Previously there were at least 5 different ways MACHINE_ARCH could be set,
some statically and some at run time, and in many cases these settings
differed, leading to issues at pkg_add time where there was conflict
between the setting encoded into the package and that used by pkg_install.
Instead, move to a single source of truth where the correct value based on
the host and the chosen (or default) ABI is determined in the bootstrap
script. The value can still be overridden in mk.conf if necessary, e.g.
for cross-compiling.
ABI is now set by default and if unset a default is calculated based on
MACHINE_ARCH. This fixes some OS, e.g. Linux, where the wrong default was
previously chosen.
As a result of the refactoring there is no need for LOWER_ARCH, with
references to it replaced by MACHINE_ARCH. SPARC_TARGET_ARCH is also
removed.
"/bin/sh: There is no process to read data written to a pipe." is logged using
system supplied shell during the scan stage of a bulkbuildi otherwise.
Reviewed by wiz@
- unit-tests/modts now works on MirBSD
- meta mode
- ApplyModifiers: when we parse a variable which is not the entire modifier
string, or not followed by ':', do not consider it as containing modifiers.
- when long modifiers fail to match, check sysV style.
- :hash - cheap 32bit hash of value
- :localtime, :gmtime - use value as format string for strftime.
- fix for use after free() in CondDoExists().
- boot-strap (TOOL_DIFF): aparently at least on linux distro
formats the output of 'type' differently - so eat any "()"
- correct sysV substitution handling of empty lhs and variable
- correct exists() check for dir with trailing /
- correct handling of modifiers for non-existant variables during evaluation
of conditionals.
- fix for incorrect .PARSEDIR when .OBJDIR is re-computed after makefiles
have been read.
- fix example of :? modifier in man page.
- sigcompat.c: convert to ansi so we can use higher warning levels.
- parse.c: SunOS 5.8 at least does not have MAP_FILE
- use mmap(2) if available, for reading makefiles
- to ensure unit-tests results match, need to control LC_ALL as well as LANG.
- if stale dependency is an IMPSRC, search via .PATH
- machine.sh: like os.sh, allow for uname -p producing useless drivel
- boot-strap: document configure knobs for meta and filemon.
- unit-tests/modts now works on MirBSD
- meta mode
- ApplyModifiers: when we parse a variable which is not the entire modifier
string, or not followed by ':', do not consider it as containing modifiers.
- when long modifiers fail to match, check sysV style.
- :hash - cheap 32bit hash of value
- :localtime, :gmtime - use value as format string for strftime.
- fix for use after free() in CondDoExists().
- boot-strap (TOOL_DIFF): aparently at least on linux distro
formats the output of 'type' differently - so eat any "()"
- correct sysV substitution handling of empty lhs and variable
- correct exists() check for dir with trailing /
- correct handling of modifiers for non-existant variables during evaluation
of conditionals.
- fix for incorrect .PARSEDIR when .OBJDIR is re-computed after makefiles
have been read.
- fix example of :? modifier in man page.
- sigcompat.c: convert to ansi so we can use higher warning levels.
- parse.c: SunOS 5.8 at least does not have MAP_FILE
- use mmap(2) if available, for reading makefiles
- to ensure unit-tests results match, need to control LC_ALL as well as LANG.
- if stale dependency is an IMPSRC, search via .PATH
- machine.sh: like os.sh, allow for uname -p producing useless drivel
- boot-strap: document configure knobs for meta and filemon.
- boot-strap (TOOL_DIFF): set this to ensure tests use the same version
of diff that configure tested
- in jobs mode, when we discover we cannot make something, call
PrintOnError before exit.
- formatting fixes for ignored errors
- ensure jobs are cleaned up regardless of where wait() was called.
- os.sh (MACHINE_ARCH): watch out for drivel from uname -p
- call PrintOnError from JobFinish when we detect an error we are not
ignoring.
- use bmake_signal() which is a wrapper around sigaction() in place of
signal()
- add .export-env to allow exporting variables to environment without
tracking (so no re-export when the internal value is changed).
- fix for .info et al being greedy.
- back to using realpath on argv[0] but only if contains '/' and does not
start with '/'.
