The current logic was storing symbols source code on a list,
not linked to the actual KdocItem. While this works fine when
kernel-doc markups are OK, on places where there is a "/**"
without a valid kernel-doc markup, it ends that the 1:1 match
between source code and KdocItem doesn't happen, causing
problems to generate the YAML output.
Fix it by storing the source code directly into the KdocItem
structure.
This shouldn't affect performance or memory footprint, except
when --yaml option is used.
While here, add a __repr__() function for KdocItem, as it
helps debugging it.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <77902dafabb5c3250486aa2dc1568d5fafa95c5b.1774256269.git.mchehab+huawei@kernel.org>
Fix check for simple table delimiters.
ReST simple tables use "=" instead of "-". I ended testing it with
a table modified from a complex one, using "--- --- ---", instead
of searching for a real Kernel example.
Only noticed when adding an unit test and seek for an actual
example from kernel-doc markups.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <dea95337c05040f95e5a95ae41d69ddef0aaa8d6.1774256269.git.mchehab+huawei@kernel.org>
Currently, there are 15 occurrences of section?_start_lines,
with 10 using the plural way.
This is an issue, as, while kdoc_output works with KdocItem,
the term doesn't match its init value.
The variable sections_start_lines stores multiple sections,
so placing it in plural is its correct way.
So, ensure that, on all parts of kdoc, this will be referred
as sections_start_lines.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <d1e0f1d3f80df41c11a1bbde6a12fd9468bc3813.1773823995.git.mchehab+huawei@kernel.org>
When reading the contents on a KdocItem using YAML, the data
will be imported into a dict.
Add a method to create a new KdocItem from a dict to allow
converting such input into a real KdocItem.
While here, address an issue that, if the class is initialized
with an internal parameter outside the 4 initial arguments,
it would end being added inside other_stuff, which breaks
initializing it from a dict.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <fafeac23d1577927e1a3c32cddfbec1e0209ac73.1773823995.git.mchehab+huawei@kernel.org>
When writing unittests for kdoc_output, it became clear that
the logic with handles a series of KdocItem symbols from
a single file belons to kdoc_output, and not to kdoc_files.
Move the code to it.
While here, also ensure that self.config will be placed
together with set.out_style.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <4ebc26e37a0b544c50d50b8077760f147fa6a535.1773823995.git.mchehab+huawei@kernel.org>
The Sphinx output from autodoc doesn't automatically break long
lines, except on spaces.
Change KernRe __repr__() to break the pattern on multiple strings,
each one with a maximum limit of 60 characters.
With that, documentation output for KernRe should now be displayable,
even on long strings.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <60c264a9d277fed655b1a62df2195562c8596090.1773823995.git.mchehab+huawei@kernel.org>
Most of the rules inside CTransforms are of the type CMatch.
Don't re-parse the source code every time.
Doing this doesn't change the output, but makes kdoc almost
as fast as before the tokenizer patches:
# Before tokenizer patches
$ time ./scripts/kernel-doc . -man >original 2>&1
real 0m42.933s
user 0m36.523s
sys 0m1.145s
# After tokenizer patches
$ time ./scripts/kernel-doc . -man >before 2>&1
real 1m29.853s
user 1m23.974s
sys 0m1.237s
# After this patch
$ time ./scripts/kernel-doc . -man >after 2>&1
real 0m48.579s
user 0m45.938s
sys 0m0.988s
$ diff -s before after
Files before and after are identical
Manually checked the differences between original and after
with:
$ diff -U0 -prBw original after|grep -v Warning|grep -v "@@"|less
They're due:
- whitespace fixes;
- struct_group are now better handled;
- several badly-generated man pages from broken inline kernel-doc
markups are now fixed.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <1cc2a4286ebf7d4b2d03fcaf42a1ba9fa09004b9.1773770483.git.mchehab+huawei@kernel.org>
Changeset 2b957decdb6c ("docs: kdoc: don't add broken comments inside prototypes")
revealed a hidden bug at split_struct_proto(): some comments there may break
its capability of properly identifying a struct.
Fixing it is as simple as stripping comments before calling it.
Fixes: 2b957decdb6c ("docs: kdoc: don't add broken comments inside prototypes")
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <dcff37b6da5329aea415de31f543b6a1c2cbbbce.1773770483.git.mchehab+huawei@kernel.org>
The previous approach were to unwind nested structs/unions.
Now that we have a logic that can handle it well, use it to
ensure that struct_group macros will properly reflect the
actual struct.
Note that the replacemend logic still simplifies the code
a little bit, as the basic build block for struct group is:
union { \
struct { MEMBERS } ATTRS; \
struct __struct_group_tag(TAG) { MEMBERS } ATTRS NAME; \
} ATTRS
There:
- ATTRS is meant to add extra macro attributes like __packed
which we already discard, as they aren't relevant to
document struct members;
- TAG is used only when built with __cplusplus.
So, instead, convert them into just:
struct { MEMBERS };
Please notice that here, we're using the greedy version of the
backrefs, as MEMBERS is actually MEMBERS... on all such macros.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <24bf2c036b08814d9b4aabc27542fd3b2ff54424.1773770483.git.mchehab+huawei@kernel.org>
The NextMatch code is complex, and will become even more complex
if we add there support for arguments.
Now that we have a tokenizer, we can use a better solution,
easier to be understood.
Yet, to improve performance, it is better to make it use a
previously tokenized code, changing its ABI.
So, reimplement NextMatch using the CTokener class. Once it
is done, we can drop NestedMatch.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <fa818ea164216b17520b588e3f12b81499b76dd7.1773770483.git.mchehab+huawei@kernel.org>
Better handle comments inside structs. After those changes,
all unittests now pass:
test_private:
TestPublicPrivate:
test balanced_inner_private: OK
test balanced_non_greddy_private: OK
test balanced_private: OK
test no private: OK
test unbalanced_inner_private: OK
test unbalanced_private: OK
test unbalanced_struct_group_tagged_with_private: OK
test unbalanced_two_struct_group_tagged_first_with_private: OK
test unbalanced_without_end_of_line: OK
Ran 9 tests
This also solves a bug when handling STRUCT_GROUP() with a private
comment on it:
@@ -397134,7 +397134,7 @@ basic V4L2 device-level support.
unsigned int max_len;
unsigned int offset;
struct page_pool_params_slow slow;
- STRUCT_GROUP( struct net_device *netdev;
+ struct net_device *netdev;
unsigned int queue_idx;
unsigned int flags;
};
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Message-ID: <f83ee9e8c38407eaab6ad10d4ccf155fb36683cc.1773074166.git.mchehab+huawei@kernel.org>
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <054763260f7b5459ad0738ed906d7c358d640692.1773770483.git.mchehab+huawei@kernel.org>
Parsing a file like drivers/scsi/isci/host.h, which contains
broken kernel-doc markups makes it create a prototype that contains
unmatched end comments.
That causes, for instance, struct sci_power_control to be shown this
this prototype:
struct sci_power_control {
* it is not. */ bool timer_started;
*/ struct sci_timer timer;
* requesters field. */ u8 phys_waiting;
*/ u8 phys_granted_power;
* mapped into requesters via struct sci_phy.phy_index */ struct isci_phy *requesters[SCI_MAX_PHYS];
};
as comments won't start with "/*" anymore.
Fix the logic to detect such cases, and keep adding the comments
inside it.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Message-ID: <18e577dbbd538dcc22945ff139fe3638344e14f0.1773074166.git.mchehab+huawei@kernel.org>
Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <12ac4a97e2bd5a19d6537122c10098690c38d2c7.1773770483.git.mchehab+huawei@kernel.org>
While the main goal for kernel-doc is to be used inside the Linux
Kernel, other open source projects could benefit for it. That's
currently the case of QEMU, which has a fork, mainly due to two
reasons:
- they need an extra C function transform rule;
- they handle the html output a little bit different.
Add an extra optional argument to make easier for the code to be
shared, as, with that, QEMU can just create a new derivated class
that will contain its specific rulesets, and just copy the
remaining kernel-doc files as-is.
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
Message-ID: <6b274ddbdcd9d438c6848e00e410a2f65ef80ec2.1772810574.git.mchehab+huawei@kernel.org>
