summaryrefslogtreecommitdiff
path: root/mk/compiler/gcc.mk
AgeCommit message (Expand)AuthorFilesLines
2004-02-06I missed changing a defined(_NEED_GCC3) into a yes/no check.jlam1-2/+2
2004-02-06Whitespace nit.jlam1-2/+2
2004-02-06If we're passing through MAKEFLAGS variables whose values may containjlam1-3/+3
2004-02-06Remove a debugging statement.jlam1-2/+1
2004-02-06Hiding the PATH from certain phases of the build only accidentally workedjlam1-6/+5
2004-02-06We only prepend a directory to the PATH if we haven't already done sojlam1-10/+22
2004-02-05GCC_REQD should be appended to, not set.jlam1-2/+2
2004-02-05Allow bsd.compiler.mk to be included by both bsd.prefs.mk and bsd.pkg.mk.jlam1-119/+133
2004-02-05Don't compute the CC_VERSION unless ${CC} exists.jlam1-1/+5
2004-02-05_CC is really meant to be evaluated from the value of CC _as found_ injlam1-4/+5
2004-02-05_GCC_PREFIX and _GCC_SUBPREFIX are expected to end in "/" if we're usingjlam1-3/+3
2004-02-05Reset the values of _GCC_PREFIX and _GCC_SUBPREFIX from the value ofjlam1-1/+5
2004-02-04Missing a "jlam1-2/+2
2004-02-04case branches are separated by ;;jlam1-3/+3
2004-02-04Try to avoid "/" as the _GCC_SUBPREFIX.jlam1-2/+6
2004-02-04Make it more apparent where certain "not_found" values are being set forjlam1-3/+3
2004-02-04Avoid a non-zero exit value in a != variable definition.jlam1-2/+2
2004-02-04Define _GCC_PKG unconditionally since its value is determined by anotherjlam1-2/+2
2004-02-04Try to be smarter about detecting whether a cc in the PATH is actuallyjlam1-37/+40
2004-02-03Reorder checks when setting _GCC_SUBPREFIX so that we always search forjlam1-6/+9
2004-02-03Remove not_found checks now that we pre-filter them out when settingjlam1-3/+1
2004-02-03Do previous in a different way to avoid make errors.jlam1-2/+3
2004-02-03Don't add to LDFLAGS if the directories are "not_found".jlam1-1/+3
2004-02-03":H", not ":T", is the correct make variable modifier to get everythingjlam1-2/+2
2004-02-03Reorder sections a bit so that related sections are together.jlam1-49/+47
2004-02-03Whitespace nit.jlam1-2/+2
2004-02-03Fix a pervasive cut-and-paste bug: change a = back into a +=. Noted byjlam1-3/+3
2004-02-03Check whether _NEED_GCC[23] is yes/no rather than just defined, sincejlam1-11/+18
2004-02-03Move the _COMPILER_LD_FLAG setting to somehere global as it applies to alljlam1-2/+4
2004-02-03Avoid using "type" to find the absolute path to the compiler. We find itjlam1-27/+31
2004-02-02Whitespace fixes.jlam1-3/+3
2004-02-02If we can't find the package using pkg_info, then use a simple heuristicjlam1-2/+10
2004-02-02G/C _GCC_ARCHSUBDIR. All references to it from gcc*/buildlink*.mk filesjlam1-5/+2
2004-02-02If we need a GCC C++ or Fortran compiler, then define USE_GCC_SHLIBS asjlam1-1/+6
2004-02-02Rename _GCC{2,3}_REQD to _NEED_GCC{2,3} which more clearly expresses whatjlam1-7/+7
2004-02-02Support a new variable in package Makefiles:jlam1-17/+30
2004-02-01Fix a broken conditional.jlam1-3/+3
2004-02-01Use the gcc3-* packages instead of gcc3. The former are newer, and usingjlam1-16/+50
2004-02-01Use _GCC_REQD where we want the strictest value (GCC_REQD can be a list).jlam1-2/+2
2004-02-01Actually, it does matter what _GCC_PKG is set to as it's re-used later asjlam1-7/+3
2004-02-01Prepend the path to the compiler to PATH even if we're using the builtinjlam1-1/+6
2004-02-01We're included by bsd.prefs.mk, so no need to include bsd.prefs.mk forjlam1-3/+1
2004-02-01New compiler specification framework, based on mk/compiler.mk. Thejlam1-0/+224