Pull request for branch yem-host-image-tools
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. | '------------------------------^-------^------------------^--------------------'
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
@@ -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
@@ -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
PATCH 2/6 package/dosfstools: add host-package selection
In reply to Yann E. MORINAdd dosfstools as a host-package selection in the menuconfig.
---
v1 -> v2
- use $(HOST_CONFIGURE_OPTS) (Thomas)
--- package/Config.in.host | 1 + package/dosfstools/Config.in.host | 6 ++++++ package/dosfstools/dosfstools.mk | 9 +++++++++ 3 files changed, 16 insertions(+), 0 deletions(-) create mode 100644 package/dosfstools/Config.in.host
package/Config.in.host +1 -0
@@ -1,6 +1,7 @@ menu "Host utilities" source "package/dfu-util/Config.in.host" +source "package/dosfstools/Config.in.host" source "package/e2fsprogs/Config.in.host" source "package/lpc3250loader/Config.in.host" source "package/omap-u-boot-utils/Config.in.host"
package/dosfstools/Config.in.host new file +6 -0
@@ -0,0 +1,6 @@ +config BR2_PACKAGE_HOST_DOSFSTOOLS + bool "host dosfstools" + help + Tools for creating and checking DOS FAT filesystems. + + http://www.daniel-baumann.ch/software/dosfstools/
package/dosfstools/dosfstools.mk +9 -0
@@ -43,4 +43,13 @@ define DOSFSTOOLS_CLEAN_CMDS -$(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
PATCH 3/6 package/libconfuse: add host variant
In reply to Yann E. MORINNeeded later by genimage.
--- package/libconfuse/libconfuse.mk | 1 + 1 files changed, 1 insertions(+), 0 deletions(-)
package/libconfuse/libconfuse.mk +1 -0
@@ -10,3 +10,4 @@ LIBCONFUSE_INSTALL_STAGING = YES LIBCONFUSE_CONF_OPT = --disable-rpath $(eval $(autotools-package)) +$(eval $(host-autotools-package))
Signature · 1 line
1.7.2.5
PATCH 4/6 package/genimage: new host-only package
In reply to Yann E. MORIN--- package/Config.in.host | 1 + package/genimage/Config.in.host | 12 ++++++++++++ package/genimage/genimage.mk | 12 ++++++++++++ 3 files changed, 25 insertions(+), 0 deletions(-) create mode 100644 package/genimage/Config.in.host create mode 100644 package/genimage/genimage.mk
package/Config.in.host +1 -0
@@ -3,6 +3,7 @@ menu "Host utilities" source "package/dfu-util/Config.in.host" source "package/dosfstools/Config.in.host" source "package/e2fsprogs/Config.in.host" +source "package/genimage/Config.in.host" source "package/lpc3250loader/Config.in.host" source "package/omap-u-boot-utils/Config.in.host" source "package/openocd/Config.in.host"
package/genimage/Config.in.host new file +12 -0
@@ -0,0 +1,12 @@ +config BR2_PACKAGE_HOST_GENIMAGE + bool "host genimage" + help + genimage is a tool to generate multiple filesystem and flash images + from a given root filesystem tree. genimage is intended to be run + in a fakeroot environment. + + Configuration is done is a config file parsed by libconfuse. Options + like the path to tools can be given in environment variables, the config + file or commandline switches. + + http://www.pengutronix.de/software/genimage/
package/genimage/genimage.mk new file +12 -0
@@ -0,0 +1,12 @@ +############################################################# +# +# genimage +# +############################################################## + +GENIMAGE_VERSION = 4 +GENIMAGE_SOURCE = genimage-$(GENIMAGE_VERSION).tar.xz +GENIMAGE_SITE = http://www.pengutronix.de/software/genimage/download/ +GENIMAGE_DEPENDENCIES = libconfuse + +$(eval $(host-autotools-package))
Signature · 1 line
1.7.2.5
PATCH 4/6 package/genimage: new host-only package
Dear Yann E. MORIN,
package/genimage/genimage.mk line 10
Quoted 4 lines · Yann E. MORIN
package/genimage/genimage.mk new file +4
-0+GENIMAGE_VERSION = 4 +GENIMAGE_SOURCE = genimage-$(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
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 = libconfuseEven though it is a host package, we could add:
GENIMAGE_LICENSE= GPLv2Unfortunately, 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. | '------------------------------^-------^------------------^--------------------'
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
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
@@ -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
@@ -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
@@ -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
@@ -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 $(host-autotools-package))
Signature · 1 line
1.7.2.5
PATCH 5/6 package/genpart: new host-only package
Dear Yann E. MORIN,
package/genpart/genpart.mk line 9
Quoted 12 lines · Yann E. MORIN
package/genpart/genpart.mk new file +9
-0@@ -0,0 +1,11 @@ +############################################################# +# +# genpart +# +############################################################# + +GENPART_VERSION = 1.0.2 +GENPART_SOURCE = genpart-$(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
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 authorsAnd 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. | '------------------------------^-------^------------------^--------------------'
PATCH 6/6 package/mtools: new host-package
In reply to Yann E. MORIN---
v1 -> v2
- remove extraneous spaces, use $(BR2_GNU_MIRROR) (Thomas)
--- package/Config.in.host | 1 + package/mtools/Config.in.host | 9 +++++++++ package/mtools/mtools.mk | 15 +++++++++++++++ 3 files changed, 25 insertions(+), 0 deletions(-) create mode 100644 package/mtools/Config.in.host create mode 100644 package/mtools/mtools.mk
package/Config.in.host +1 -0
@@ -6,6 +6,7 @@ source "package/e2fsprogs/Config.in.host" source "package/genimage/Config.in.host" source "package/genpart/Config.in.host" source "package/lpc3250loader/Config.in.host" +source "package/mtools/Config.in.host" source "package/omap-u-boot-utils/Config.in.host" source "package/openocd/Config.in.host" source "package/sam-ba/Config.in.host"
package/mtools/Config.in.host new file +9 -0
@@ -0,0 +1,9 @@ +config BR2_PACKAGE_HOST_MTOOLS + bool "host mtools" + help + Mtools is a collection of utilities to access MS-DOS disks from + Unix without mounting them. It supports Win'95 style long file + names, OS/2 Xdf disks and 2m disks (store up to 1992k on a high + density 3 1/2 disk). + + http://www.tux.org/pub/tux/mtools/
package/mtools/mtools.mk new file +15 -0
@@ -0,0 +1,15 @@ +############################################################# +# +# dosfstools +# +############################################################# + +MTOOLS_VERSION = 4.0.18 +MTOOLS_SOURCE = mtools-$(MTOOLS_VERSION).tar.bz2 +MTOOLS_SITE = $(BR2_GNU_MIRROR)/mtools/ + +MTOOLS_CONF_OPT = --enable-xdf \ + --enable-vold \ + --enable-new-vold + +$(eval $(host-autotools-package))
Signature · 1 line
1.7.2.5
PATCH 6/6 package/mtools: new host-package
Yann, Luca,
Cc'ing Luca, there is a licensing infrastructure question below.
package/mtools/mtools.mk line 9
Quoted 3 lines · Yann E. MORIN
package/mtools/mtools.mk new file +3
-0+MTOOLS_VERSION = 4.0.18 +MTOOLS_SOURCE = mtools-$(MTOOLS_VERSION).tar.bz2 +MTOOLS_SITE = $(BR2_GNU_MIRROR)/mtools/
MTOOLS_LICENSE = GPLv3+MTOOLS_LICENSE_FILES = COPYING
However, this makes me realize that we don't distinguish host packages from target packages in the manifest.csv file. For example, once I had the licensing informations for this mtools file, I have the following line in my manifest.csv:
"mtools","4.0.18","GPLv3+","COPYING","mtools-4.0.18.tar.bz2"
And then, someone might think "damn, I have some GPLv3 software in my system, I now need to comply with the additional special requirements of GPLv3", while in fact it's not the case (as far as I know) because mtools is not distributed on the embedded device. Maybe we need to explicit which packages are on the target, which packages are on the host, and which packages are on both, no?
Best regards,
Thomas
Signature · 4 lines
Thomas Petazzoni, Free Electrons Kernel, drivers, real-time and embedded Linux development, consulting, training and support. http://free-electrons.com
PATCH 6/6 package/mtools: new host-package
Thomas, Luca, All,
Quoted 8 lines · On Sunday 10 March 2013 Thomas Petazzoni wrote:
package/mtools/mtools.mk line 9
Quoted 3 lines · Yann E. MORIN
package/mtools/mtools.mk new file +3
-0+MTOOLS_VERSION = 4.0.18 +MTOOLS_SOURCE = mtools-$(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. | '------------------------------^-------^------------------^--------------------'
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
PATCH 6/6 package/mtools: new host-package
In reply to Yann E. MORINQuoted 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= COPYINGDone.
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
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= COPYINGDone.
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
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
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
Thomas Petazzoni
Tested-byDear 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