines · Yann E. MORIN

Thomas, Luca, All,

Quoted 8 lines · On Sunday 10 March 2013 Thomas Petazzoni wrote:

package/mtools/mtools.mk line 9

Quoted 3 lines · Yann E. MORIN
package/mtools/mtools.mk new file +3 -0
+MTOOLS_VERSION  = 4.0.18
+MTOOLS_SOURCE   = mtools-$(MTOOLS_VERSION).tar.bz2
+MTOOLS_SITE     = $(BR2_GNU_MIRROR)/mtools/

MTOOLS_LICENSE = GPLv3+
MTOOLS_LICENSE_FILES = COPYING

Done.

Quoted 5 lines

However, this makes me realize that we don't distinguish host packages from target packages in the manifest.csv file. For example, once I had the licensing informations for this mtools file, I have the following line in my manifest.csv:

Quoted 8 lines

"mtools","4.0.18","GPLv3+","COPYING","mtools-4.0.18.tar.bz2"

And then, someone might think "damn, I have some GPLv3 software in my system, I now need to comply with the additional special requirements of GPLv3", while in fact it's not the case (as far as I know) because mtools is not distributed on the embedded device. Maybe we need to explicit which packages are on the target, which packages are on the host, and which packages are on both, no?

Quoted 4 lines

Yes, fully agreed; this would be very puzzling.

I would suggest that we not install licensing information about host packages, unless they are also used on the target.

It would probably be sound in most cases, but not compliant to some licenses, at least if there isa GPL package on the target. Take what the GPLv2 saysin clause 3:

Quoted 5 lines

The source code for a work means the preferred form of the work for making modifications to it. For an executable work, complete source code means all the source code for all modules it contains, plus any associated interface definition files, plus the scripts used to control compilation and installation of the executable.

What do the "scripts used to control compilation and installation"? We use a toolchain, the autotools, CMake, Python etc on the host to build packages. While these tools are arguably more complex than the average "script", they are "used to control the cpmpilation and installation".

Reading on:

Quoted 6 lines

However, as a
special exception, the source code distributed need not include anything that is normally distributed (in either source or binary form) with the major components (compiler, kernel, and so on) of the operating system on which the executable runs, unless that component itself accompanies the executable.

While this is saying you do not need to redistribute the Debian compiler, it is using "compiler" and "kernel" as two examples of the "scripts used to control compilation and installation". Hence my understanding that the toolchain should be redistributed, along with all of the host tools that are used to build anything on the target.

As far as the mtools are concerned, they are probably not used to produce the executables on the target, so they actually do not need to be in the redistribution list. But most other host packages are different.

Indeed, in the practice I'm retty sure almost nobody modify autotools, or CMake or other host packages, so even redistributing them would not disclose much industrial secret.

Bottom line, I think it would be useful to save in the package manifest an additional column stating whether the package is on the target, on the host, or both. But skipping host-only packages is not safe.

Luca

#16 lore raw

PATCH 6/6 package/mtools: new host-package

Hi,

Quoted 68 lines · Luca Ceresoli
Quoted 28 lines · Yann E. MORIN

Thomas, Luca, All,

Quoted 8 lines · On Sunday 10 March 2013 Thomas Petazzoni wrote:

package/mtools/mtools.mk line 9

Quoted 3 lines · Yann E. MORIN
package/mtools/mtools.mk new file +3 -0
+MTOOLS_VERSION  = 4.0.18
+MTOOLS_SOURCE   = mtools-$(MTOOLS_VERSION).tar.bz2
+MTOOLS_SITE     = $(BR2_GNU_MIRROR)/mtools/

MTOOLS_LICENSE = GPLv3+
MTOOLS_LICENSE_FILES = COPYING

Done.

Quoted 5 lines

However, this makes me realize that we don't distinguish host packages from target packages in the manifest.csv file. For example, once I had the licensing informations for this mtools file, I have the following line in my manifest.csv:

Quoted 8 lines

"mtools","4.0.18","GPLv3+","COPYING","mtools-4.0.18.tar.bz2"

And then, someone might think "damn, I have some GPLv3 software in my system, I now need to comply with the additional special requirements of GPLv3", while in fact it's not the case (as far as I know) because mtools is not distributed on the embedded device. Maybe we need to explicit which packages are on the target, which packages are on the host, and which packages are on both, no?

Quoted 4 lines

Yes, fully agreed; this would be very puzzling.

I would suggest that we not install licensing information about host packages, unless they are also used on the target.

It would probably be sound in most cases, but not compliant to some licenses, at least if there isa GPL package on the target. Take what the GPLv2 saysin clause 3:

Quoted 5 lines

The source code for a work means the preferred form of the work for making modifications to it. For an executable work, complete source code means all the source code for all modules it contains, plus any associated interface definition files, plus the scripts used to control compilation and installation of the executable.

What do the "scripts used to control compilation and installation"? We use a toolchain, the autotools, CMake, Python etc on the host to build packages. While these tools are arguably more complex than the average "script", they are "used to control the cpmpilation and installation".

Reading on:

Quoted 6 lines

However, as a
special exception, the source code distributed need not include anything that is normally distributed (in either source or binary form) with the major components (compiler, kernel, and so on) of the operating system on which the executable runs, unless that component itself accompanies the executable.

While this is saying you do not need to redistribute the Debian compiler, it is using "compiler" and "kernel" as two examples of the "scripts used to control compilation and installation". Hence my understanding that the toolchain should be redistributed, along with all of the host tools that are used to build anything on the target.

As far as the mtools are concerned, they are probably not used to produce the executables on the target, so they actually do not need to be in the redistribution list. But most other host packages are different.

Quoted 7 lines

Indeed, in the practice I'm retty sure almost nobody modify autotools, or CMake or other host packages, so even redistributing them would not disclose much industrial secret.

Bottom line, I think it would be useful to save in the package manifest an additional column stating whether the package is on the target, on the host, or both. But skipping host-only packages is not safe.

I follow Luca's reasoning here. As we discussed on the Buildroot Developer Day 2012, we will include the toolchain/host sources precisely for the above reasons, and so they should be described in a manifest. I also agree with Thomas' point about the confusion.

But, instead of an extra column, wouldn't it be more clear to separate the legal-license manifest files for host and target packages? This would result in a manifest-host.csv and manifest.csv. Packages that appear on both host and target, would be present in both files. An end user that is only interested in getting a list of target packages, can find that info in a glance.

Best regards,
Thomas

#17 lore raw

PATCH 6/6 package/mtools: new host-package

Quoted 3 lines · Thomas De Schampheleire

Hi,

On Tue, Mar 12, 2013 at 6:16 PM, Luca Ceresoli <[email protected]> wrote:

[snip]

Quoted 3 lines
Quoted 3 lines

Bottom line, I think it would be useful to save in the package manifest an additional column stating whether the package is on the target, on the host, or both. But skipping host-only packages is not safe.

I agree. And I have a second reason for that: I have one customer who
distributes an "SDK", i.e. host packages.

Quoted 11 lines

I follow Luca's reasoning here. As we discussed on the Buildroot Developer Day 2012, we will include the toolchain/host sources precisely for the above reasons, and so they should be described in a manifest. I also agree with Thomas' point about the confusion.

But, instead of an extra column, wouldn't it be more clear to separate the legal-license manifest files for host and target packages? This would result in a manifest-host.csv and manifest.csv. Packages that appear on both host and target, would be present in both files. An end user that is only interested in getting a list of target packages, can find that info in a glance.

Very good idea.

Regards,
Arnout

Signature · 6 lines
Arnout Vandecappelle                          arnout at mind be
Senior Embedded Software Architect            +32-16-286500
Essensium/Mind                                http://www.mind.be
G.Geenslaan 9, 3001 Leuven, Belgium           BE 872 984 063 RPR Leuven
LinkedIn profile: http://www.linkedin.com/in/arnoutvandecappelle
GPG fingerprint:  7CB5 E4CC 6C2E EFD4 6E3D A754 F963 ECAB 2450 2F1F
#18 lore raw

PATCH 6/6 package/mtools: new host-package

Dear Arnout Vandecappelle,

Quoted 8 lines · Arnout Vandecappelle
Quoted 6 lines

But, instead of an extra column, wouldn't it be more clear to separate the legal-license manifest files for host and target packages? This would result in a manifest-host.csv and manifest.csv. Packages that appear on both host and target, would be present in both files. An end user that is only interested in getting a list of target packages, can find that info in a glance.

Very good idea.

I don't know if creating a separate file is needed. Maybe just a need field in the CSV that says if it's a host package only, target package only, or both. But that's a minor detail, I don't have a strong opinion between a single CSV file, and two CSV files, as long as it is possible to identify what is on the target and what isn't.

Best regards,

Thomas

Signature · 4 lines
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com
#19 lore raw
In reply to Yann E. MORIN

Dear Yann E. MORIN,

Quoted 7 lines · Yann E. MORIN

Yann E. MORIN (6):
package/e2fsprogs: add host-package selection package/dosfstools: add host-package selection package/libconfuse: add host variant package/genimage: new host-only package package/genpart: new host-only package package/mtools: new host-package

For the six patches:

Signature · 4 lines
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com