From owner-freebsd-ports@FreeBSD.ORG Tue Dec 16 23:38:07 2008 Return-Path: Delivered-To: freebsd-ports@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id D67AA1065679 for ; Tue, 16 Dec 2008 23:38:07 +0000 (UTC) (envelope-from daniel@roe.ch) Received: from calvin.ustdmz.roe.ch (calvin.ustdmz.roe.ch [IPv6:2001:41e0:ff17:face::26]) by mx1.freebsd.org (Postfix) with ESMTP id F1D488FC23 for ; Tue, 16 Dec 2008 23:38:06 +0000 (UTC) (envelope-from daniel@roe.ch) Received: from roe (ssh-from [10.23.42.101]) by calvin.ustdmz.roe.ch (envelope-from ) with LOCAL id 1LCjRH-000KHw-PQ ; Wed, 17 Dec 2008 00:34:55 +0100 Date: Wed, 17 Dec 2008 00:34:55 +0100 From: Daniel Roethlisberger To: Guido Falsi Message-ID: <20081216233455.GC76371@hobbes.ustdmz.roe.ch> Mail-Followup-To: Guido Falsi , freebsd-ports@freebsd.org, bg1tpt@gmail.com References: <20081215152824.GA18862@megatron.madpilot.net> <20081215173812.GA42598@hobbes.ustdmz.roe.ch> <20081216112320.GB29493@megatron.madpilot.net> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20081216112320.GB29493@megatron.madpilot.net> User-Agent: Mutt/1.4.2.3i Cc: bg1tpt@gmail.com, freebsd-ports@freebsd.org Subject: Re: rrdtool 1.3.3 and dejavu dependency X-BeenThere: freebsd-ports@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Porting software to FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 16 Dec 2008 23:38:08 -0000 Guido Falsi 2008-12-16: > On Mon, Dec 15, 2008 at 06:38:12PM +0100, Daniel Roethlisberger wrote: > > > Even as the submitter of the PR in question, I fully agree. > > This is a good starting point :P > > > > I read the PR but don't really see such effects.. I'm using > > > mrtg and smokeping, with mrtg.cgi [1] and mailgraph.cgi. They > > > all work fine, mailgraph at least seems to use the > > > functionality described in the PR. > > > > > > Are we sure it isn't something else which is depending on > > > dejavu? > > > > Yes, pretty much. Note that rrdtool can fall back to other > > fonts, such as Bitstream Vera, but if no fonts are found, using > > `rrdtool graph` or language bindings such as RRDs::graph will > > fail. I just used Tobi's tutorial [2] to verify this again. > > > > Can you try to reproduce this with no fonts installed under > > /usr/local/lib/X11/fonts and using the tutorial [2] as a test? > > If it works for you, can you try to find out where it takes the > > font from? > > > > I guess it might be possible to make `rrdtool graph` work without > > the full install of x11-fonts/dejavu and all it's dependencies by > > either creating a lightweight dejavu font port without X11, or by > > adding dejavu to the rrdtool port somehow. > > > > I'm all open to better ideas than the current run-time dependency > > on x11-fonts/dejavu. I kind of hoped that the maintainer would > > be able to find a better solution, but unfortunately, there was > > no reaction from the maintainer. > > I made a few experiments on a spare PC at work(this is an AMD64 8.0 on > which I was testing ZFS, grat work to all the developers for that). > > I reverted your change on the machine, defined WITHOUT_X11 and made sure > all rrdtools dependencies where configured to require the least x11 > pieces(cairo requires disabling both xcb and glitz support). > > The machine had no ports installed. I ended up with the following ports > installed: > > bdftopcf-1.0.1 Convert X font from BDF to PCF > bitstream-vera-1.10_4 Bitstream Vera TrueType font collection > cairo-1.6.4_3,1 Vector graphics library with cross-device output > support > encodings-1.0.2,1 X.Org Encoding fonts > expat-2.0.1 XML 1.0 parser written in C > font-bh-ttf-1.0.0 X.Org Bigelow & Holmes TTF font > font-misc-ethiopic-1.0.0 X.Org miscellaneous Ethiopic font > font-misc-meltho-1.0.0_1 X.Org miscellaneous Meltho font > font-util-1.0.1 Create an index of X font files in a directory > fontcacheproto-0.1.2 Fontcache extension headers > fontconfig-2.5.0,1 An XML-based font configuration API for X Windows > fontsproto-2.0.2 Fonts extension headers > freetype2-2.3.7 A free and portable TrueType font rendering engine > gamin-0.1.9_2 A file and directory monitoring system > gettext-0.17_1 GNU gettext package > gio-fam-backend-2.16.5 FAM backend for GLib's GIO library > glib-2.16.5_1 Some useful routines of C programming (current > stable versi > gmake-3.81_3 GNU version of 'make' utility > icu-3.8.1_1 International Components for Unicode (from IBM) > libXfont-1.3.1_3,1 X font libary > libfontenc-1.0.4 The fontenc Library > libiconv-1.11_1 A character set conversion library > libtool-1.5.26 Generic shared library support script > libxml2-2.6.32_2 XML parser library for GNOME > mkfontdir-1.0.3_1 Create an index of X font files in a directory > mkfontscale-1.0.3 Creates an index of scalable font files for X > pango-1.20.5 An open-source framework for the layout and > rendering of i1 > pcre-7.8 Perl Compatible Regular Expressions library > perl-5.8.8_1 Practical Extraction and Report Language > pixman-0.10.0_2 Low-level pixel manipulation library > pkg-config-0.23_1 A utility to retrieve information about installed > libraries > png-1.2.33 Library for manipulating PNG images > python25-2.5.2_3 An interpreted object-oriented programming language > rrdtool-1.3.3_2 Round Robin Database Tools > xorg-fonts-truetype-7.3 X.Org TrueType fonts > xproto-7.0.10_1 X11 protocol headers > xtrans-1.0.4 Abstract network code for X > > As you can see it already sucked in various things and some fonts. I > tried the tutorial and it works flawlessly in such a setup. I think the > trick is this port: bitstream-vera-1.10_4. If there were some way to > install dejavu without sucking in all of libX11 it would be good too. > > could you make your tests without dejavu but with bitstream-vera and see > if it works? My tests show that it falls back to bitstream-vera, but I haven't tested it on a clean system. I think what we need is something along the lines of depending on (dejavu OR bitstream-vera). Or even any TTF font. There are other possibilities, like making graphing support optional in rrdtool, or replacing the dependency with a pkg-message, asking the user to install some font if graphing is to be used. > I see pango is depending on x11-fonts/xorg-fonts-truetype, which depends > on some fonts pacjkages but not on libX11. > > Hope my feedback helps. Due to ENOTIME, I wont be able to devote any meaningful amount of time to this any time soon. Can you submit the information in this thread as a PR against rrdtool so it doesn't get lost? -- Daniel Roethlisberger http://daniel.roe.ch/