From openpandora at free.fr Fri Mar 1 21:25:39 2013 From: openpandora at free.fr (openpandora at free.fr) Date: Fri, 01 Mar 2013 22:25:39 +0100 Subject: [ARMedslack] Jumping cursor with touchscreen. Message-ID: <1362173139.51311cd311de4@imp.free.fr> Hi ! Sometimes, the cursor jumps to the top left corner of the screen when i'm manipulating particular things on the screen. Things like a scrollbar, an object in Thunar, a slider, the menu's up and down, so it's not a hardware or a calibration problem (same behavior on my two Pandora units). As i've no idea about where to begin (gtk bug ?), i'm asking there. beagleboard.googlecode.com/files/evtest.c reports no "jumping" coordinates. Are Xfce sliders and so on handled by gtk+, gtk+2 ? LS From openpandora at free.fr Sun Mar 31 17:06:44 2013 From: openpandora at free.fr (openpandora at free.fr) Date: Sun, 31 Mar 2013 19:06:44 +0200 (CEST) Subject: [ARMedslack] Using the x-toolchain. In-Reply-To: <493607550.4769340.1364749480394.JavaMail.root@zimbra53-e8.priv.proxad.net> Message-ID: <1217739984.4769980.1364749604331.JavaMail.root@zimbra53-e8.priv.proxad.net> Hi ! So i've suceessfully installed the x-toolchain on 32 and 64 bit hosts, and used distccd from an ARM machine. Now i would like to use the toolchain directly on my x86 64 bit host. For instance, i'd like to compile a kernel for ARM. I'd also like to use the other ARM distcc on my network. How do i setup the pathes, launch the stuff etc ? From dhm at mersenne.com Sun Mar 31 17:14:42 2013 From: dhm at mersenne.com (David Madden) Date: Sun, 31 Mar 2013 10:14:42 -0700 Subject: [ARMedslack] Using the x-toolchain. In-Reply-To: <1217739984.4769980.1364749604331.JavaMail.root@zimbra53-e8.priv.proxad.net> References: <1217739984.4769980.1364749604331.JavaMail.root@zimbra53-e8.priv.proxad.net> Message-ID: <51586F02.20606@mersenne.com> On 3/31/2013 10:06, openpandora at free.fr wrote: > How do i setup the pathes, launch the stuff etc ? FWIW, OpenWRT has a fairly reliable cross-build system. It automagically downloads the necessary packages, compiles all the tools, then builds a kernel and selectable system utilities. The only problem is, it's fairly all-or-nothing. You can use the cross tools to compile random packages or your own code, but I don't think it's straightforward to pull out pieces of the toolchain for use outside an OpenWRT build tree. -- Mersenne Law LLC ? www.mersenne.com ? +1-503-679-1671 - Small Business, Startup and Intellectual Property Law - 1500 SW First Ave. ? Suite 1170 ? Portland, Oregon 97201 -------------- next part -------------- A non-text attachment was scrubbed... Name: smime.p7s Type: application/pkcs7-signature Size: 4003 bytes Desc: S/MIME Cryptographic Signature URL: From openpandora at free.fr Sun Mar 31 17:25:25 2013 From: openpandora at free.fr (openpandora at free.fr) Date: Sun, 31 Mar 2013 19:25:25 +0200 (CEST) Subject: [ARMedslack] Using the x-toolchain. In-Reply-To: <51586F02.20606@mersenne.com> Message-ID: <1180772258.4785221.1364750725300.JavaMail.root@zimbra53-e8.priv.proxad.net> Well, i'm compiling a kernel for the Open Pandora. From m-lists at biscuit.org.uk Sun Mar 31 17:44:51 2013 From: m-lists at biscuit.org.uk (Stuart Winter) Date: Sun, 31 Mar 2013 18:44:51 +0100 (BST) Subject: [ARMedslack] Using the x-toolchain. In-Reply-To: <1217739984.4769980.1364749604331.JavaMail.root@zimbra53-e8.priv.proxad.net> References: <1217739984.4769980.1364749604331.JavaMail.root@zimbra53-e8.priv.proxad.net> Message-ID: > Now i would like to use the toolchain directly on my x86 64 bit host. > For instance, i'd like to compile a kernel for ARM. I'd also like to use > the other ARM distcc on my network. > > How do i setup the pathes, launch the stuff etc ? You don't. It's not a full cross compiler - as it says in the script: it builds enough to work with distcc. Eric has made a set of scripts to build a full cross compiler http://taper.alienbase.nl/mirrors/alienarm/bootstrap/ I used it a few months ago to cross compile a kernel and it was fine. David - the only reason to use the cross compiler that I've supplied (when building natively with distcc) is so that the cross toolchain is the exact same source base + patches, kernel headers and glibc that's running natively -- as there'd most likely be issues if you're compiling using mismatched toolchains. However, if you're just cross compiling kernels entirely non-natively then it doesn't matter if the toolchains don't match. From nsollars at gmail.com Sun Mar 31 18:07:21 2013 From: nsollars at gmail.com (Nigel Sollars) Date: Sun, 31 Mar 2013 14:07:21 -0400 Subject: [ARMedslack] Quick question Message-ID: Hi, Just a quick one, is there a top level script / slackbuild esq script that starts a build of a complete Slackware distro ?.. Nige -- ?Science is a differential equation. Religion is a boundary condition.? Alan Turing -------------- next part -------------- An HTML attachment was scrubbed... URL: From m-lists at biscuit.org.uk Sun Mar 31 19:09:21 2013 From: m-lists at biscuit.org.uk (Stuart Winter) Date: Sun, 31 Mar 2013 20:09:21 +0100 (BST) Subject: [ARMedslack] Quick question In-Reply-To: References: Message-ID: > Just a quick one, is there a top level script / slackbuild esq script that > starts a build of a complete Slackware distro ?.. The build order of Slackware is not deterministic - the distribution evolves upon itself. There is no sort of 'make world' type script. Eric's documenting his build process and order for the ARM hard float port though, so you might find interesting. http://alien.slackbook.org/blog/armport/ From m-lists at biscuit.org.uk Sun Mar 31 19:16:32 2013 From: m-lists at biscuit.org.uk (Stuart Winter) Date: Sun, 31 Mar 2013 20:16:32 +0100 (BST) Subject: [ARMedslack] Quick question In-Reply-To: References: Message-ID: > The build order of Slackware is not deterministic - the distribution evolves upon itself. > There is no sort of 'make world' type script. I was just looking at the comments I put in the bootstrap script that I wrote for Slackware ARM's x11 packages: As you can see, it's not a simple affair. Since Slackware x86 is built on a full system (not a minimal chroot), this means that in order to get your packages compatible with Slackware x86, you need to rebuild many of the packages tens of times in order to link against the other packages. http://ftp.arm.slackware.com/slackwarearm/slackwarearm-current/source/x/x11/bootstrap-build From ottavio2006-usenet2012 at yahoo.com Sun Mar 31 20:35:37 2013 From: ottavio2006-usenet2012 at yahoo.com (Ottavio Caruso) Date: Sun, 31 Mar 2013 21:35:37 +0100 Subject: [ARMedslack] Quick question In-Reply-To: References: Message-ID: On Mar 31, 2013 8:09 PM, "Stuart Winter" wrote: > > > > Just a quick one, is there a top level script / slackbuild esq script that > > starts a build of a complete Slackware distro ?.. > > The build order of Slackware is not deterministic - the distribution evolves upon itself. > There is no sort of 'make world' type script. > > Eric's documenting his build process and order for the ARM hard float port > though, so you might find interesting. > http://alien.slackbook.org/blog/armport/ > ____________________________________________ Incidentally Eric mentions the Chromebook as one of the reasons to port the hf. I did install a full rootfs of Armedslack (sf) on the Chromebook and it did work. I then gave up because of lack of feedback and because my unit was faulty. -------------- next part -------------- An HTML attachment was scrubbed... URL: From m-lists at biscuit.org.uk Sun Mar 31 21:09:07 2013 From: m-lists at biscuit.org.uk (Stuart Winter) Date: Sun, 31 Mar 2013 22:09:07 +0100 (BST) Subject: [ARMedslack] Quick question In-Reply-To: References: Message-ID: > Incidentally Eric mentions the Chromebook as one of the reasons to port the > hf. I did install a full rootfs of Armedslack (sf) on the Chromebook and it > did work. I then gave up because of lack of feedback and because my unit > was faulty. Yep I'm writing this using WindowMaker with Slackware ARM on the Chomebook. I just used the miniroot and some Chromebook tweaks. I'm planning on making Slackware ARM installable on the Chromebook - I've already started laying the groundwork. Slackware ARM will run on any 32-bit machine that's armv4 + thumb or greater. Slackware needs a hard float port as we're the only distro without one, and given that ARM is becoming more prevalent it's (IMO) important that Slackware is able to deliver max performance on the machines with a hardware floating point unit -- allowing them to be able to render multi media efficiently. That said, for general usage (non multi media) I really don't think there's much in it with soft vs hard float. Running Firefox and some other stuff on this Chromebook is *fast* even with the default Slackware ARM packages. -- Stuart Winter Slackware ARM: http://arm.slackware.com From ottavio2006-usenet2012 at yahoo.com Sun Mar 31 21:12:55 2013 From: ottavio2006-usenet2012 at yahoo.com (Ottavio Caruso) Date: Sun, 31 Mar 2013 22:12:55 +0100 Subject: [ARMedslack] Quick question In-Reply-To: References: Message-ID: Are you using the default chrome os kernel or is it a native one? On Mar 31, 2013 10:09 PM, "Stuart Winter" wrote: > > > Incidentally Eric mentions the Chromebook as one of the reasons to port > the > > hf. I did install a full rootfs of Armedslack (sf) on the Chromebook and > it > > did work. I then gave up because of lack of feedback and because my unit > > was faulty. > > Yep I'm writing this using WindowMaker with Slackware ARM on the > Chomebook. I just used the miniroot and some Chromebook tweaks. > I'm planning on making Slackware ARM installable on the Chromebook - I've > already started laying the groundwork. > > Slackware ARM will run on any 32-bit machine that's armv4 + thumb or > greater. Slackware needs a hard float port as we're the only distro > without one, and given that ARM is becoming more prevalent it's (IMO) > important that Slackware is able to deliver max performance on the > machines with a hardware floating point unit -- allowing them to be able > to render multi media efficiently. > > That said, for general usage (non multi media) I really don't think > there's much in it with soft vs hard float. Running Firefox and some > other stuff on this Chromebook is *fast* even with the default Slackware > ARM packages. > > -- > Stuart Winter > Slackware ARM: http://arm.slackware.com > _______________________________________________ > ARMedslack mailing list > ARMedslack at lists.armedslack.org > http://lists.armedslack.org/mailman/listinfo/armedslack > -------------- next part -------------- An HTML attachment was scrubbed... URL: From ottavio2006-usenet2012 at yahoo.com Sun Mar 31 21:15:29 2013 From: ottavio2006-usenet2012 at yahoo.com (Ottavio Caruso) Date: Sun, 31 Mar 2013 22:15:29 +0100 Subject: [ARMedslack] Quick question In-Reply-To: References: Message-ID: I wish I hadn't thrown away my work in progress tutorial about slackware on chrome book as it contained some good tips about audio hacking. On Mar 31, 2013 10:12 PM, "Ottavio Caruso" wrote: > Are you using the default chrome os kernel or is it a native one? > On Mar 31, 2013 10:09 PM, "Stuart Winter" wrote: > >> >> > Incidentally Eric mentions the Chromebook as one of the reasons to port >> the >> > hf. I did install a full rootfs of Armedslack (sf) on the Chromebook >> and it >> > did work. I then gave up because of lack of feedback and because my unit >> > was faulty. >> >> Yep I'm writing this using WindowMaker with Slackware ARM on the >> Chomebook. I just used the miniroot and some Chromebook tweaks. >> I'm planning on making Slackware ARM installable on the Chromebook - I've >> already started laying the groundwork. >> >> Slackware ARM will run on any 32-bit machine that's armv4 + thumb or >> greater. Slackware needs a hard float port as we're the only distro >> without one, and given that ARM is becoming more prevalent it's (IMO) >> important that Slackware is able to deliver max performance on the >> machines with a hardware floating point unit -- allowing them to be able >> to render multi media efficiently. >> >> That said, for general usage (non multi media) I really don't think >> there's much in it with soft vs hard float. Running Firefox and some >> other stuff on this Chromebook is *fast* even with the default Slackware >> ARM packages. >> >> -- >> Stuart Winter >> Slackware ARM: http://arm.slackware.com >> _______________________________________________ >> ARMedslack mailing list >> ARMedslack at lists.armedslack.org >> http://lists.armedslack.org/mailman/listinfo/armedslack >> > -------------- next part -------------- An HTML attachment was scrubbed... URL: From m-lists at biscuit.org.uk Sun Mar 31 21:32:48 2013 From: m-lists at biscuit.org.uk (Stuart Winter) Date: Sun, 31 Mar 2013 22:32:48 +0100 (BST) Subject: [ARMedslack] Quick question In-Reply-To: References: Message-ID: > Are you using the default chrome os kernel or is it a native one? The default one at the moment but if I make the work public, I'll have built my own 3.4.0 from the ChromeOS Git repo. I do keep my notes in the 'slackwarearm-devtools/chromebook' dir on the FTP site though. It's a bit random at the moment, but it'll start to take shape in the next few months when I have time to dedicate to it in May and August. If you have some good info, please let me know on list. The first thing I want to figure out how to get the installer working on it - which will involve getting WiFI into the installer and figuring some other logistical things out. For me, that's really the fun part of maintaining a port; I quite like hacking the installer to add ARM support. I may just keep it as a side project and not put it back into Slackware ARM though, as I'd recommend Eric's hard float port to make maximum use of the device. From ottavio2006-usenet2012 at yahoo.com Sun Mar 31 21:41:26 2013 From: ottavio2006-usenet2012 at yahoo.com (Ottavio Caruso) Date: Sun, 31 Mar 2013 22:41:26 +0100 Subject: [ARMedslack] Quick question In-Reply-To: References: Message-ID: The default method now is to untar a rootfs, there's no point using an installer. That is a remainder of the pc world. I did try it though and may recall it didn't work. -------------- next part -------------- An HTML attachment was scrubbed... URL: From m-lists at biscuit.org.uk Sun Mar 31 21:55:50 2013 From: m-lists at biscuit.org.uk (Stuart Winter) Date: Sun, 31 Mar 2013 22:55:50 +0100 (BST) Subject: [ARMedslack] Quick question In-Reply-To: References: Message-ID: On Sun, 31 Mar 2013, Ottavio Caruso wrote: > The default method now is to untar a rootfs, there's no point using an > installer. That is a remainder of the pc world. I did try it though and may > recall it didn't work. For me, the reasons to use the installer are: 1 - Someone has to provide rootfs's and if you want your users to be able to install a -current, you need to keep them up to date. Or if you don't update it, people will have to upgrade manually (or with a tool). But for over 10 years of doing this, I have realised that many people install a -current and *never* update it. Or they pick and choose packages to upgrade, and end up with a broken OS; then tell you it's all broken. You spend ages looking at it and realise it's just cause they didn't update properly ;-) If you can tell them "reinstall with the installer", and that's how you (as a developer do it), then you know that they will get a sane environment. 2 - If instead of providing rootfs's you supply scripts, you need to be cognisant that if your users has no ARM machine, then they cannot installpkg every package successfully since some packages chroot into the installation and execute the binaries. THis obviously will not result in a properly installed package if it's installed from anything other than the target architecture. This is why the Slackware ARM miniroots are built natively even though it'd be far faster to do on an x86. 3- You still have to do some OS configuration - and IMO, the installer is the best place to do this up front. I mean look at Fedora on the Trimslice - they provide a mini root but Slackware ARM is installed using the regular installer without any real effort. You just boot the installer as normal, install and reboot into the OS. From ottavio2006-usenet2012 at yahoo.com Sun Mar 31 22:03:37 2013 From: ottavio2006-usenet2012 at yahoo.com (Ottavio Caruso) Date: Sun, 31 Mar 2013 23:03:37 +0100 Subject: [ARMedslack] Quick question In-Reply-To: References: Message-ID: On the other hand if you release a new rootfs say every month, all that a smart user has to do is umount /home and untar the new rootfs over the old one and some post install cleanup. Alternatively they can be educated to update properly. On Mar 31, 2013 10:55 PM, "Stuart Winter" wrote: > On Sun, 31 Mar 2013, Ottavio Caruso wrote: > > > The default method now is to untar a rootfs, there's no point using an > > installer. That is a remainder of the pc world. I did try it though and > may > > recall it didn't work. > > For me, the reasons to use the installer are: > > 1 - Someone has to provide rootfs's and if you want your users to be able > to install a -current, you need to keep them up to date. > Or if you don't update it, people will have to upgrade manually (or with a > tool). But for over 10 years of doing this, I have realised that many > people install a -current and *never* update it. Or they pick and choose > packages to upgrade, and end up with a broken OS; then tell you it's all > broken. You spend ages looking at it and realise it's just cause they > didn't update properly ;-) > If you can tell them "reinstall with the installer", and that's how you > (as a developer do it), then you know that they will get a sane > environment. > > 2 - If instead of providing rootfs's you supply scripts, you need to be > cognisant that if your users has no ARM machine, then they cannot > installpkg every package successfully since some packages chroot into the > installation and execute the binaries. THis obviously will not result in > a properly installed package if it's installed from anything other than > the target architecture. This is why the Slackware ARM miniroots are > built natively even though it'd be far faster to do on an x86. > > 3- You still have to do some OS configuration - and IMO, the installer is > the best place to do this up front. > > I mean look at Fedora on the Trimslice - they provide a mini root but > Slackware ARM is installed using the regular installer without any real > effort. You just boot the installer as normal, install and reboot into the > OS. > _______________________________________________ > ARMedslack mailing list > ARMedslack at lists.armedslack.org > http://lists.armedslack.org/mailman/listinfo/armedslack > -------------- next part -------------- An HTML attachment was scrubbed... URL: From m-lists at biscuit.org.uk Sun Mar 31 22:18:04 2013 From: m-lists at biscuit.org.uk (Stuart Winter) Date: Sun, 31 Mar 2013 23:18:04 +0100 (BST) Subject: [ARMedslack] Quick question In-Reply-To: References: Message-ID: > On the other hand if you release a new rootfs say every month, all that a > smart user has to do is umount /home and untar the new rootfs over the old > one and some post install cleanup. Alternatively they can be educated to > update properly. Untarring would lead to cruft lying around which would be problematic. You'd also have to provide instructions about how to update. See MySQL->MariaDB for example - you'd have document the upgradepkg package name switch (upgradepkg oldpkg%newpkg) - otherwise someone'd have bits of both packages installed at once. THe Slackware ARM build machines are reinstalled every few days. For -current, an installer is ideal IME. As the developer, it's *so easy* to boot the installer and let it install, press ENTER a few times and reboot; as distinct from having to put in an SD Card, unpack this and that, load up an editor to edit fstab and so on. For stable releases there's probably not much in it though, to be honest - there'd be one rootfs for the release, and anything else ends up in /patches as usual. From stanley at stanleygarvey.com Sun Mar 31 22:24:34 2013 From: stanley at stanleygarvey.com (stanley garvey) Date: Sun, 31 Mar 2013 23:24:34 +0100 Subject: [ARMedslack] Quick question In-Reply-To: References: Message-ID: <20130331222434.BB2BC7BDE0FC2@bmail04.one.com> On Mar 31, 2013 23:03 "Ottavio Caruso" wrote: > On the other hand if you release a new rootfs say every month, all > that a smart user has to do is umount /home and untar the new rootfs > over the old one and some post install cleanup. Alternatively they can > be educated to update properly. > > > > 1) That would be a lot of work and .. > > 2) You assume the user is smart, Slackware is K.I.S.S. for a reason. > > 3) Education is not an exact science. > > > 4) Installers provide a uniform experience. > > > > > > -------------- next part -------------- An HTML attachment was scrubbed... URL: From ottavio2006-usenet2012 at yahoo.com Sun Mar 31 22:37:20 2013 From: ottavio2006-usenet2012 at yahoo.com (Ottavio Caruso) Date: Sun, 31 Mar 2013 23:37:20 +0100 Subject: [ARMedslack] Quick question In-Reply-To: <20130331222434.BB2BC7BDE0FC2@bmail04.one.com> References: <20130331222434.BB2BC7BDE0FC2@bmail04.one.com> Message-ID: An installer assumes that you have a removable media from which you can boot, which is not the case for most commercial devices on the market, as they are boot-locked. On Mar 31, 2013 11:24 PM, "stanley garvey" wrote: > > > On Mar 31, 2013 23:03 "Ottavio Caruso" wrote: > > On the other hand if you release a new rootfs say every month, all that a > smart user has to do is umount /home and untar the new rootfs over the old > one and some post install cleanup. Alternatively they can be educated to > update properly. > > > > 1) That would be a lot of work and .. > > 2) You assume the user is smart, Slackware is K.I.S.S. for a reason. > > 3) Education is not an exact science. > > 4) Installers provide a uniform experience. > > > > > _______________________________________________ > ARMedslack mailing list > ARMedslack at lists.armedslack.org > http://lists.armedslack.org/mailman/listinfo/armedslack > > -------------- next part -------------- An HTML attachment was scrubbed... URL: