| Age | Commit message (Collapse) | Author |
|
Sourcing a file with an absolute path and using rsource in it triggered
a relpath() between the absolute path and $srctree. Since e.g.
symlink/../bar/ = bar/ is not guaranteed for symlinks, this could lead
to the rsource'd file not being found if $srctree pointed to a symlink.
Switch to a simpler, more textual method for stripping $srctree from
glob results, which should be robust against symlink shenanigans. This
also makes the code a bit easier to follow.
Discovered by Marc Herbert.
Piggyback some minor cleanup.
|
|
Remove some 'Note:'s, update some stuff.
|
|
Before writing a configuration file or header file, compare the old
contents of the file against the new contents. If there's no change,
skip the write, to avoid updating the file modification time.
This might avoid triggering redundant rebuilds depending on how the
build system is set up, and could allow for a simpler setup.
|
|
Gives a better overview.
|
|
Gives *nix rename() semantics on Windows, with overwriting of the target
file.
Also limit the scope of the catch-all try-except.
|
|
|
|
Mirrors commit aff11cd983ec ("kconfig: Terminate menu blocks with a
comment in the generated config") in the kernel.
This makes the compatibility tests pass again, and is handy.
|
|
Flatten, and unscrunch the bool/tristate case.
|
|
Consistent with the other ones.
|
|
Not just for syntax errors when raised from Kconfig.__init__().
|
|
|
|
|
|
Pro: Looks cleaner and matches how definitions are usually written.
Con: Makes it harder to discover that 'bool "foo"' is just shorthand for
'bool' + 'prompt "foo"' (though it's documented).
The pros probably win.
|
|
These are constant symbols, which is why the quotes showed up.
Special-case them to remove the quotes, to match what people expect.
Internally, n/m/y without quotes is taken as a shorthand for
"n"/"m"/"y".
|
|
Commit e81a77b ("Consistently put direct deps. last when propagating")
makes the position of the direct deps. in property conditions
predictable after dependency propagation, making it easy to strip them
as needed.
Use this to implement MenuNode.orig_{prompt,defaults,selects,implies,ranges},
which work like the non-orig_* versions but omit the direct deps. Use
those in turn to omit the direct deps in Symbol/Choice.__str__().
The direct deps. (with propagated parent deps.) can still be seen after
'depends on ...', so there is no loss of information.
This unclutters Kconfig definitions shown in menuconfig/guiconfig and in
any generated documentation. The old output also had duplicated
dependencies, though it doesn't matter for evaluation.
Before:
config A
bool
prompt "A" if DEP
default y if FOO && DEP
depends on DEP
After:
config A
bool
prompt "A"
default y if FOO
depends on DEP
|
|
Preparation for some later additions. Previously, the 'visible if' deps
appeared to the right of the 'depends on' deps. Now, the direct deps
always appear last.
With this change, the prompt tuple is only updated once for any given
menu node too, which should be a tiny bit faster.
|
|
KconfigSyntaxError is just an older (bad) name. pydoc generates
confusing documentation due to the "KconfigSyntaxError = KconfigError"
assignment, so mention it in the docstring.
|
|
Clearer.
|
|
|
|
- Mention that having no direct dependencies corresponds to having y
direct dependencies
- Remove the note about how direct_dep is used internally. It's also
used for the select-with-unsatisfied-deps warning.
- Remove the note about how checking direct_dep might be "redundant"
since it's propagated. direct_dep can be useful in scripts still, and
maybe it could scare people away from it.
- Say 'surrounding menus and ifs' instead of of 'parent' in the 'dep'
docstring
- Fix two s/config/kconfig/ typos
|
|
- Point out why dependencies from surrounding if's and menus are
included
- Clarify that .referenced isn't transitive. Only "direct" references
are included.
- Give a hint about direct dependencies and expr_items()
Prompted by https://github.com/ulfalizer/Kconfiglib/issues/69.
|
|
It's undocumented and unguessable that Kconfig.top_node.filename/linenr
gets set to the location of the 'mainmenu' statement (if any).
Always set it to the first line of the top-level Kconfig instead, which
should be more expected.
|
|
This is a graphical configuration interface written in Tkinter. Like
menuconfig.py, it supports showing all symbols (with invisible symbols
in red) and jumping directly to symbols. Symbol values can also be
changed directly from the jump-to dialog.
When single-menu mode is enabled, a single menu is shown at a time, like
in the terminal menuconfig. Only this mode distinguishes between symbols
defined with 'config' and symbols defined with 'menuconfig'.
Compatible with both Python 2 and Python 3. Has been tested on X11,
Windows, and macOS.
To avoid having to carry around a bunch of GIFs, the image data is
embedded in guiconfig.py. To use separate GIF files instead, change
_USE_EMBEDDED_IMAGES to False. The image files can be found in
https://github.com/ulfalizer/Kconfiglib/tree/screenshots/guiconfig.
A new `make guiconfig` target in the Makefile patch will run the GUI
configuration interface on the Linux kernel.
|
|
In retrospect, trying to be "helpful" by saving the old version of a
$KCONFIG_CONFIG that does not start with a '.' as e.g. '.config.old'
instead of 'config.old' is a bad idea, because it means that scripts
can't rely on the backup file simply being called $KCONFIG_CONFIG.old.
I spotted this causing compatibility issues in
https://github.com/automate-lfs/jhalfs/commit/a645174fd43ba4eee84089965df85785878e7aa6.
I had Vim backup files and the like in mind originally, but .config.old
is much more likely to be processed by scripts.
This is a small backwards compatibility break, so the major version will
be increased to 11.
|
|
Temporarily changing the working directory is iffy if threads are
involved, since other code will see the change. Seems like a generally
rude thing for a random library function to do.
dir_fd (open() relative to directory) is another option, but it's Python
3 only, only available on the low-level os.open() interface, and might
not be universally supported.
|
|
Says the same thing the comment above it says.
|
|
These are clear in context without the module name, and it speeds up the
'source' code a tiny bit.
|
|
|
|
Print a warning suggesting to add quotes for things like
source foo/bar/Kconfig
menu title
prompt unquoted
Example warning:
Kconfig:32: warning: style: quotes recommended around 'lib/Kconfig.debug' in 'source lib/Kconfig.debug'
That quoteless syntax is supported for compatibility with old versions
of the C tools. It only works for a single word.
|
|
|
|
Make kconfiglib.py consistent with the utilities.
|
|
Have BOOL/TRISTATE/... match _T_BOOL/_T_TRISTATE/... so that _set_type()
doesn't have to do any conversion in the common case. Conversion is
still needed for _T_DEF_BOOL/_T_DEF_TRISTATE/..., but those are rare.
Also make use of UNKNOWN being falsy.
Piggyback some rearranging of global constants so that related stuff
appears closer together, for readability.
|
|
|
|
Saves around 10 ms of startup time when they aren't, as measured with
'-X importtime'.
|
|
|
|
Make the exception argument to _KconfigIOError() harder to miss.
|
|
The current error message talks a lot about $srctree, but $srctree is
seldom the culprit in practice. More common is
'source "$(SOME_ENV_VAR)/foo"`, where SOME_ENV_VAR hasn't been set.
Include the complete 'source ...' line for missing Kconfig files, and
mention unset environment variables as a hint. Only mention $srctree
briefly.
Also shorten the message when a .config can't be a found a bit. This
message would usually only be seen when working directly with the
library.
|
|
|
|
Use a set comprehension.
|
|
|
|
Came up in https://github.com/ulfalizer/Kconfiglib/issues/67.
|
|
|
|
Due to an oversight, '# CONFIG_FOO is not set' with FOO undefined only
triggered a warning about assigning an undefined symbol if FOO was never
referenced inside the Kconfig files.
|
|
Previously, menuconfig.py only prompted for saving the configuration if
.config didn't exist or the user changed symbol values within the
interface.
Also make it prompt for save if Kconfig symbols have been added,
removed, or have had their defaults changed, provided it would make the
saved .config differ from the loaded one.
This usually won't matter for correctness, because loading an outdated
configuration performs an implicit olddefconfig, but it's less
confusing.
Also add a Kconfig.missing_syms attribute that records all assignments
to undefined symbols in the most recently loaded .config file. This is
needed to implement the check for whether the saved .config would be
different.
As an unrelated change, always prompt for saving if a .config has been
loaded from within the menuconfig interface. The intention is probably
often to save the configuration somewhere else, even if it isn't
modified.
|
|
Also add quotes around the filename.
|
|
The functions must be defined before they can be put into 'functions'.
|
|
Say what was expected in the error message.
|
|
Gets rid of some lookups.
A bigger hotspot is all the _next_line() calls though. Wonder if there's
some nice restructuring to avoid them. Getting rid of the _reuse_tokens
flag would be nice too.
|
|
self._line is only used for error reporting, and the empty string
returned at EOF would never be shown. Move the assignment to after
readline() and use a local variable to hold the line instead, to get rid
of some property lookups.
|
|
More common. isspace() returns False for empty strings.
|