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

re: "mysql" port outdated?
by James Cox 22 May '07

22 May '07

On 22 May 2007, at 16:13, James Berry wrote: > On May 22, 2007, at 8:09 AM, Chris Pickel wrote: > >> There were just two people on the irc channel trying to install >> mysql +server. This didn't work because the mysql port depends on >> DarwinPortsStartup, which they couldn't fetch. >> >> This makes me wonder: should the "mysql" port still be around, or >> ought it be removed? Both it and mysql3 report their version as >> 3.23.58. The differences are: > > > Personally, I'd be in favor if ditching the mysql port if it > duplicates mysql3. But personally I use mysql5, so I may be missing > something... ;) > I'm not sure there is a very good reason to maintain a working mysql3 port - it's not recommended for use anymore. do we have a policy about upstream'ed out-of-date software ? -james

2 1

"mysql" port outdated?
by Chris Pickel 22 May '07

22 May '07

There were just two people on the irc channel trying to install mysql +server. This didn't work because the mysql port depends on DarwinPortsStartup, which they couldn't fetch. This makes me wonder: should the "mysql" port still be around, or ought it be removed? Both it and mysql3 report their version as 3.23.58. The differences are: "mysql3" port: * lists more official-looking mirrors * has some more options passed to configure * uses startupitem for +server "mysql" port: * uses DarwinPortsStartup for +server * claims support for sunos. Chris

2 1

Fwd: zsh-devel +mp_completion variant thougts
by N_Ox 21 May '07

21 May '07

I'm really going to hate myself for not hitting 'reply all' once again. Début du message réexpédié : > De : N_Ox <n.oxyde(a)gmail.com> > Date : 20 mai 2007 18:57:52 HAEC > À : Emmanuel Hainry <ehainry(a)free.fr> > Objet : Rép : zsh-devel +mp_completion variant thougts > > > Le 20 mai 07 à 16:58, Emmanuel Hainry a écrit : > >> Citando N_Ox : >>> I don't understand what you are saying at the end of your mail (the >>> only constraint part). >> >> What I mean is that completion functions are installed in >> /opt/local/share/4.3.4/functions for zsh-devel, the number part is >> different for zsh and also for future versions, hence the need to >> pay a >> little attention. > > I'll install it in share/zsh/site-functions. > >>> >>> How would you make it part of macports itself? >>> >> >> Probably put it in a contrib or share directory in macports and >> make it >> be installed by selfupdate. > > Don't think everybody wants this file. I still think a port would > be better. > >>> Finally, i'll take a look at the patch you linked. >>> I'm thinking about adding descriptions and all in it, as you've >>> maybe >>> already seen in the symfony >>> autocomplete. >>> >>> Le 19 mai 07 à 19:54, Emmanuel Hainry a écrit : >>> >>>> Citando Charlie Allom : >>>>> On Sat, May 19, 2007 at 03:40:48PM +0200, N_Ox wrote: >>>>>> Hello, >>>>>> >>>>>> I've just taken over zsh-devel port and one thing seems idiot to >>>>>> me: >>>>>> What if there is a macports update and that i want to update the >>>>>> autocompletion port command file? I'd have to do a new zsh-devel >>>>>> revision, and everybody would have to rebuild it just for this >>>>>> tiny >>>>>> update. >>>>>> >>>>>> So here's my proposal: Shouldn't we make a new "macports" >>>>>> category, >>>>>> with all of those files which do belong to other ports, but are >>>>>> macports specific? I think there are other ports like zsh out >>>>>> there >>>>>> which add files like that. >>>>>> >>>>>> So in the end, we would have zsh-devel without the mp_completion >>>>>> variant, and another port named macports/zsh-completion which >>>>>> would >>>>>> depend on path:${prefix}/bin/zsh:zsh. Nice, isn't it? >>>>> >>>>> just put it in shells/ ala bash-completion,as a sep. port. >>>>> >>>> >>>> the _port completion file does not merit a port for itself, as an >>>> emacs >>>> mode for portfile would not. In fact I think that the problem of >>>> emacs mode is the same as the one for zsh completions. Those >>>> files can >>>> be installed in the tree independently of the installation of the >>>> port. >>>> Then one can add the path to their fpath (for zsh) or emacs >>>> modes dir >>>> even for apple's zsh or emacs... The only constraint is that when >>>> installing zsh, the versioned function takes this file into >>>> account. >>>> >>>> So I think _port should be part of macports, not of a port. >>>> >>>> By the way http://trac.macports.org/projects/macports/ticket/11494 >>>> >>>> Best regards, >>>> >>>> Emmanuel >

1 0

email address anti-spam encoding in portfiles
by James Berry 21 May '07

21 May '07

I thought I'd run by everybody the novel concept of anti-spam encoding email addresses in portfiles. For all other cases we could obfuscate the address, but with Portfiles we're sortof stuck, since we need to be able to make these available in raw form in various places (svn, mpwa, etc). What if we adopt the convention in the maintainer field of using user/ domain instead of user@domain. I have a feeling the spambots won't find that, and it's pretty easy to recognize as a user (or to reconstitute as a machine). So to be completely clear, I propose that we would encode my email address (jberry(a)macports.org) as jberry/macports.org Feedback? James

7 18

Update ports to use gcc42?
by Ryan Schmidt 20 May '07

20 May '07

Now that gcc42 is a final release, what should be done with all the ports that depend on gcc41? Should they be updated to require gcc42 instead? It's not very many ports: devel/lua-numlua gnustep/gnustep-base gnustep/gnustep-make lang/ftidy python/py-numpy/Portfile textproc/pdftk There are also a few still using gcc40: lang/ftidy math/GiNaC math/R math/fftw-3-single math/fftw-3 math/nestedsums math/octave-forge math/octave science/libnc-dap science/xloops www/varnish x11/fluxbox Assuming gcc42 is capable of building each of the above ports, and that the port authors update these ports to use gcc42, is there a need to keep these old versions of gcc around? It's annoying enough for the user to have to spend a couple hours on an Intel Mac, or many many hours on a slower PowerPC, to compile a single version of gcc. It would be even more annoying if the user wanted to use various different ports, each of which required a different version of gcc.

5 5

zsh-devel +mp_completion variant thougts
by N_Ox 20 May '07

20 May '07

Hello, I've just taken over zsh-devel port and one thing seems idiot to me: What if there is a macports update and that i want to update the autocompletion port command file? I'd have to do a new zsh-devel revision, and everybody would have to rebuild it just for this tiny update. So here's my proposal: Shouldn't we make a new "macports" category, with all of those files which do belong to other ports, but are macports specific? I think there are other ports like zsh out there which add files like that. So in the end, we would have zsh-devel without the mp_completion variant, and another port named macports/zsh-completion which would depend on path:${prefix}/bin/ zsh:zsh. Nice, isn't it? My two cents, N_Ox.

3 4

#11960: py-paramiko - fix maintainer email address
by Adam Mercer 19 May '07

19 May '07

Hi I submitted a ticket fixing my email address in the maintainers field of the py-paramiko port a few days ago, could someone with commit access please commit this? http://trac.macports.org/projects/macports/ticket/11960 Cheers Adam

2 2

Fwd: [25313] trunk/dports/shells/zsh-devel
by N_Ox 18 May '07

18 May '07

I must be dumb... /me notes to himself: 'HIT REPLY ALL.' Début du message réexpédié : > De : N_Ox <n.oxyde(a)gmail.com> > Date : 18 mai 2007 23:47:21 HAEC > À : Vincent Lefevre <vincent-opdarw(a)vinc17.org> > Objet : Rép : [25313] trunk/dports/shells/zsh-devel > > Those files are already installed in ${prefix}/share/doc/${name}-$ > {version} > > % ls /opt/local/share/doc/zsh-4.3.4 > ChangeLog FEATURES LICENCE NEWS README dvi html pdf > > Le 18 mai 07 à 23:22, Vincent Lefevre a écrit : > >> On 2007-05-18 22:14:05 +0200, N_Ox wrote: >>> I've yet another time forgot to reply to all~~ >>> >>> Well, this mail is not for nothing: >>> i'm pleased to tell you i've fixed this problem by making a soelim >>> (1) >>> system call. >> >> Great! >> >>> By the way, I'll add an 'examples' variant before commiting it — >>> well, >>> i'm also to wait for >>> commit rights. >> >> You should also install the main doc files into >> ${prefix}/share/doc/${name}. >> >> For instance, on Debian: >> >> -rw-r--r-- 1 938 2006-09-08 02:52:33 BUGS.gz >> -rw-r--r-- 1 4941 2006-09-08 02:52:33 CONTRIBUTORS.gz >> -rw-r--r-- 1 49311 2004-07-02 22:49:03 ChangeLog-3.0.gz >> -rw-r--r-- 1 146735 2004-07-02 22:49:03 ChangeLog-3.1.gz >> -rw-r--r-- 1 52430 2004-07-02 22:49:03 ChangeLog-4.1.gz >> -rw-r--r-- 1 2441 2005-08-05 21:10:04 FEATURES.gz >> -rw-r--r-- 1 769 2006-09-08 02:52:33 FTP-README.gz >> -rw-r--r-- 1 4304 2006-08-09 20:24:02 MACHINES.gz >> -rw-r--r-- 1 1869 2007-05-10 17:55:34 META-FAQ.gz >> -rw-r--r-- 1 213 2007-05-10 17:55:34 NEWS.Debian.gz >> -rw-r--r-- 1 12787 2006-08-09 20:24:02 NEWS.gz >> -rw-r--r-- 1 303 2007-05-10 17:55:34 README.Debian.gz >> -rw-r--r-- 1 4746 2007-05-08 16:25:09 README.gz >> -rw-r--r-- 1 223 2006-09-08 02:52:34 README.zftp.gz >> -rw-r--r-- 1 142 2006-09-08 02:52:33 TODO.gz >> -rw-r--r-- 1 11015 2007-05-10 17:55:34 changelog.Debian.gz >> -rw-r--r-- 1 59789 2007-05-10 16:33:27 changelog.gz >> -rw-r--r-- 1 7836 2006-09-08 02:52:33 completion-style-guide.gz >> -rw-r--r-- 1 1829 2007-05-10 17:55:34 copyright >> -rw-r--r-- 1 113 2006-09-08 02:52:33 examples/Functions/acx.gz >> -rw-r--r-- 1 163 2006-09-08 02:52:33 examples/Functions/cat.gz >> -rw-r--r-- 1 116 2006-09-08 02:52:33 examples/Functions/cx.gz >> -rw-r--r-- 1 175 2006-09-08 02:52:33 examples/Functions/proto.gz >> -rw-r--r-- 1 273 2006-08-13 03:50:07 examples/Functions/pushd.gz >> -rw-r--r-- 1 109 2006-09-08 02:52:33 examples/Functions/ >> randline.gz >> -rw-r--r-- 1 52 2006-09-08 02:52:33 examples/Functions/yp.gz >> -rw-r--r-- 1 58 2006-09-08 02:52:33 examples/Functions/yu.gz >> -rw-r--r-- 1 681 2006-09-08 02:52:33 examples/Functions/zless.gz >> -rw-r--r-- 1 839 2006-09-08 02:52:33 examples/Functions/zls.gz >> -rw-r--r-- 1 253 2006-09-08 02:52:33 examples/Functions/zpgrep.gz >> -rw-r--r-- 1 1032 2006-09-08 02:52:34 examples/Misc/ >> bash2zshprompt.gz >> -rw-r--r-- 1 1746 2006-09-08 02:52:34 examples/Misc/c2z.gz >> -rw-r--r-- 1 10803 2006-09-08 02:52:34 examples/Misc/compctl- >> examples.gz >> -rw-r--r-- 1 1709 2006-07-26 02:33:47 examples/Misc/globtests.gz >> -rw-r--r-- 1 933 2006-09-08 02:52:34 examples/Misc/ >> globtests.ksh.gz >> -rw-r--r-- 1 552 2006-09-08 02:52:34 examples/Misc/job-control- >> tests.gz >> -rw-r--r-- 1 3958 2006-09-08 02:52:34 examples/Misc/lete2ctl.gz >> -rw-r--r-- 1 1741 2006-09-08 02:52:34 examples/Misc/make-zsh- >> urls.gz >> -rw-r--r-- 1 893 2006-09-08 02:52:35 examples/Util/ >> check_exports.gz >> -rw-r--r-- 1 1108 2006-09-08 02:52:35 examples/Util/difflog.pl.gz >> -rw-r--r-- 1 1867 2006-09-08 02:52:35 examples/Util/helpfiles.gz >> -rw-r--r-- 1 829 2006-09-08 02:52:35 examples/Util/ >> mkdisttree.sh.gz >> -rw-r--r-- 1 234 2006-09-08 02:52:35 examples/Util/preconfig.gz >> -rw-r--r-- 1 3882 2006-09-08 02:52:35 examples/Util/reporter.gz >> -rw-r--r-- 1 1069 2007-05-10 17:55:34 examples/ssh_completion.gz >> -rw-r--r-- 1 550 2007-05-10 17:55:34 examples/ssh_completion2.gz >> -rw-r--r-- 1 396 2006-09-08 02:52:35 examples/zlogin.gz >> -rw-r--r-- 1 870 2006-09-08 02:52:35 examples/zshenv.gz >> -rw-r--r-- 1 2474 2006-09-08 02:52:35 examples/zshrc.gz >> -rw-r--r-- 1 10573 2006-09-08 02:52:33 zsh-development-guide.gz >> -rw-r--r-- 1 314037 2007-05-10 17:55:34 zsh.texi.gz >> >> I don't think the zsh.texi and the old ChangeLog files are useful, >> though. >> >> -- >> 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) >

1 0

test.env not being evaled any more
by Blair Zajac 18 May '07

18 May '07

It looks like test.env is not being evaled any more. We use this in the Subversion test suite to find all the dylibs so that the programs can run: test.env DYLD_LIBRARY_PATH=`find ${worksrcpath}/subversion -type d -name .libs | xargs echo | sed -e 's/ /:/g'` CLEANUP=true This used to work, but now no longer does. Can we get this behavior back? And more importantly, how do we make changes to the port system without breaking existing ports? Do we need to set up a test suite of ports that we build to ensure that we don't break anything? Regards, Blair

1 0

Fwd: [25313] trunk/dports/shells/zsh-devel
by N_Ox 18 May '07

18 May '07

I've yet another time forgot to reply to all~~ Well, this mail is not for nothing: i'm pleased to tell you i've fixed this problem by making a soelim(1) system call. By the way, I'll add an 'examples' variant before commiting it — well, i'm also to wait for commit rights. Regards. Début du message réexpédié : > De : N_Ox <n.oxyde(a)gmail.com> > Date : 18 mai 2007 21:33:03 HAEC > À : Vincent Lefevre <vincent-opdarw(a)vinc17.org> > Objet : Rép : [25313] trunk/dports/shells/zsh-devel > > You are so right, and i'm so dumb! I do have zsh on Gentoo, where > the manpages are gzipped too. > I'll take a look. > > Le 18 mai 07 à 21:01, Vincent Lefevre a écrit : > >> On 2007-05-18 06:27:58 -0700, source_changes(a)macosforge.org wrote: >>> +post-destroot { >>> + # This manpage only includes every zsh manpage; as manpages >>> are gzipped in MacPorts, >>> + # includes fail, so this one is useless. >>> + delete ${destroot}${prefix}/share/man/man1/zshall.1 >> >> It would be better to fix the bug. FYI, the manpages are gzipped >> under >> Linux (Debian) too, and there's no problem with the zshall man page. >> >> But IMHO, it is probably better not to compress the manpages (that >> should be a way to solve the problem). Another problem with the >> gzipped man pages is that the man command sometimes freezes. And disk >> space on desktop machines are no longer a problem for things like man >> pages (if it is on some machines, then there should be an option[*] >> not to include the manpages). >> >> [*] e.g. a variant with a common name. >> >> Also note that deleting the zshall man page will give the user wrong >> information since "man zshall" will load the zshall man page that >> comes with Mac OS X. >> >> -- >> 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) >

2 1