Having Boolean aliased to int creates ambiguities since int is widely
used. Allow to occasionally compile make with -DUSE_DOUBLE_BOOLEAN to
check that the type definitions still agree.
A string like OP_DEPENDS|OP_OPTIONAL|OP_PRECIOUS is much easier to read
and understand than the bit pattern 00000089.
The implementation in enum.h looks really bloated and ugly, but using
this API is as simple and natural as possible. That's the trade-off.
In enum.h, I thought about choosing the numbers in the macros such that
it is always possible to combine two of them in order to reach an
arbitrary number, because of the "part1, part2" in the ENUM__SPEC macro.
The powers of 2 are not these numbers, as 7 cannot be expressed as the
sum of two of them. Neither are the fibonacci numbers since 12 cannot
be expressed as the sum of 2 fibonacci numbers. I tried to find a
general pattern to generate these minimal 2-sum numbers, but failed.
Theoretically this should show the enum constant names when displaying
the field in the debugger. This only happens for GNode.flags though,
but not for GNode.type. Not sure what the exact cause is.
This allows debuggers to print symbolic names for the type flags.
The names of these constants need to be revised. These flags probably
started as the 3 dependency operators (:, !, ::), and at that time, the
prefix OP_ made sense. The other flags are not related to the type of
the dependency operator and could have been placed in GNode.flags as
well.
Having a numbered list above the type was not helpful since the numbers
didn't serve any purpose, they just consumed screen space.
Several of these list items didn't have an obvious relationship to the
struct fields. It's better to have just a rough introduction as the
type level documentation, leaving the details to the individual fields.
Converting the types and flags and other constants into separate types
and defining them outside the struct leaves more space to see the
relationship of the struct fields.
Limiting the documentation of each field to a single line, as suggested
by the end-of-line comments, reduces clarity since several of the fields
need way more documentation to be properly understood.
The previous documentation mentioned Str_Concat, but str_concat has been
written in lowercase for years. The "flags" are not flags since they
cannot be combined, not even when they are written in hex.
When parsing variable assignments other than := and if
value contains '$' attempt Var_Subst the same as for :=,
if the value does not parse correctly, we get a fatal error
including file an line number.
This can greatly help with finding the cause of problems.
Reviewed by: christos
Allow tracking of max chain length, to see how well the hash
tables are working.
Pull the actual hash operation into a marco so it can be
easily changed - for experimenting.
The current hash, is pretty good.
Reviewed by: christos
There are several types of flags for variables, and these cannot be
mixed. To prevent accidental typos, these are defined in separate enum
types. Clang warns about direct assignments between distinct types, but
not about mixing distinct types in binary expressions like A | B. GCC
does not warn at all.
When analyzing a build log, the first 'stopped' output
from make, is the end of interesting output.
Normally when a build fails deep down in a parallel build
the log ends with many blockes of error output from make,
with all but the fist being unhelpful.
We add a function dieQuietly() which will return true
if we should supress the error output from make.
If the failing node was a sub-make, we want to die quietly.
Also when we read an abort token we call dieQuietly telling we
want to die quietly.
This behavior is suppressed by -dj or
setting .MAKE.DIE_QUIETLY=no
Reviewed by: christos
debug was declared extern, but debug_file was not; correct this and define
debug_file in main.c (as debug is) to fix the -fno-common build.
-fno-common will become the default with GCC10/LLVM11.
Patch from kevans at freebsd
well as *interrupted* targets to be deleted. The name and behavior of
the variable matches gmake.
Also fix a glitch in newline output on error in compat mode that I
discovered while doing it.
Closes PR 51376.
realpath(3) is expensive, and meta mode at least uses it extensively.
We use cached_realpath() to save the result of successful calls to
realpath(3) in a private variable context.
This improves the worst case performance (eg examining libc with
nothing to do) by a factor of 4.
Reviewed by: christos
At least GNU Hurd lacks MAXHOSTNAMELEN/MAXPATHLEN/PATH_MAX, so make sure
those have some arbitrary semi-sane values.
from Robert Millan in PR misc/50166, with some modifications from myself
Quite extensive rewrite of the Suff module. Some ripple effects into
Parse and Targ modules too.
Dependency searches in general were made to honor explicit rules so
implicit and explicit sources are no longer applied on targets that
do not invoke a transformation rule.
Archive member dependency search was rewritten. Explicit rules now
work properly and $(.TARGET) is set correctly. POSIX semantics for
lib(member.o) and .s1.a rules are supported.
.SUFFIXES list maintenance was rewritten so that scanning of existing
rules works when suffixes are added and that clearing the suffix list
removes single suffix rules too. Transformation rule nodes are now
mixed with regular nodes so they are available as regular targets too
if needed (especially after the known suffixes are cleared).
The .NULL target was documented in the manual page, especially to
warn against using it when a single suffix rule would work.
A deprecation warning was also added to the manual and make also
warns the user if it encounters .NULL.
Search for suffix rules no longer allows the explicit dependencies
to override the selected transformation rule. A check is made in
the search that the transformation that would be tried does not
already exist in the chain. This prevents getting stuck in an infinite
loop under specific circumstances. Local variables are now set
before node's children are expanded so dynamic sources work in
multi-stage transformations. Make_HandleUse() no longer expands
the added children for transformation nodes, preventing triple
expansion and allowing the Suff module to properly postpone their
expansion until proper values are set for the local variables.
Directory prefix is no longer removed from $(.PREFIX) if the target
is found via directory search.
The last rule defined is now used instead of the first one (POSIX
requirement) in case a rule is defined multiple times. Everything
defined in the first instance is undone, but things added "globally"
are honored. To implement this, each node tracks attribute bits
which have been set by special targets (global) instead of special
sources (local). They also track dependencies that were added by
a rule with commands (local) instead of rule with no commands (global).
New attribute, OP_FROM_SYS_MK is introduced. It is set on all targets
found in system makefiles so that they are not eligible to become
the main target. We cannot just set OP_NOTMAIN because it is one of
the attributes inherited from transformation and .USE rules and would
make any eligible target that uses a built-in inference rule ineligible.
The $(.IMPSRC) local variable now works like in gmake: it is set to
the first prerequisite for explicit rules. For implicit rules it
is still the implied source.
The manual page is improved regarding the fixed features. Test cases
for the fixed problems are added.
Other improvements in the Suff module include:
- better debug messages for transformation rule search (length of
the chain is now visualized by indentation)
- Suff structures are created, destroyed and moved around by a set
of maintenance functions so their reference counts are easier
to track (this also gets rid of a lot of code duplication)
- some unreasonably long functions were split into smaller ones
- many local variables had their names changed to describe their
purpose instead of their type
with versions prefixed by MAKE_ATTR_* to avoid modifying the
implementation namespace. Make sure they are available in all places
using nonints.h to fix bootstrap on Linux.
Use a common variable (savedEnv) to track that to avoid wasting memory.
Also, if providing setenv and unsetenv, do getenv too to ensure a consistent
set.
meta_oodate: if a file that was written or linked within our bailiwick,
but outside of .OBJDIR is missing, add it to missingFiles.
If we get to the end of the .meta file without seeing it [re]moved,
then consider the target out-of-date.
In this mode, a .meta file is created for each target, capturing
the expanded commands used, any command output, and if filemon(9)
is available, a record of system calls which are of interest.
Not enabled unless USE_META=yes is set when building make.
Also, if FILEMON_H exists, meta.c will be compiled to use filemon(9).
