Evening everyone! I am glad to announce that we've finally finished the 1.5.0 edition of MacPorts, readily available through all our regular distribution means: -) binary installer for both 10.4 (Universal) and 10.3, http:// svn.macports.org/repository/macports/downloads/MacPorts-1.5.0/ -) source tarballs, both gz and bz2 formats, same location as above; -) selfupdate. Of all the features and bug fixes in the changelog quoted below, I'd like to point out two in particular: 1) The MacPorts project no longer mixes its old DarwinPorts name with the new one at any level, whether that is in chosen installation paths, file names, strings we output to the user and many more things. This implies that upon installing 1.5.0 several things in a standard MacPorts installation will silently change to new naming and installation locations, as thoroughly detailed in the Wiki page that tracked this work as it was being developed: http://trac.macports.org/ projects/macports/wiki/MacPortsRenaming. Please read it carefully before getting scared about your configuration files or other stuff being missing ;-) 2) Even though still not complete, extensive work on the generation of rpm's off MacPorts Portfiles has pushed forward our ability to generate real binary packages, improving both the building of the rpm tool itself on Mac OS X and the "rpm" MacPorts target (renamed from "rpmpackage") that creates the corresponding rpm files/packages. These packages still need to be generated by users off their own MacPorts installations ("port rpm <port>"), and flawless functionality for each of them is still not guaranteed, unfortunately. For instance, one long standing shortcoming still with us is that rpm installed software (even if the package is generated through MacPorts) is separate from MacPorts installed software, they are still two different universes that don't talk to each other even if installed into the same /opt/local prefix. However, strong developer momentum on the packaging front raises our hope that the goal of providing fully working and ironed out binary packages will be reached eventually! So, without much further ado, enjoy MacPorts 1.5.0! Biggest thanks to all those who worked to produce this solid release, kudos! Regards,... -jmpp PS: Full 1.5.0 ChangeLog following: Release 1.5 (09-July-2007 tagged at r26818 by jmpp): - Many documentation updates and improvements, mostly thanks to Maun Suang and Brian Campbell who both started work on both improving our man pages and reviving the long forgotten MacPorts official guide, now nightly regen'd on an automated basis and hosted at a temporary location, http://geeklair.net/ macports_guide/, all thanks to Daniel J. Luke (dluke). - Improve built-in "daemondo" program logging (jberry r26537 & r26569). - rpm target now knows how to also build universal packages if requested (afb r26470). - The "rpmpackage" target was renamed to "rpm", and a new target "srpm" added that allows packaging a Portfile (and files/distfiles) up as a source package. Upgraded RPM to version 4.4.9 and changed OpenDarwin/ DarwinPorts to MacPorts. The default build location is now ${prefix}/src/macports instead of ${prefix}/src/apple (jmpp & afb in r26465, r26496, r26536, r26543). - Fixed a potential crasher in the delete command (ticket #12149, eridius r26397). - 'port delete --work' now removes $portbuiltpath instead of $workpath, effectively prunning empty dirs that up until now were left behind in the build directory (jmpp r26394). - gnustep portgroup for gnustep-make-2.0 (yves r26266). - Adding backwards compatibility glue for clients of the public (darwin|mac)ports1.0 API that use now deprecated procs naming (DarwinPorts namespace) (jmpp r26392). - Merged the dp2mp-move branch into trunk to finally move away from all DarwinPorts related strings and conventions both internally and externally, implying big changes in MacPorts pathnames with respect to user visible stuff. Read http://trac.macports.org/projects/macports/wiki/ MacPortsRenaming and http://trac.macports.org/projects/macports/log/branches/dp2mp- move/base for more full details and information (jmpp r26177). - Fix livecheck to properly de-escape livecheck.url (eridius r26041). - Add warning when it looks like PortIndex file is corrupt (eridius r26040). - Make portindex use stderr for errors (eridius r26038, ticket #11585). - Fix port variants calculation to properly account for negated variants and to detect problems between required and negated variants (ticket #11920, eridius r26036, patch by gwhitney). - Add global methods lpush, lpop, lshift, and lunshift. Works similar to lappend (in fact lpush is just lappend) and do what they sound like. Also add ldindex. Works like lindex, but deletes the element from the list. Documented in portfile.7 (eridius r26034). - Add global methods try and throw. Implemented based on the specification in TIP #89. Documented in portfile.7 (eridius r25979). - Handle encodings properly now. All Portfiles and .conf files are assumed to be utf-8, and reading them or calling portindex(1) should now work the same on all locales (ticket #11978, eridius r25975). - Add support for decoding of obscured maintainer addresses. We support two types of obscured addresses in Portfile maintainers field: (1) username ==> implies username@macports.org (2) subdomain.tld:username ==> implies username@subdomain.tld These are unobscured by port info, and by port submit during the submission process, but are left obscured in the Portfile to avoid accidental disclosure. (jberry r25795). - Update port.1 to reflect what "port dependents" really does (markd r25787, ticket #11898). - Distinguish a pseudo portname that evaluates to nothing from no arguments at all, improving behavior of 'installed', 'active', 'list', and 'search' pseudo portnames. (jberry in r25789, ticket #10674). - Bugfix startup item generation so that launchd.plists are disabled by default, as we claim they are in our documentation. (jberry r25785). - Update adduser/addgroup to use dscl instead of niutil. Also add support for realname key in addgroup (ticket #11012, eridius r25586). - Make a correction to port(1)'s determination of whether or not a port has been updated by making it compare ${version} and then $ {revision} rather than a single comparison of the compound ${version}_$ {revision}; the latter reports 2.01_2 to be newer than 2.01.01a25_0 because, at the fifth character, `_' comes after `.' in ASCII lexicographic order. (boeyms in r25310)