CVS log for src/usr.bin/xlint/xlint/xlint.c

Up to [cvs.NetBSD.org] / src / usr.bin / xlint / xlint

Request diff between arbitrary revisions

Keyword substitution: kv
Default branch: MAIN

lint: warn about do-while macros that end with a semicolon

lint: add wrapper for <ctype.h> functions, for strict bool mode

When using the Clang preprocessor (with MKLLVM=yes), the preprocessor
output does not indicate which tokens come from a system header and
which tokens come from the user code.  Lint's strict bool mode relies on
this information to treat the character classification functions from
<ctype.h> as if their return type were bool instead of int.

These wrapper functions are only used when their argument is indeed a
'char', but not when the argument might be 'EOF or representable as an
unsigned char', such as when reading a byte from the input.

lint: remove unused flag '-u' from lint1

This flag is completely handled by lint2.

lint: remove custom wrappers around <ctype.h> functions

lint: clean up

lint: clean up comments

No functional change.

lint: in declarations, replace tab with space

Previously, in some cases, the keywords 'const' or 'struct' were
followed by a tab, which doesn't make sense.

No functional change.

lint: re-wrap comments

No functional change.

lint: indent statement continuations consistently

No binary change.

lint: reduce number of negations, fix variable name

No functional change.

lint: merge redundant variables

No functional change.

lint: make variable names more expressive, remove redundant comments

No functional change.

lint: indent copyright lines consistently

lint: sync usage messages with reality

lint: allow to keep the preprocessor output on success as well

lint: sort functions in calling order

No functional change.

lint: indent local variables consistently

No binary change.

lint: make basename simpler

There is no need to handle trailing slashes since lint only handles
regular files in diagnostics, not directories. Furthermore, only the
last '/' was ignored, but multiple trailing slashes would not.

lint: repurpose the '-d' option to specify DESTDIR

Previously, passing '-nostdinc -isystem $dir' only searched the given
directory but not any compiler-specific directories.

Discovered by fontconfig, which includes <stdatomic.h> from C11, which
lives in /usr/include/gcc-10 instead of /usr/include.

Change the preprocessor options to '--sysroot' instead, to align them
with how the compiler is invoked using build.sh.

lint: catch more write errors when copying files

lint: remove unnecessary abstraction layer for passing arguments

The abstraction layer of the pass_to functions wasn't worth its weight.
After inlining the functions, the code is even clearer than before.

No functional change.

lint: keep output from cpp if cpp fails

Previously, lint only kept the cpp output if lint1 failed.  Due to
subtle differences in the preprocessing, there can be cases where the
preprocessor already fails, for example when it takes a different
'#ifdef' branch and therefore tries to include nonexistent files.  In
such a case, preserving the output of the preprocessor is helpful to
compare it with the regular output of cpp, so keep it.

xlint: fix null pointer dereference for lint -V (since today)

Building the argument lists further away from the vfork call in xlint.c
1.100 had the side effect that the trailing null pointer was added
outside run_child.

lint: clean up main function

No functional change.

lint: extract run_cpp and run_lint1 to separate functions

Move run_lint2 further to the top, to save a forward declaration.

No functional change.

lint: reduce number of local variables in lbasename

No functional change.

lint: move argument list modification further away from vfork

Just to be on the safe side.

While here, don't allocate argument strings needlessly.

No functional change.

lint: turn O(n^2) to O(n) for list of arguments to lint child processes

Previously, adding an argument to the lint child processes (cpp, lint1,
lint2) each time searched the list of arguments for the terminating
null pointer and then reallocated the memory for storing the strings.

Replace the above with a standard resizable array implementation and
give it a proper name, to avoid 'char ***' in the code.

The terminating null pointer in the lists is only required when passing
the list to execvp.  In all other cases it's not needed, so drop it.

No functional change.

lint: rename local functions to be more readable

No functional change.

lint: explicitly ignore errors when printing status messages

No binary change.

lint: remove unnecessary MBLKSIZ, use stack buffer in xlint

No functional change.

lint: remove redundant cast

Since C90 prototypes, integer arguments are converted to the correct
parameter type.

No functional change.

lint: add additional queries that are not enabled by default

In the last 18 months, several lint warnings have been made adjusted to
allow common usage patterns.  For example, lint no longer warns about a
constant condition in the statement 'do { ... } while (false)' (message
161), as this pattern is well-known in statement-like macros, making it
unlikely that the 'false' is a mistake.  Another example is casts
between unequal pointer types (message 247) for a few well-known
patterns that are unlikely to be bugs.

Occasionally, it is useful to query the code for patterns or events that
would not justify a warning.  These patterns are modeled as predefined
queries that can be selected individually, in addition to and
independently of the existing warnings and errors.

New queries can be added as needed, in the same way as new warnings.
Queries that are deemed no longer used can be deactivated in the same
way as warnings that are no longer used.

As long as none of the queries is enabled, they produce a minimal
overhead of querying a single global variable.  Computations that are
more expensive than a few machine instructions should be guarded by
any_query_enabled.

https://mail-index.netbsd.org/source-changes-d/2022/06/28/msg013716.html

ok christos@

lint: clean up usage messages

The two synopsis forms differed in the spelling of 'file ...'.

The options string for getopt does not start with ':', which led to a
duplicate message 'unknown option -- ?' followed by 'Unknown flag ?'.

Be more specific when calling 'lint file.c -u'; the message 'Unknown
argument' was not helpful as it didn't pinpoint that there are two
different phases for parsing options.  In the second phase, only the
options '-L' and '-l' are recognized.

In the manual page, mention the difference between the two synopsis
forms as early as possible.  The two synopsis forms are very similar and
both have far to many options to see the difference at a glance.

lint: use __RCSID in lint mode as well

Since 1995-10-02, lint supports __asm statements and __asm modifiers.

No binary change.

lint: skip duplicate -llib and -Llibdir options

This removes many of the currently 26.000 'multiply defined' lines from
a typical NetBSD build.  Duplicate libraries happen a lot due to
overlapping transitive dependencies.  An example is usr.bin/telnet,
which links to -lcrypto and several other libraries twice.

lint: in verbose mode, quote shell commands in output

lint: fix typo in comment

No binary change.

lint: do not output "Lint pass2:"

A colon at the end of a line requires at least 1 follow-up line, but
xlint cannot know whether lint2 will find anything to complain about.
Having a colon followed by nothing creates unnecessary confusion.

Also accept -W to pass to cpp

remove extra line

Explain the usage errors.

lint: rename list_add to list_add_ref and list_add_copy to list_add

The function for copying the string is used more often, so use the
simpler name for it.

No functional change.

lint: remove redundant function prototypes

lint: hide irrelevant type information from lint2

No functional change.

lint: fix some more lint warnings

The only remaining warnings are in scan.c, which is a generated Flex
scanner.

No functional change.

lint: explicitly ignore the return value of close

mkdep: make argument of findcc const

Previously, findcc modified its argument string, even though it had been
declared as 'const char *'.  This triggered a lint warning that "strchr
effectively discards 'const char *' from argument", in fact, this code
caused the lint check to be implemented in the first place.

The first attempt at fixing it by removing the 'const' from the
parameter type was a bad idea since it made the API of that function
more complicated.

Revert back to making the parameter a 'const char *' and duplicate that
string internally as necessary.  Add a few more tests for absolute
pathnames since these had been missing before.  There are no tests yet
for snprintf with too long strings, but the current change does not
modify that part of the code.

mkdep: fix prototype of findcc

A function that modifies a string argument must not declare that
argument as 'const char *', even if all callers (mkdep and lint) always
pass it a modifiable string.

No functional change.

lint: remove redundant variable Bflag

No functional change.

lint: remove unused static variable

lint: simplify code for adding '-l' and '-L' to libraries

No functional change.

lint: merge duplicate code for passing flags to subprocesses

No functional change.

lint: uniformly pass command line flags to subprocesses

No functional change.

lint: remove commented-out preprocessor options

These have been commented out since 2002.

lint: group global variables by the subprocess

No functional change.

lint: rename list_free to list_clear

The previous name suggested that the function would free the list
itself, which it doesn't.

lint: remove dead code for reading from stdin

Since xlint.c 1.27 from 2002-01-31, it has not been possible to read
from stdin anymore.  Apparently nobody missed this feature in the last
19 years.

lint: inline list_add_defines

No functional change.

lint: rename functions for handling string lists

No functional change.

lint: make data flow in findlibs simpler

No functional change.

lint: make memory management in xlint simpler

No functional change.

lint: make xasprintf simpler

No functional change.

lint: on request, keep the preprocessor output

Previously, the simplest way of getting the preprocessed translation
unit was to run lint with the additional flag -V, which outputs the
command line of the C preprocessor, among other things.  That command
line does not include the proper quoting though, so it cannot be used
verbatim as a shell command if the command line contains spaces such as
in -Du64="unsigned long long".

In the common situation where lint is run via a Makefile, the option -V
had to be added in the Makefile itself since there is no make variable
for additional user-settable lint flags.  This is not straight-forward
enough.

Adding another command line option for this purpose would reduce the
remaining namespace for options.  Most of the 52 letters are already
used up.

To make this situation as simple as possible, preserve the output of the
C preprocessor depending on an environment variable.

lint: remove WARNS=3, falling back to the default WARNS=5

It's strange that GCC does not warn about the nonliteral format strings
in lint1/err.c, lint2/msg.c and lint2/read.c, despite -Wformat=2, but
Clang does.

lint: un-abbreviate code for passing options to cpp, lint1 and lint2

No functional change.

lint: add option to accept C11 features

The list of available letters for the command line options gets shorter
and shorter.  Most of the interesting letters are already used for some
warning categories.  Curiously, -A, -W and -E were all still available.

The option -A nicely matches the intention of the option, which is to
allow a certain set of language features.  To keep the option available
for further extensions, define -Ac11 as the currently only valid option
of that kind.  This allows straight-forward extension for C17 and future
language standards, as well as independent feature-sets.  The options -W
and -E may someday complement the -A option, using the allow/warn/error
categories.

lint: arrange options in the usage messages

* lowercase the word 'usage'
* group the preprocessor flags on a line of their own
* add missing closing ']' for '-X'
* properly indent the lines

lint: add parentheses after sizeof, as required by share/misc/style

No functional change.

lint: in malloc calls, use 'sizeof *ptr' instead of 'sizeof(type)'

No functional change.

lint: replace integer constant expressions with true and false

LINTFLAGS=-gST make lint, with manual review.

The error messages from lint are all correct, they are not complete
though.  The return value of a function returning bool may still be
compared to the integer 0.

lint: in strict bool mode, integer constants do not have type bool

Previously, lint1 allowed integer constants such as 0 and 1 to be used
as bool constants.  This was only half-baked since after fixing all
error messages from that strict mode, there may still be integer
literals in the code that should be replaced with true or false.  This
would stop a migration from int to bool in the middle, leaving
inconsistent code around.

To find the remaining type inconsistencies, treat integers and bool as
completely incompatible, even for compile time constants.

lint: replace 0 and 1 with false and true, where appropriate

Change in behavior: Passing the option -h exactly 4294967296 times or
any multiple thereof is no longer equivalent to passing it never at all,
it is now equivalent to passing it once.  See main2.c, hflag++ for the
actual change.

Other than that, no functional change intended.

A very large portion of the code already conformed to the requirements
of the strict bool mode.  The only missing thing was using the constant
literals false and true instead of 0 and 1.  For sure there are some
integer literals left that can be converted.  For now, all literals that
appeared in the form " = 0" or " = 1" have been replaced.

lint: add new check for strict bool mode

In strict bool mode, bool is considered incompatible with all other
scalar types, just as in Java, C#, Pascal.

The controlling expressions in if statements, while loops, for loops and
the '?:' operator must be of type bool.  The logical operators work on
bool instead of int, the bitwise operators accept both integer and bool.
The arithmetic operators don't accept bool.

Since <stdbool.h> implements bool using C preprocessor macros instead of
predefining the identifiers "true" and "false", the integer constants 0
and 1 may be used in all contexts that require a bool expression.
Except from these, no implicit conversion between bool and scalar types
is allowed.

See usr.bin/tests/xlint/lint1/d_c99_bool_strict.c for more details.

The command line option -T has been chosen because all obvious choices
(-b or -B for bool, -s or -S for strict) are already in use.  The -T may
stand for "types are checked strictly".

The default behavior of lint doesn't change.  The strict bool check is
purely optional.

An example program for strict bool mode is usr.bin/make, which has been
using explicit comparisons such as p != NULL, ch != '\0' or n > 0 in
most places for a long time now, even before the refactoring in 2020.

lint: fix typos and other minor stylistic issues

lint: remove redundant parentheses around return value

Use -Z to pass arguments to ccp(1) directly. Can be used as:
    -Z-include -Zfile.h

Merge changes from current as of 20200406

use asprintf().

usr.bin/xlint: Fix -Werror=format-overflow= error.

Replace sprintf(3) with snprintf(3).

Error was reported when build.sh was run with MKLIBCSANITIZER=yes flag.

Reviewed by: kamil@

Sync with HEAD

Undefine SSE stuff; we don't want portable code to depend on it and we
don't provide the necessary headers.

Sync with HEAD.  (Note that most of these changes are simply $NetBSD$
tag issues.)

Add -R (source filename remapping) for MKREPRO

Drop main() prototype.

Fix lseek(2) swapped arguments.

convert -idirafter to -isystem, as recommended by cgd in PR 11843.
most of the list in that PR were already done.

Sync with HEAD.

Third (and last) commit. See http://mail-index.netbsd.org/source-changes/2009/05/13/msg221222.html

Lukemify (WARNS=4)

fix -Wcast-qual issues

Sync with wrstuden-revivesa-base-2.

Undefine __PCC__ the same as __GNUC__ inside lint.

enable -P

sync with HEAD

Convert to using raise_default_signal(3).

appease gcc -Wuninitialized

Only pass -Wtraditional to cpp for lint -t ...
Stops warnings about pre-processor constructs like #elif - which there is
no point detecting now that we've changed much of the code to require an
ANSI C compiler.

Apply patch (requested by dbj in ticket 1724):
Fix problems when building the netbsd-1-6 branch with GCC 3.3.3.

Pull up revision 1.35 (requested by jmc in ticket #527):
Completely rework how tools/compat is done. Purge all uses/references to
_NETBSD_SOURCE as this makes cross building from older/newer versions of
NetBSD harder, not easier (and also makes the resulting tools 'different')
Wrap all required code with the inclusion of nbtool_config.h, attempt to
only use POSIX code in all places (or when reasonable test w. configure and
provide definitions: ala u_int, etc).
Reviewed by lukem. Tested on FreeBSD 4.9, Redhat Linux ES3, NetBSD 1.6.2 x86
NetBSD current (x86 and amd64) and Solaris 9.
Fixes PR's: PR#17762 PR#25944

Completely rework how tools/compat is done. Purge all uses/references to
_NETBSD_SOURCE as this makes cross building from older/newer versions of
NetBSD harder, not easier (and also makes the resulting tools 'different')

Wrap all required code with the inclusion of nbtool_config.h, attempt to
only use POSIX code in all places (or when reasonable test w. configure and
provide definitions: ala u_int, etc).

Reviewed by lukem. Tested on FreeBSD 4.9, Redhat Linux ES3, NetBSD 1.6.2 x86
NetBSD current (x86 and amd64) and Solaris 9.

Fixes PR's: PR#17762 PR#25944

Pass -MD through to cpp so that it can generate a depend file.
(Actually pass any -M <arg> through, but -MD is most useful.)

Overhaul how `build.sh tools' are used:

    *	Rename "config.h" to "nbtool_config.h" and
	HAVE_CONFIG_H to HAVE_NBTOOL_CONFIG_H.
	This makes in more obvious in the source when we're using
	tools/compat/config.h versus "standard autoconf" config.h

    *	Consistently move the inclusion of nbtool_config.h to before
    	<sys/cdefs.h> so that the former can provide __RCSID() (et al),
    	and there's no need to protect those macros any more.

These changes should make it easier to "tool-ify" a program by adding:
	#if HAVE_NBTOOL_CONFIG_H
	#include "nbtool_config.h"
	#endif
to the top of the source files (for the general case).

synchronous, not syncronous.

Skip passing -$ to the preprocessor, for now.

add -S flag for c99 support.

sweep of errx/warnx, remove unnecessary trailing \n

Use findcc() from ../mkdep so ${CC} can contain multiple tokens instead
of trying to exec the whole contents of ${CC}.

* Use setprogname() in main().
* Don't grossly abuse getopt(); allow only -l and -L after filenames are
  encountered, and do the parsing of these options manually.

Add a missing "break;" statement so that this actually works again.

Add hooks for compiling on non-NetBSD hosts.

Add support for putting prefixes on the name of lint(1)'s front
and back ends.  This means that it can be built as a cross tool.
Part of the fix for bin/14280.

cleanup (prior to more adding more features):
- convert to ANSI KNF
- remove trailing whitespace
- translate some comments from german into english

code compiles and runs clean, and tested by running "make lint" against
xlint source using previous and this lint produces same results.

convert to use getprogname()

Handle arguments in a slightly more standard way (avoid resetting optind).

pass -X <id>[,<id>]... and -w to lint1

Pull up rev. 1.19:
Use execvp() instead of execv(). Deals with the case where CC, the compiler,
is not a full path. For instance, "cc".

Use execvp() instead of execv(). Deals with the case where CC, the compiler,
is not a full path. For instance, "cc".

Sync w/ netbsd-1-5-base.

fix up NetBSD RCS Ids to match the standard, and the leading comment as
to match as well.  No functional changes.

Add parsing of the CC env variable, and a -Bpath flag (for /usr/libexec)
so cross-building of lint libraries can be made possible.  Tested
building a powerpc libc via make build on an alpha.

Define __LINT__, per XCU5.

Slight improvement to last fix:  don't need to reset the cpp tempfile for
files which don't need to be run through cpp.  Also, use vfork() instead of
fork().

Both from Anders Hjalmarsson (Anders.Hjalmarsson@economics.gu.se)

Apply fix from PR bin/8328 by Anders Hjalmarsson (Anders.Hjalmarsson@economics.gu.se):

work properly again when passwd multiple files on the command line.

More cosmetics in the usage message...

Now that we use cc -E, -U__GNUC__ to avoid gnu extensions. Nuke the
__attribute__ and __extension__ workarounds.
Our invariant is: No gcc extensions if __GNUC__ is not defined, so lint
should not be playing around trying to pretend it is gcc.

Deal with gcc __extension__({ }). We assume that it returns a value 0, and
we prepend a /*NOSTRICT*/ comment to avoid constant in conditional context
warning.

use "cc -E" not cpp directly.  this allows lint(1) to be unaware of any
machine-dependant defines that cc(1) may normally define, and also
means that the mips ports work again.

Use the C preprocessor define "__mips__" instead of "__mipse{b,l}__"
if we are running on a mipseb or mipsel platform..  This is the same
behaviour as the C compiler.

NOTE: This is only applied to the 1.4 branch.  A proper long-term fix
      using the equivalent of "${CC} -E" instead of lint(1) trying to
      guess what to define will be coming.

pass -CC instead of -C to cpp so that we preserve macro comments in place.

Eliminate memory leaks.
Use  xrealloc()  rather than  xmalloc/memcpy/free  trio.

Add -d option, so lint can be used safely when we're not building "in
place" (e.g., when DESTDIR is set).  This causes the lint driver to
pass -nostdinc -idirafter <dir> to cpp, causing it to ignore
/usr/include and look somewhere else instead..

WARNSify

Use S_IS*(), not S_IF*.

* accept the file name '-' to indicate that standard input is to be
  used as lint1 input.  That involves having lint pass the '-' through
  to the cpp which preprocesses the lint1 input, and having lint1's
  scanner recognize a cpp filename "" as "{standard input}".

don't rely on macros predefined by cpp

RCS id cleanup

lint(1) implementation, by Jochen Pohl.  named 'xlint' for a similar
reason to why 'install' is named 'xinstall'.

Initial revision