"throw" is a keyword defined in macports1.0/macports_util.tcl. Go read about it. It's basically an error with different order of arguments. The portfile used to be using exit 1, which is a bad idea. BTW, you really need to top-post if you're quoting that much text. I should be able to see your reply start on the first page. -Kevin Ballard On Oct 26, 2007, at 3:53 AM, Ryan Schmidt wrote:
On Oct 23, 2007, at 18:45, source_changes@macosforge.org wrote:
Revision: 30293 http://trac.macosforge.org/projects/macports/changeset/30293 Author: eridius@macports.org Date: 2007-10-23 16:45:28 -0700 (Tue, 23 Oct 2007)
Log Message: ----------- Update clisp 2.42, and at the same time make significant changes so this will actually build under Leopard
Modified Paths: -------------- trunk/dports/lang/clisp/Portfile
Removed Paths: ------------- trunk/dports/lang/clisp/files/
Modified: trunk/dports/lang/clisp/Portfile =================================================================== --- trunk/dports/lang/clisp/Portfile 2007-10-23 22:46:03 UTC (rev 30292) +++ trunk/dports/lang/clisp/Portfile 2007-10-23 23:45:28 UTC (rev 30293) @@ -1,10 +1,11 @@ # $Id$
PortSystem 1.0 + name clisp -version 2.40 +version 2.42 categories lang -maintainers nomaintainer +maintainers eridius openmaintainer platforms darwin description The Clisp Common Lisp Implementation long_description \ @@ -21,9 +22,9 @@ homepage http://clisp.cons.org/ master_sites gnu:clisp/release/${version} \ sourceforge -checksums sha1 4e64b3d04c91b16bc2626f5d89938dbcd4d71a56 +checksums md5 3fc28342e446a62c1eb4f65277929888 \ + sha1 4c39a5549d2821b1d5ab1cfca172527e9053af1c use_bzip2 yes -patchfiles patch-documentation.lisp
depends_lib port:readline \ port:gettext \ @@ -41,82 +42,74 @@ }
platform darwin 6 { - ui_msg "clisp is not supported on Jaguar (OS X 10.2.x)" - exit 1 + throw UNSUPPORTED "clisp is not supported on Jaguar (OS X 10.2.x)" }
[snip]
Hm! I've never seen "throw" in a portfile before. Interesting. Seems to work though. (I made a minor fix in r30386). Is "throw" how we should be handling unsupported platforms? What else should we be throwing, instead of 'return -code error "message"' which is what I thought was recommended before?
-- Kevin Ballard http://kevin.sb.org eridius@macports.org http://www.tildesoft.com