Date: Wed, 12 May 2004 19:55:29 +0200 From: Herve Quiroz <herve.quiroz@esil.univ-mrs.fr> To: java@FreeBSD.org Subject: Re: ports/66342: [PATCH] fix ECHO_MSG breakage in java ports Message-ID: <20040512175529.GA52412@arabica.esil.univ-mrs.fr> In-Reply-To: <200405121310.i4CDAUY3033287@freefall.freebsd.org> References: <200405121310.i4CDAUY3033287@freefall.freebsd.org>
next in thread | previous in thread | raw e-mail | index | archive | help
On Wed, May 12, 2004 at 06:10:30AM -0700, Roman Neuhauser wrote: > # herve.quiroz@esil.univ-mrs.fr / 2004-05-12 12:33:39 +0200: > > > > http://www.freshports.org/ports-ignore.php > > what kind of problem? you don't tell me what I should be looking for. > Sorry for that. For instance if you look at devel/ZendOptimizer: ZendOptimizer 2.5.1-1 / devel IGNORE: :\n Because of licensing restrictions, you must fetch the source distribution\n manually.\n Please access http://www.zend.com/store/free_download.php?pid=13\n with a web browser, read the license and click the \"I ACCEPT\" button.\n Download the source file, ZendOptimizer-2.5.1-FreeBSD4.0-i386.tar.gz,\n and place it in /usr/home/dan/ports/distfiles.\n You see the '\n' characters are in the output. > > - The first one is the "correct" regular expression but it doesn't work > > (that's why it is commented). > > right. you can't use the "\n" escape in replacement in the s command. > When I invoke sed from command-line (thus not from make) it works anyway. > > - The second is some kind of cheat, replacing '\n' by '%' in the regular > > expression, and then filtering through tr(1) to produce 'n'. This is > > quite ugly if you ask me but it works. > > it breaks ports whose IGNORE variable contains %. > Indeed, that's an issue. If there's a way to get rid of tr(1), it should work anyway. > another problem is that the third subst, if it worked, would break > long lines inside words, and would not reflow the text, producing > short line overflows. Indeed. > moreover, this patch, goes against my recent push to reduce > unnecessary commands in ${PORTSDIR}/Mk/*, so I guess it's not > surprising I'm not terribly excited. You're right. Actually the only problem I see with your solution is the same as the existing one (the output at FreshPorts.org), so it's not a big issue. I did not say I disagree with your patch BTW, just wanted to know your opinion about "the other way"... It's probably more a problem with FreshPorts.org than the ports tree itself. Herve
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20040512175529.GA52412>