September 2007 - macports-dev - lists.macosforge.org

Re: [28702] trunk/base/HACKING
by markd＠macports.org 07 Sep '07

07 Sep '07

macports-dev(a)lists.macosforge.org writes: >Added: trunk/base/HACKING (0 => 28702) > > >--- trunk/base/HACKING (rev 0) >+++ trunk/base/HACKING 2007-09-07 18:34:33 UTC (rev 28702) >@@ -0,0 +1,13 @@ >+Whitespace rules as discussed on the list: >+ >+* All source code files MUST use soft tabs at a tabstop of 4. No hard >tabs are allowed. >+* All source code files SHOULD have the following as the first line of >the file: >+ # -*- Mode: Tcl; tab-width: 4; indent-tabs-mode: nil; >c-basic-offset: 4 -*- vim:et:sw=4:ts=4:sts=4 >+ This is a modeline that works for both emacs and vim. >+* Portfiles SHOULD use soft tabs at a tabstop of 4, but implementation >of this is left up to the discretion of the maintainer. >+* Portfiles SHOULD use the given modeline >+* Makefiles MUST use tabs as it is required by the syntax. Makefiles >SHOULD use a tab stop of 8. >+* Makefiles MAY use a modeline. The following works for emacs and vim: >+ # -*- Mode: Makefile; tab-width: 8; indent-tabs-mode: t -*- >vim:noet:sw=8:ts=8 >+* All other files (documentation, etc) SHOULD use soft tabs at a tabstop >of 4 if the document format allows. >+* All other files (documentation, etc) SHOULD NOT use a modeline as it >is probably meant for human consumption. I think that separately sprinkling documentation around in svn is something that shouldn't be done anymore. They are parallel documentation trains. I think they should be put into Trac tickets for inclusion in official docs, which are the guide and man pages. I just included some things from base/doc/INTERNALS into the guide, and it should really be deleted now. An email in a few minutes will spell out more details. Mark

1 0

Re: [28665] trunk/dports/editors/scite
by Ryan Schmidt 07 Sep '07

07 Sep '07

On Sep 6, 2007, at 07:27, source_changes(a)macosforge.org wrote: > Revision: 28665 > http://trac.macosforge.org/projects/macports/changeset/28665 > Author: nox(a)macports.org > Date: 2007-09-06 05:27:36 -0700 (Thu, 06 Sep 2007) > > Log Message: > ----------- > scite: Added universal support! Don't worry, be happy! > > Modified Paths: > -------------- > trunk/dports/editors/scite/Portfile > trunk/dports/editors/scite/files/patch-scite-gtk-makefile > > Modified: trunk/dports/editors/scite/Portfile > =================================================================== > --- trunk/dports/editors/scite/Portfile 2007-09-06 12:22:19 UTC > (rev 28664) > +++ trunk/dports/editors/scite/Portfile 2007-09-06 12:27:36 UTC > (rev 28665) > @@ -42,7 +42,22 @@ > > pre-build { > cd ${workpath}/scintilla/gtk > - system "make GTK2=yes" > + > + if {[variant_isset universal]} { > + foreach {arch} {i386 ppc} { > + system "make GTK2=yes UNIVERSAL_CFLAGS=\"-isysroot /Developer/ > SDKs/MacOSX10.4u.sdk -arch ${arch}\"" > + > + file mkdir ../bin/${arch} > + move ../bin/scintilla.a ../bin/${arch}/scintilla.a Shouldn't these paths be absolute paths? I didn't think the cwd was guaranteed to be anything in particular. I'm surprised that the "cd" command above works. > + > + system "make clean" > + } > + } else { > + system "make GTK2=yes" > + } > + > + cd ../bin > + system "lipo -create -arch i386 i386/scintilla.a -arch ppc ppc/ > scintilla.a -output scintilla.a" Does this really work for the non-universal build? Don't you only want to lipo if the universal variant has been requested? > } > > destroot.destdir prefix=${destroot}${prefix} > @@ -51,7 +66,17 @@ > xinstall -m 0444 ${worksrcpath}/../doc/scite.1 ${destroot}$ > {prefix}/share/man/man1 > } > > +variant universal { > + post-patch { > + cd ${workpath} > + > + reinplace -E "/^CXXBASEFLAGS/s|\$| \$(UNIVERSAL_CFLAGS)|" > scintilla/gtk/makefile > + reinplace -E "/^CXXBASEFLAGS/s|\$| $ > {configure.universal_cflags}|" \ > + scite/gtk/makefile > + } > + > + build.args-append "UNIVERSAL_LDFLAGS=\"$ > {configure.universal_ldflags}\"" > +} > + > livecheck.distname SciTE > > -universal_variant no > - [snip]

3 2

ghostscript and X11
by N_Ox 06 Sep '07

06 Sep '07

Hello, As the recent discussions on this list concluded, each port should enable by default most of the features common user may want. Another thing i believe would be good is that lib ports with doxygen documentation should have a doc variant that build it. But, one thing i don't understand is that if you want to build doxygen, you must install X11 for the ghostscript dependency. Shouldn't ghostscript X11 support be a variant of the port? I don't think users who are just willing to build some API documentation want X11 installed. Regards, -- Anthony Ramine, the infamous MacPorts Trac slave. nox(a)macports.org

3 4

daemondo defeats purpose of launchd?
by Ryan Schmidt 06 Sep '07

06 Sep '07

I have this problem: the lighttpd web server on my MacBook Pro unexpectedly dies rather frequently (much more frequently than it did on my PowerBook). When this happens, the daemondo process is still there, and I have to use launchctl to unload and then load the lighttpd plist to get the server running again. The point of launchd is that it notices when processes unexpectedly quit, and restarts them automatically so close to no downtime is observed. But launchd is monitoring daemondo, which does not exit, so launchd does not restart anything. And daemondo doesn't seem to have the same process-death-noticing feature so daemondo is not automatically restarting the failed lighttpd. Comments?

5 12

Re: [28660] trunk/base/src/port1.0/portbuild.tcl
by Weissmann Markus 06 Sep '07

06 Sep '07

While thinking about it: Wouldn't it be best in the case of a given build-style to use the full path to the chosen tool? Like "/usr/bin/ bsdmake" etc. else there could be nasty conflicts with ports not naming their executables exactly like the platform does (which at least is the case on different operating systems than OS X). -Markus On 06.09.2007, at 12:28, source_changes(a)macosforge.org wrote: > Revision 28660 Author afb(a)macports.org Date 2007-09-06 03:28:48 > -0700 (Thu, 06 Sep 2007) Log Message fix gnu/bsd make defaults > Modified Paths > trunk/base/src/port1.0/portbuild.tcl > Diff > Modified: trunk/base/src/port1.0/portbuild.tcl (28659 => 28660) --- > trunk/base/src/port1.0/portbuild.tcl 2007-09-06 10:07:07 UTC (rev > 28659) +++ trunk/base/src/port1.0/portbuild.tcl 2007-09-06 10:28:48 > UTC (rev 28660) @@ -61,13 +61,17 @@ bsd { if {[option os.platform] > == "darwin"} { return bsdmake + } elseif {[option os.platform] == > "freebsd"} { + return make } else { - return make + return > pmake } } gnu { if {[option os.platform] == "darwin"} { return > gnumake + } elseif {[option os.platform] == "linux"} { + return > make } else { return gmake } > _______________________________________________ > macports-changes mailing list > macports-changes(a)lists.macosforge.org > http://lists.macosforge.org/mailman/listinfo/macports-changes --- Markus W. Weissmann http://www.mweissmann.de/

2 1

The Emperor's New Clothes!
by Juan Manuel Palacios 06 Sep '07

06 Sep '07

Thanks a great deal to Chris for pushing his css mockup into trunk/ www! Even if temporary, this design is looking pretty good in my opinion. And if we all end up liking it enough.. maybe we can even drop the "temporary" moniker ;-) You can always look at the new design live on my server, http:// apollo.homeunix.net/macports , which always mirrors the contents of trunk/www at any moment. Content of the pages is still the same, though, Chris only worked on the aesthetics. I'm hoping with these initiatives we can mull people's interest to start contributing some new content for the new site. In the mean time... enjoy! -jmpp

1 0

Problem with patchfiles (same name for different versions)
by Vincent Lefevre 05 Sep '07

05 Sep '07

MacPorts stores patchfiles in /opt/local/var/macports/distfiles/${name} e.g. /opt/local/var/macports/distfiles/mpfr contains here: -rw-r--r-- 1 root admin 787634 2007-07-24 16:10:58 mpfr-2.2.1.tar.bz2 -rw-r--r-- 1 root admin 872947 2007-08-29 15:38:57 mpfr-2.3.0.tar.bz2 -rw-r--r-- 1 root admin 1826 2007-07-24 16:10:49 patch01 -rw-r--r-- 1 root admin 1687 2007-07-24 16:10:52 patch02 -rw-r--r-- 1 root admin 11655 2007-07-24 16:10:52 patch03 -rw-r--r-- 1 root admin 7509 2007-07-24 16:10:53 patch04 -rw-r--r-- 1 root admin 16992 2007-07-24 16:10:53 patch05 The problem is that patches are not stored in a subdirectory that depends on the version. So, there is a clash if I want to declare patch01 as a new patchfile for mpfr-2.3.0. What is the recommended solution? -- Vincent Lefèvre <vincent(a)vinc17.org> - Web: <http://www.vinc17.org/> 100% accessible validated (X)HTML - Blog: <http://www.vinc17.org/blog/> Work: CR INRIA - computer arithmetic / Arenaire project (LIP, ENS-Lyon)

5 9

Re: [28625] trunk/dports/databases/mysql5-devel/Portfile
by Jyrki Wahlstedt 05 Sep '07

05 Sep '07

On 5.9.2007, at 16.41, source_changes(a)macosforge.org wrote: > Revision 28625 Author jwa(a)macports.org Date 2007-09-05 06:41:03 > -0700 (Wed, 05 Sep 2007) Log Message > version bump to 5.1.21, add 5 to binaries in configuration, remove > patch as unnecessary, change startup, add violate_mtree due to /sql- > bench, didn't find a setting to put it somewhere else /sql-bench there means really /opt/local/sql-bench, wrote ${prefix}, but that of course did not work from shell (should perhaps use apostrophes instead of quotes). ! ! Jyrki Wahlstedt ! http://www.wahlstedt.fi/jyrki/ ! ! Our life is no dream; but it ought to become one and perhaps will. ! PGP key ID: 0x139CC386 fingerprint: F355 B46F 026C B8C1 89C0 A780 6366 EFD9 139C C386

2 1

Re: [28595] trunk/dports/math/bc/Portfile
by Ryan Schmidt 05 Sep '07

05 Sep '07

On Sep 4, 2007, at 03:34, source_changes(a)macosforge.org wrote: > Revision: 28595 > http://trac.macosforge.org/projects/macports/changeset/28595 > Author: mww(a)macports.org > Date: 2007-09-04 01:34:17 -0700 (Tue, 04 Sep 2007) > > Log Message: > ----------- > take maintainership, add sha1 checksum & livecheck [snip] > + > +livecheck.check regex > +livecheck.url http://ftp.gnu.org/gnu/bc/?C=M&O=D > +livecheck.regex bc-(\[0-9\.\]+\[0-9\])\\. > + You don't need a backslash before a period in a character class. So livecheck.regex bc-(\[0-9.\]+\[0-9\])\\. would be better.

2 2

Re: New website! ... not yet, really (Fwd: [28303] trunk/www)
by Juan Manuel Palacios 05 Sep '07

05 Sep '07

On Sep 1, 2007, at 7:08 PM, mark.duling(a)biola.edu wrote: > I like your ideas. It sounds like a good way forward. > > Mark > So, from your message and Ryan's, and the silence that resulted on the list thereon, I can assume that either everyone liked my ideas and therefore didn't feel any need to reply... or that only you two read my message ;-) In any case, I think that with the consensus we have so far we can start putting something together to jump-start this effort, and adapt as we go along if other people do chime in with input that might change our course. And to finally get started, I would love it if we already started adding new content to the trunk/www svn space, based on my outline of what the new site should say/have. I'll start this myself ASAP but everyone else should feel free to jump in. There is also a 'www' component in new ticket submissions, so people could also submit patches if either they don't confident enough to commit and want to get some feedback or if they don't have commit access altogether. Mark, how would you feel about renaming the "Documentation" milestone to something like "Documentation and www"... or whatever to that effect? (feel free to suggest a new name ;-) Lastly, I'll coordinate with Chris to include his css as the new default one. But that doesn't have to be final, a redesign for that could also be open to discussion (and maybe even an open contest too ;-) Regards,.... -jmpp

2 1