Now, both the download and source-check code are iterating over
<pkg>_SOURCE, <pkg>_PATCH and <pkg>_EXTRA_DOWNLOADS elements, figuring
out whether they contain full URLs or not. Instead of doing this
repeatdly, this patch introduces an internal <pkg>_ALL_DOWNLOADS
variable, which contains the list of everything that needs to be
downloaded, with URLs already expanded to take into account <pkg>_SITE
if needed.
This allows to simplify quite significantly the .stamp_download and
source-check implementation.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
The .stamp_downloaded target is now only being used to really
download, and no longer for other activities like "source check" or
"external deps". So the check on DL_MODE being equal to DOWNLOAD is no
longer necessary.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Tested-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Acked-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
The only reason for the .stamp_rsync_sourced fake stamp file target to
exist was to handle the SOURCE_CHECK operation on packages using the
OVERRIDE_SRCDIR mechanism. Now that this is handled directly inside
$(1)-source-check, there is no longer any need for this part of the
code.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Tested-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
This commits extends the common package infrastructure with the
$(1)-source-check and $(1)-all-source-check targets.
The $(1)-source-check target simply calls the newly added
SOURCE_CHECK macro on all items to be downloaded.
The $(1)-all-source-check target will depend on the
$(1)-all-source-check targets of all dependent packages and the
$(1)-source-check target of the current package, which allows to do a
recursive source-check in the dependency tree.
[Thomas: move the code around a bit to avoid repeating the
<pkg>_OVERRIDE_SRCDIR condition, as suggested by Yann.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Now that the external-deps implementation relies on the per-package
<pkg>-all-external-deps and <pkg>-external-deps targets and no longer
on the 'source' target with a custom DL_MODE, we can get rid of the
support for the SHOW_EXTERNAL_DEPS DL_MODE value in the pkg-download
logic.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
The $(1)-all-{source,external-deps,legal-info} targets currently only
take care of the dependencies in <pkg>_DEPENDENCIES, but not
<pkg>_PATCH_DEPENDENCIES. This patch fixes that by introducing a
<pkg>_FINAL_ALL_DEPENDENCIES. It also reworks the $(1)-show-depends
target to make use of <pkg>_FINALL_ALL_DEPENDENCIES.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Using a colon or a space in a make target doesn't work, so they have to
be filtered out of the PKG_VERSION variable just like the / currently
already is.
This will be needed for date-based CVS versions.
Add a helper macro 'sanitize' in pkg-utils.mk to implement this. This
also moves the strip to the inner level for the target branch.
Signed-off-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Cc: Fabio Porcedda <fabio.porcedda@gmail.com>
Cc: Yann E. MORIN <yann.morin.1998@free.fr>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
For clarity, this commit renames the TARGETS variable to the more
meaningful PACKAGES variable. Indeed, only packages (handled by one of
the package infrastructures) should be listed in this variable, and
not other random non-package targets.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Some packages need to vampirise files from one or more other packages.
This is the case, for example, of the Linux kernel and its /extensions/.
Add a new type of dependencies, that are guaranteed to be extracted and
patched before a package is patched.
[Thomas: remove <pkg>-show-build-depends and <pkg>-show-patch-depends,
since they don't seem to really be necessary and very useful.]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This commit improves the pkg-generic package infrastructure so that it
declares all the <pkg>-<something> targets as PHONY, which they
actually are.
[Thomas:
- add some missing phony targets, noticed by Yann E. Morin;
- put one target per line, order alphabetically. Makes it easier to
spot missing targets.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
In order to implement the global 'make source', 'make legal-info' and
'make external-deps' using the package infrastructure logic, this
commit introduces a set of per-package targets that allow to
recursively run the source/legal-info/external-deps actions on the
dependencies of a given package.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
This commit implements a new <pkg>-external-deps target in the
pkg-generic package infrastructure. This target simply displays the
list of file names that are needed for <pkg> to build (source tarball,
patches, extra downloads).
In the case of the usage of <pkg>_OVERRIDE_SRCDIR, the path to the
overriden source directory is displayed, prepended by file://. This
matches what the code is currently doing in .stamp_rsync_sourced.
This new target is not currently being used in this commit, but
followup commits will make use of it.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Currently, a 'make clean' leaves the graphs/ subdirectory in the
output directory. This commit defines a GRAPHS_DIR variable, used by
the different graph-generating targets, and which gets cleaned up in
the 'clean' target.
[Thomas: use the new GRAPHS_DIR variable in more places, as suggested
by Yann.]
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The current logic for <pkg>_EXTRA_DOWNLOADS assumes that it is a list
of files, all hosted at <pkg>_SITE. However, just like for
<pkg>_PATCH, it may be useful to specify <pkg>_EXTRA_DOWNLOADS entries
that are hosted on a different site than the package <pkg>_SITE.
This commit implements this, by re-using the same logic as the one
used for <pkg>_PATCH.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: Romain Naour <romain.naour@openwide.fr>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
In the .stamp_downloaded hook, the logic to decide whether or not to
display the "Downloading" message is treating the check of
<pkg>_SOURCE as a special case. But in fact, there is no real reason
to do so: the existing loop used for <pkg>_PATCH and
<pkg>_EXTRA_DOWNLOADS could work just as well.
This commit therefore refactors this piece of code, to have a single
loop checking <pkg>_SOURCE, <pkg>_PATCH and <pkg>_EXTRA_DOWNLOADS.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
The .stamp_downloaded target displays the "Downloading" message even
if there is really something to download. However, this logic only
checks for <pkg>_SOURCE and <pkg>_PATCH: it does not check if
something needs to be downloaded in <pkg>_EXTRA_DOWNLOADS.
This commit fixes that, which makes sure that the "Downloading"
message is displayed if one of the items in <pkg>_EXTRA_DOWNLOADS has
not yet been downloaded.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: Romain Naour <romain.naour@openwide.fr>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Even though the .stamp_downloaded target is executed for each package
being built, the pkg-generic infrastructure tries to not display the
"Downloading" message when there is in fact nothing to download.
Unfortunately, the logic was incorrect for the patch download: it
forgot the fact that <pkg>_PATCH can contain either file names (in
which case we assume the patch should be downloaded from <pkg>_SITE),
or full URLs. The latter case was not properly handled, as we were
checking if $(DL_DIR)/<full URL> existed, while we should be testing
if $(DL_DIR)/`basename <full URL>` exists.
This patch fixes that, which makes sure the "Downloading" message is
displayed only when necessary.
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Reviewed-by: Romain Naour <romain.naour@openwide.fr>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
virtual packages are found by their version,
so we retrieve the version of all packages
Signed-off-by: Francois Perrad <francois.perrad@gadz.org>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Tested-by: Luca Ceresoli <luca@lucaceresoli.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Add a <pkg>-reinstall target, just like <pkg>-rebuild and
<pkg>-reconfigure, but that only retrigger the install step of the
package.
Signed-off-by: Doug Kehn <rdkehn@yahoo.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
When FOO_SOURCE is non-empty, FOO_MANIFEST_TARBALL is always set.
When FOO_SOURCE is empty, FOO_MANIFEST_TARBALL is not set, but also
never used, due to the if below which defuses the whole legal-info
processing for packages that have FOO_SOURCE explicitly set to an
empty string.
So get rid of the default assignment to "not saved".
Do it for FOO_MANIFEST_SITE as well: it is pointless to have
FOO_MANIFEST_SITE with an empty FOO_SOURCE in a package. A quick
grep session in the sources confirmed this assumption is indeed
true for the current code.
Signed-off-by: Luca Ceresoli <luca@lucaceresoli.net>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
If the BR2_GLOBAL_PATCH_DIR is misspelled in the .config or if somebody
renamed the existing directory with patches, buildroot happily builds
everything but without the patches. As this can lead to surprising
results, it is better to fail and give a user a message that
BR2_GLOBAL_PATCH_DIR is wrongly configured.
Signed-off-by: Michal Sojka <sojka@merica.cz>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
To easy up adding optional parameters when calling the
"apply-patches.sh" add and use the "APPLY_PATCHES" variable to execute
the script.
Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The manual gives this example for using the github macro:
FOO_VERSION = v1.0 # tag or full commit ID
Unfortunately, people copy/pasting this example will face weird make errors,
because it leads the FOO_VERSION variable to end with a space. Similar
problems can happen when testing a version bump or similar, so strip
leading/trailing spaces before the version is used to construct the build
directory path.
Reported-by: Edd Robbins <edd.robbins@gmail.com>
Cc: Edd Robbins <edd.robbins@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
[Thomas: fix duplicate -> duplicating.]
Reported-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Acked-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Suggested-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This patch introduces a check for the usage of deprecated variable
FOO_BAR_OPT, and suggests to use FOO_BAR_OPTS instead.
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This patch introduces a helper check-deprecated-variable, that throws an
error when a package defines a deprecated variable. This helps in detecting
transition errors when changing the buildroot infrastructures, and also
helps users that have created packages not part of mainline buildroot.
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Danomi Manchego <danomimanchego123@gmail.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998 at free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Since cset 54456cc (infra: consistently use double dollar signs...), it
is no longer possible to download versions with a slash in it, because
the _DL_VERSION variable is second-expanded, but the _VERSION variable
is immediately-expanded to have '/' substitued with '_'.
So, _DL_VERSION is only expanded at the time it is evaluated, and by
that time, we've lost the slashes in _VERSION.
Make the _DL_VERSION variables immediately-evaluated.
Fixes:
https://bugs.busybox.net/show_bug.cgi?id=7328
Reported-by: Vincent Stehlé <vincent.stehle@freescale.com>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Vincent Stehlé <vincent.stehle@freescale.com>
Acked-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Tested-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
In English, unlike in French, almost all usages of the word 'information'
are uncountable, meaning that 'informations' is invalid.
This patch fixes this typo throughout the tree, except in CHANGES and
docs/news.html (historic text).
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
In English, unlike in French, the 'h' in 'host' is pronounced, meaning that
the article should be 'a' instead of 'an'.
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
I would like to propose adding the site to the legal-info manifest
files. This gives a little more information on where the sources came
from without adding much overhead. Please note that is is only for
packages where the source is not local or set with OVERRIDE_SRCDIR.
Signed-off-by: Clayton Shotwell <clayton.shotwell@rockwellcollins.com>
Reviewed-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Because the configure stamp file depends on the rsync stamp file,
removing it by the "<pkg>-clean-for-rebuild" rule trigger the configure
step. To avoid triggering the configure step use an order-only dependency.
Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
Reported-by: Cédric Marie <cedric.marie@openmailbox.org>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Switch legal-info generation dependency to patch rather than extract
so we can use debian/copyright from debian-maintained packages,
because the *orig* file alone doesn't contain the appropiate file.
Signed-off-by: Gustavo Zacarias <gustavo@zacarias.com.ar>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
The init hooks may reference files installed by the package. For
example, when a package installs systemd unit files, and wants them
enabled/disabled by default, the init hook may move/symlink/rename the
unit files.
Currently, the init hooks are called before the package is installed,
thus breaking the use-case above.
Just inverse the install order, so that a package is installed before
its init hooks are called.
Fixes: http://autobuild.buildroot.net/results/5b03174318d8a6dc3112053ff141b70a90869c20/
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Eric Le Bihan <eric.le.bihan.dev@free.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
If the user sets a custom BR2_HOST_DIR, then STAGING_DIR is not under
BASE_DIR. The .pc/.la file fixup commands incorrectly made this assumption
and thus should be corrected.
Additionally, this patch:
- rewords the comment in pkg-autotools that explains the replacements.
- removes a for statement in favor of xargs
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
This will be needed to get the hash file, to check the
downloaded files.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Reviewed-by: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
pkg-generic and pkg-autotools both perform sed replacements of /usr to
$(STAGING_DIR)/usr, in .pc and .la files respectively. However, if buildroot
itself is located in /usr, then STAGING_DIR also starts with /usr, and naive
replacements of '/usr' can result in paths of the form:
/usr/buildroot/.../sysroot/usr/buildroot/.../sysroot/...
This patch makes the replacements more careful, by first replacing the known
paths BASE_DIR and STAGING_DIR to tokens, then performing the actual
replacement, and finally replacing the tokens with the corresponding paths.
Additionally it removes the unnecessary making of a backup copy in
pkg-autotools.
Partially fixes bug #5750 (https://bugs.busybox.net/show_bug.cgi?id=5750)
[Thomas: improve pkg-autotools.mk by including the explanation comment
proposed by Arnout.]
Partially-based-on-a-patch-from: Ben Stoltz <benstoltz@gmail.com>
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
This helper was called when none of the sources or license
files were saved.
Now we handle license files separately from the sources,
this is no longer the case: they are only called when the
sources are not saved.
Rename the handler and change the warning message accordingly.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Luca Ceresoli <luca@lucaceresoli.net>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Fabio Porcedda <fabio.porcedda@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
As the legal-info infra only (rightfully) saves the tarballs of packages
that:
- we want to redistribute,
- and are not local,
- and are not overriden,
add a comment stating so.
This should clarify the code-block, which although trivial to read,
was not easy to interpret without thinking thouroughly about it.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Luca Ceresoli <luca@lucaceresoli.net>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Fabio Porcedda <fabio.porcedda@gmail.com>
Reviewed-by: Luca Ceresoli <luca@lucaceresoli.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Even if we do not save the sources for local or overridden packages because
it is too complex, we can still quite easily save the license files.
Also, having the license files is a very important part of complying with
the licenses.
Move the copy of license files out of the non-local, non-overridden package
case, but still in the case where packages have a _SOURCE defined, to
avoid catching packages bundled in Buildroot (eg. mkpasswd et al.)
Reported-by: Luca Ceresoli <luca@lucaceresoli.net>
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Luca Ceresoli <luca@lucaceresoli.net>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Fabio Porcedda <fabio.porcedda@gmail.com>
Reviewed-by: Luca Ceresoli <luca@lucaceresoli.net>
Tested-by: Luca Ceresoli <luca@lucaceresoli.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Currently, if a package is marked _REDISTRIBUTE = NO, then legal-info
will not try to extract it first.
If that package also declares some _LICENSE_FILES, legal-info fails
if it is the only action we're trying to run:
$ cat defconfig
BR2_arm=y
BR2_TOOLCHAIN_BUILDROOT_EGLIBC=y
BR2_PACKAGE_LIBFSLCODEC=y
$ make BR2_DEFCONFIG=$(pwd)/defconfig defconfig
$ make legal-info
[--SNIP--]
cat: /home/ymorin/dev/buildroot/O/build/libfslcodec-3.5.7-1.0.0/EULA: No such file or directory
Fix this by always having legal-info extract the archives if one or
more _LICENSE_FILES are specified.
We do this for all types of packages: overridden, local or 'normal'
remote packages. Even though we do not save the sources for the
overridden or local packages, we need to save their licensing info,
so we need to extract them.
This implies that we now need only PKG-source, not PKG-extract anymore,
as a dependency of legal-info for packages we want to save (ie.
redistributable, non-local and non-overriden packages.)
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Luca Ceresoli <luca@lucaceresoli.net>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Fabio Porcedda <fabio.porcedda@gmail.com>
Reviewed-by: Luca Ceresoli <luca@lucaceresoli.net>
Tested-by: Luca Ceresoli <luca@lucaceresoli.net>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
To be able to check the "dot" command availability in
"<pkg>-graph-depends" move the check to the "graph-depends-requirements" rule.
Also don't use a subshell for the exit command to be sure that the error
will be returned by the shell.
Signed-off-by: Fabio Porcedda <fabio.porcedda@gmail.com>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Yann E. MORIN <yann.morin.1998@free.fr>
Tested-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
As the rules with respect to variable and function references and the need
for single or double dollar signs are not trivial, add a comment in
pkg-generic.mk describing them.
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Acked-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Reviewed-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
The inner-xxx-targets in the buildroot package infrastructures are
evaluated using $(eval) which causes variable references to be a bit
different than in regular make code. As we want most references to be
expanded only at the time of the $(eval) we should not use standard
references $(VAR) but rather use double dollar signs $$(VAR). This includes
function references like $(call), $(subst), etc. The only exception is the
reference to pkgdir/pkgname and numbered variables, which are parameters to
the inner block: $(1), $(2), etc.
This patch introduces consistent usage of double-dollar signs throughout the
different inner-xxx-targets blocks.
In some cases, this would potentially cause circular references, in
particular when the value of HOST_FOO_VAR would be obtained from the
corresponding FOO_VAR if HOST_FOO_VAR is not defined. In these cases, a test
is added to check for a host package (the only case where such constructions
are relevant; these are not circular).
Benefits of these changes are:
- behavior of variables is now again as expected. For example, setting
$(2)_VERSION = virtual in pkg-virtual.mk will effectively work, while
originally it would cause very odd results.
- The output of 'make printvars' is now much more useful. This target shows
the value of all variables, and the expression that led to that value.
However, if the expression was coming from an inner-xxx-targets block, and
was using single dollar signs, it would show in printvars as
VAR = value (value)
while if double dollar signs are used, it would effectively look like
VAR = value (actual expression)
as is intended.
This improvement is for example effective for FOO_DL_VERSION, FOO_RAWNAME,
FOO_SITE_METHOD and FOO_MAKE.
The correctness of this patch has been verified using 'make printvars',
'make manual' and 'make legal-info' before and after applying this patch,
and comparing the output.
Insight-provided-by: Arnout Vandecappelle <arnout@mind.be>
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Currently, we just use what a package declares as its dependencies.
But some packages may declare the same depdency more than once. For
example, php has two options to add SQL support: 'mysql' or 'mysqli',
which are not exclusive. So, php.mk has mysql twice as a dependency.
Although that does not cause any grievance for make, we end up generating
dependency graphs where this duplicate dependency is visible.
Add an intermediary variable which contains the $(sort)-ed list of the
dependencies, thus eliminating any duplicates.
This has the side effect of also sorting the list, which is probably
good for reproducibility anyway.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Kids nowaday seem to prefer a left-to-right drawing rather than the
more conventional and historical top-down drawing.
Rather than multiply the number of environment variables, just add
a single one where the user can pass arbitrary dot options, such as:
make BR2_GRAPH_DOT_OPTS=-Grankdir=LR graph-depends
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
In pkg-generic.mk, an entry would be added to each of the permissions,
devices and users tables, even if FOO_PERMISSIONS/DEVICES/USERS is empty. In
that last case, the entry would contain only the separator, which is
substituted to '\n' in fs/common.mk.
For configurations with many packages, this would render the build output a
bit odd, containing many \n instances (even though the end result in the
target would of course be the same).
This patch cleans up the build output by only adding to these tables when
the package actually specified contents for them.
Signed-off-by: Thomas De Schampheleire <thomas.de.schampheleire@gmail.com>
Tested-by: Romain Naour <romain.naour@openwide.fr>
Acked-by: Romain Naour <romain.naour@openwide.fr>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
Currently, it is possible that more than one provider of a virtual package
is selected in the menuconfig.
This leads to autobuild failures, and we do not protect the user from
making a mistake in the configuration. The failure is then hard to
troubleshoot in any case.
We can't use kconfig constructs to prevent this, since kconfig does not
tell how many options did a select on another option.
This change introduces a new variable a provider *must* define to include
all the virtual packages it is an implementation of. Then, when evaluating
the package's rules, we check that the provider is indeed the declared one
for each virtual package it claims to be an implementation of.
This works by taking advantage that when more than one provider is
selected, only one of them will 'win' in setting the _PROVIDES_FOO
option. Thus any provider just has to check it is indeed the declared
provider. If not, it means that one or more other provider is selected.
This gives the opportunity to the user to change its configuration, and
we can match the error message in the autobuilders to skip those failures
(we can skip them instead of reporting them, since they are obviously
configuration errors that should not happen in the first place.)
[Note: kudos to Arnout for suggesting this actual implementation. :-)]
Fixes:
http://autobuild.buildroot.org/results/285/2851069d6964aa46d26b4aabe7d84e8c0c6c72cehttp://autobuild.buildroot.net/results/9b7/9b7870354d70e27e42d3d9c1f131ab54706bf20e
[...]
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Peter Korsgaard <jacmet@uclibc.org>
Cc: Thomas De Schampheleire <patrickdepinguin@gmail.com>
Cc: Arnout Vandecappelle <arnout@mind.be>
Cc: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
There will soon be new options to the graph-depends script, which we
can only sanely pass via environment variables.
Currently, we use such an environment variable to pass the maximum depth
of the dependency graph; the name of that variable is explicit that it
contains just the depth.
However, there has been so far no release of Buildroot which would make
use of that variable, so no user should have come to rely on it.
Rename that variable so it is less specific, and more generic, so it can
be used to pass more options to graph-depends.
Signed-off-by: "Yann E. MORIN" <yann.morin.1998@free.fr>
Cc: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
Cc: Maxime Hadjinlian <maxime.hadjinlian@gmail.com>
Cc: Samuel Martin <s.martin49@gmail.com>
Signed-off-by: Peter Korsgaard <peter@korsgaard.com>
