> nmake knows about ast because ast follows a rational naming convention > PACKAGE_ast=$INSTALLROOT > implies > PACKAGE_ast_BIN = $INSTALLROOT/bin > PACKAGE_ast_INCLUDE = $INSTALLROOT/include/ast > PACKAGE_ast_LIB = $INSTALLROOT/lib > any other package with similar naming conventions will be "known" too
Good and sensible. Many packages even work this way! > for most non-ast packages a :PACKAGE: assertion at the top of > the makefile should take care of the -I -L -l option generation > e.g., for X11 builds > > :PACKAGE: - X11 > > the "-" means that X11 is not the main package > the main package controls where headers are installed > this will also add -lX11 as a prerequisite to all command :: assertions How is this functionality implemented? What part of nmake's source code should I inspect? > if :PACKAGE: doesn't find a package you can help it with PACKAGE_foo, > PACKAGE_foo_INCLUDE, PACKAGE_foo_LIB path name variables in the environment The process of crawling the filesystem to find the values of these variables is part of what I'm trying to automate... Norman > On Mon, 1 Jun 2009 21:01:06 -0400 (EDT) [email protected] (Norman > Ramsey) wrote: > > I'm working on porting some of my own code to use iffe and nmake. > > But my code relies on other code that will not be configured using > > these tools. I therefore face the usual problem of figuring out > > what -I options have to be passed to the compiler and what -L and -l > > options have to be passed to the linker in order to compile my code > > against non-AST libraries. > > > Reading the documentation for iffe(1), I cannot see how iffe might > > help with this problem. I realize that AST itself does not have this > > problem because it is self-contained: nmake sees all and knows all. > > > Does anyone have suggestions for how to use the AST tools to build > > something that depends on non-AST software? I hate GNU autoconf > > and refuse to go in that direction, but I'm hoping something in the > > AST suite would be better than starting from scratch. _______________________________________________ ast-users mailing list [email protected] https://mailman.research.att.com/mailman/listinfo/ast-users
