| Age | Commit message (Collapse) | Author |
|---|
|
getVarFlag() now defaults to expanding by default, thus remove the
True option from getVarFlag() calls with a regex search and
replace.
Search made with the following regex:
getVarFlag ?\(( ?[^,()]*, ?[^,()]*), True\)
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
getVar() now defaults to expanding by default, thus remove the True
option from getVar() calls with a regex search and replace.
Search made with the following regex: getVar ?\(( ?[^,()]*), True\)
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The old style bb.data.getVar/setVar API has long since been deprecated in
favour of d.getVar/setVar and friends.
Now we're about to change the default expansion parameter, drop the old APIs
to simplify the transition and ensure everyone is using the new style functions.
Conversion is trivial if there are remaining stragglers.
I've left bb.data.expand() for now since its more widely used but would make a good
follow up patch series.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Fix a bug where a totally wrong value of a variable would be exported if
an exception happened during d.getVar(). Also, print a warning if an
exception happends instead of silently ignoring it. It would probably be
best just to raise the exception, instead, but use the warning for now
in order to avoid breaking existing builds.
[YOCTO #10393]
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The value used for exclusion was always being expanded. This is actually
a bad idea since in most cases you'd want to exclude an unexpanded
value and makes it impossible to use the variable as intended.
This adjusts things so the value is not expanded and we can correctly
remove things from checksums much more easily.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
python deprecated logger.warn() in favour of logger.warning(). This is only
used in bitbake code so we may as well just translate everything to avoid
warnings under python 3. Its safe for python 2.7.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This avoids a common issue where PACKAGECONFIG is emitted as a function in
bitbake -e when the 'python' flag exists. It isn't a python function unless
both 'func' and 'python' are set. This aligns with the behavior of
emit_func_python.
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Right now, if you have some python code like:
X = "a"
def somefunction(d):
d.setVar("X", "b")
d.setVar("Y", "${X}")
then any sane person would expect that Y = "b" at the end of the
function. This is not the case, Y = "a".
This is due to the python function being expanded before execution, the
executed code would read d.setVar("Y", "a"). This understandably
confuses people, it also makes it near impossible to write ${} in a
python function without unintended things happening.
I think there is general agreement we should fix this and standardise
on non-expansion of python functions. We already don't expand anonymous
python (mostly).
I've checked OE-Core with buildhistory before and after this change and
there were a small number of issues this exposed which I've sent
patches for.
I propose we default to not expanding python code and then deal with
any consequences from that if/as/where identified. This will improve
new user understanding and usability of the system, it also allows
several long standing weird expansion issues to be fixed.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Expanding python functions for variable dependencies doesn't really make sense,
not least since this causes execution of any inline python, it also makes it
impossible to write expressions like d.expand("${X}") of d.setVar("X", "${Y}")
which may have the wrong values if expanded now.
This starts to standardise the approach across bitbake for handling python code.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This sets the scene for removing the default False for expansion from
getVarFlag. This would later allow True to become the expand default.
On the most part this is an automatic translation with:
sed -e 's:\(\.getVarFlag([^,()]*, [^,()]*\)):\1, False):g' -i `grep -ril getVar *`
There should be no functional change from this patch.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
A newline is always appended to the function body when it's written
out, so strip any trailing newlines which may be there already.
Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This codepath can be triggered by a python indentation error for example.
Showing it as an ExpansionError is misleading.
Change the code to add a warning about where the failure came from (in
particular giving the variable key name that triggered it) but raise the
proper exception.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
If parse_python() fails, the output is confusing. Passing in the extra
file/line data isn't expensive and improves readability significantly.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Calling record() for each override alteration is slow. Since we now expand
overrides dynamically we don't have to record the log data at each alteration,
we can instead print it directly from the existing data stores at variable
history print time using the exact same data stores.
This massively improves performance of the data store when parsing
with bitbake -e for example, it will improve memory overhead as well.
The only downside is that VariableHistory has to poke into the datastore
for some of its data but that seems an acceptable tradeoff rather than
double caching.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
We should use the d.xxxVar syntax rather than the older function style.
Also replace 0/1 with the more pythonic True/False.
No functionality changes.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Minor formatting improvement in generated shell scripts.
Add a newline after the last export emitted by emit_func. Also, remove
chance of putting an extra newline in the middle of the exports, since
we never want to do that.
Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Rather than just use d.getVar(X), use the more explict d.getVar(X, False)
since at some point in the future, having the default of expansion would
be nice. This is the first step towards that.
This patch was mostly made using the command:
sed -e 's:\(getVar([^,()]*\)\s*):\1, False):g' -i `grep -ril getVar *`
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
At least in theory, the order the keys are expanded in can make
a difference, particularly if there is key overlap.
We also want to ensure that any underlying base key is processed
before any overridden version of that variable (FOO before FOO_x)
which helps the update_data removal code I've been testing.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This was incorrect, it *must* be done next to the rename as a previous
may overlap with the current one and we need to detect this case.
I'll add a test case to better catch this problem in future.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This comparison is interesting even in the case of empty vales. Enabling
this warning actually found a bug in the metadata in avahi. Make the
code handle None specifically and also remove the dead code path in the
second if statement.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
By doing this we can take advantage of the expansion cache before
starting write operations on the data store.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Running bitbake inside make results in the exported environment variable
MAKEOVERRIDES="${-*-command-variables-*-}", which the shell chokes on
when trying to expand it. But of course, it probably shouldn't have been
trying to expand it in the first place -- so just escape the dollar
sign.
Signed-off-by: Richard Tollerton <rich.tollerton@ni.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The shellshock patches changed the way bash functions are exported.
Unfortunately different distros used slightly different formats,
Fedora went with BASH_FUNC_XXX()=() { echo foo; } and Ubuntu went with
BASH_FUNC_foo%%=() { echo foo; }.
The former causes errors in dealing with out output from emit_env,
the functions are not exported in either case any more.
This patch handles things so the functions work as expected in either
case.
[YOCTO #6880]
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Currently the output in the python task/function run files is rather
incomplete and effectively useless. This enhances the code to take
advantage of the bitbake's dependency tracking and extend the output to
include dependencies. This makes the files more usable for debugging
purposes. Since this only happens at python function execution time, the
overhead is minimal in the grand scheme of things.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
On rare occasions it's useful to be able to exclude a part of a
variable's value from the variable's signature; for example if you want
to add an item to a list sometimes and not have the signature of the
variable change depending on whether the item is in the list or not. The
initial intended use case for this in OpenEmbedded is to allow adding a
function to SSTATEPOSTINSTFUNCS in buildhistory.bbclass and not have
that change any task signatures (so adding and removing
INHERIT += "buildhistory" won't lead to any rebuilds).
Part of the fix for [YOCTO #5897].
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
pre/postfuncs were not being added to checksums. This meant that when reconfiguration
occurred, tasks were not always being rerun when they should. This include
sstate functions as well as systemd's do_install function in the OE metadata.
With the addition of postfuncs, its possible a shell task can have a python
pre/postfunc so we have to guard against this when generating shell output
in emit_func.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
If we don't sort the data, the values can reorder changing the
signatures meaning we get confused builds and significant cache
misses.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
One of the current frustrations with the sstate checksums is that
code like base_contains('X', 'y',...) adds a full dependency on X
and varies depend even on whitespace changes in X.
This patch adds special handling of the contains functions to expand
the first parameter and check for the flag specified by the second
parameter (assuming its a string).
The result is then appended to the value of the variable with a "Set"
or "Unset" status. If the flag is added/removed, the stored variable
value changes and hence the checksum changes. No dependency on X
is added so it is free to change with regard to other flags or
whitespace.
This code is far from ideal, ideally we'd have properly typed variables
however it fixes a major annoyance of the current checksums and
is of enough value its worth adding in a stopgap solution. It shouldn't
significantly restrict any propely typed variable implementation in
future.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
VAL = "" (not shown)
VAL = " " (shown as "")
VAL = " x" (shown as "x")
would all show up rather differently to what would be expected in the
bitbake -e output. This fixes things so they appear consistently.
The output for running some shell functions may also change slightly
but shouldn't change in a way that is likely to cause problems.
[YOCTO #5507]
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Move the variable lookup to the outer loop for performance, replacing
a now unneeded parameter (after the previous changes).
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
When looking up flag variable dependencies, large chunks of the function
aren't needed. Optimise the function flow accordingly for speed.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Instead of multiple calls to getVarFlag, make one call to getVarFlags, only expanding
the flags that need to be expanded. This improves performance.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Compute a cache of the list of potential export variables so
that we don't have to compute the list from scratch.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Profiling shows the creation of keys() has overhead and we're better using
an iterator rather than the memory associated with the huge list of keys
when iterating the whoe datastore. We minimise the number of times
we do this to twice only per recipe.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The if statement current causes the size of parent to be calcuated which
is like a len() operation on a datastore. Since we're only interested
whether the value is none, checking explictly for this gives a
small performance gain.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
It was necessary to append ${TOPDIR}/recipes/images to BBFILES.
Implemented the mechanism to append a value to a variable: a command and
the method in cooker.
[YOCTO #4193]
Signed-off-by: Cristiana Voicu <cristiana.voicu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Currently we account of the top level function's vardeps but not
those of any subfunction. This would imply we'd have to manually
write the dependencies of all parent functions which would be crazy.
This patch adds the dependencies to fix the issue.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
BitBake commit 7c568132c54a21161de28907159f902462f1e2bb resulted in a
fairly serious performance regression during parsing, almost doubling
the time taken to do a full parse and almost certainly impacting
performance during building. The expandKeys function is called
frequently, and if we avoid using keys() and instead just use the normal
variable lookup mechanism, performance is restored.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
When two variables are defined as:
${var} = "bar"
foo = "foobar"
The value of 'foo' when ${var} == foo becomes indeterminate. We
want to warn a user when this situation has been encountered so they
can take corrective actions.
In the above example usually foo == bar, unless multilibs are enabled.
Then ml-foo = "ml-foobar".
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This patch adds tracking of the history of variable assignments.
The changes are predominantly localized to data_smart.py and
parse/ast.py. cooker.py and data.py are altered to display the
recorded data, and turn tracking on for the bitbake -e case.
The data.py update_data() function warns DataSmart.finalize()
to report the caller one further back up the tree.
In general, d.setVar() does what it used to do. Optionally,
arguments describing an operation may be appended; if none
are present, the operation is implicitly ignored. If it's
not ignored, it will attempt to infer missing information
(name of variable, value assigned, file and line) by examining
the traceback. This slightly elaborate process eliminates a
category of problems in which the 'var' member of the keyword
arguments dict is set, and a positional argument corresponding
to 'var' is also set. It also makes calling much simpler for
the common cases.
The resulting output gives you a pretty good picture of what
values got set, and how they got set.
RP Modifications:
a) Split from IncludeHistory to separate VariableHistory
b) Add dedicated copy function instead of deepcopy
c) Use COW for variables dict
d) Remove 'value' loginfo value and just use 'details'
e) Desensitise code for calling order (set 'op' before/after
infer_caller_details was error prone)
f) Fix bug where ?= "" wasn't shown correctly
g) Log more set operations as some variables mysteriously acquired
values previously
h) Standardise infer_caller_details to be triggered from .record()
where at all possible to reduce overhead in non-enabled cases
i) Rename variable parameter names to match inference code
j) Add VariableHistory emit() function to match IncludeHistory
k) Fix handling of appendVar, prependVar and matching flag ops
l) Use ignored=True to stop logging further events where appropriate
Signed-off-by: Peter Seebach <peter.seebach@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Currently, calling a shell function using exec_func may result in
dependent functions being missing, if the shell parser can't detect
the dependency, even if it was specified with the vardeps flag.
This patch ensures the function looks at the flag and considers it
when deciding which other functions need to be output.
[YOCTO #3561]
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The inherits cache contains duplicate entries, some with the full patch, some
just starting classes/xxx. This is a waste of parse time and potentially
error prone. This patch fixes various pieces of code so the absolute paths are
always preferred and work correctly. The inherits_class function did not work
with full paths so the patch fixes this.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Instead of logging the function/variable separately as a NOTE when
failing to expand, re-raise ExpansionError with more contextual
information. This means that the full details are reported in Hob as
well as actually reporting the original error message in any UI where
we previously did not. For example, we used to get this with tab/space
indentation issues in a python function:
NOTE: Error expanding variable populate_packages
ERROR: Unable to parse /path/to/recipename.bb
Now, we will get this:
ERROR: ExpansionError during parsing /path/to/recipename.bb: Failure
expanding variable populate_packages: IndentationError: unindent does
not match any outer indentation level (<string>, line 4)
Fixes [YOCTO #3162].
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Mixed spacing in python functions can cause subtle issues and
generally confuses users. We've standardised on 4 space indentation,
adding this warning helps ensure consistency and avoid bugs. It
also makes _prepend and _append operations on python functions slightly
less risky.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Add a dependency to each variable on a filtered list of its varflags.
This is intended to catch things such as SRC_URI checksums, varflags
controlling extra functionality from classes (e.g. the recently updated
update-alternatives class in OE-Core), etc. and ensure their values
influence the sstate checksums.
There is an exclusion list which needs to be set via bitbake.conf
(BB_SIGNATURE_EXCLUDE_FLAGS), if this is not set then the functionality
is disabled. The existing vardepsexclude mechanism can also be used to
exclude undesired varflags, but they must be fully specified, e.g.:
do_patch[vardepsexclude] += "do_patch[someflag]"
Implements [YOCTO #2517].
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
(From Poky rev: affd0907253387cee0ba0c7e9f92bcc109895446)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
Joshua Lock
Richard Purdie
Markus Lehtonen
Christopher Larson
Andre McCurdy
Richard Tollerton
Robert P. J. Day
Paul Eggleton
Cristiana Voicu
Mark Hatle
Peter Seebach