Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
Erast Benson [EMAIL PROTECTED] wrote: If you are willing to participate in further SchilliX development, I would be happy to help your project too. virtually, all our distros based on the same core bits - sunw*. We are planning quite a bit of work there. So, all these beginnings, (i.e. SchiliX, BeliniX) are very good for overall progress. After all, thanks to GPL/CDDL. It forces/stimulates developers to exchange patches between the projects. Do you use the patches I provide at ftp://ftp.berlios.de/pub/schillix/patches/ We will launch pilot GNU/Solaris developer's program by the end of this week. So, if you didn't get invitation yet, send e-mail to me, and hold your breath a little longer. Thank you. I have enough work to do for the whole day, so I will not ask to force something... Jörg -- EMail:[EMAIL PROTECTED] (home) Jörg Schilling D-13353 Berlin [EMAIL PROTECTED](uni) [EMAIL PROTECTED](work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
Venky [EMAIL PROTECTED] wrote: The biggest compatibility problem of OpenSolaris (compared to Sun Solaris) is the fact that libm is not part of OpenSolaris. I agree. And now, there are (at least) two separate attempts to develop a replacement: one by you and one by Moinak for BeleniX. Do you think it makes sense to start a project for this and pool efforts? We could host the development repository on sourceforge. It'd have potential to become the de-facto libm for any OpenSolaris distribution and cut down some of the potential incompatibilities we'll eventually have to grapple with. if we like to host things, we should do it on a non-commercial platform like Berlios ;-) As already sayd: I need to clean up my sources in order to publish them. They currenly have a set of makefiles that work with smake, so they may be compiled on SchilliX out of the box. Jörg -- EMail:[EMAIL PROTECTED] (home) Jörg Schilling D-13353 Berlin [EMAIL PROTECTED](uni) [EMAIL PROTECTED](work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
On Wed, 2005-10-26 at 18:42 +0200, Joerg Schilling wrote: virtually, all our distros based on the same core bits - sunw*. We are planning quite a bit of work there. So, all these beginnings, (i.e. SchiliX, BeliniX) are very good for overall progress. After all, thanks to GPL/CDDL. It forces/stimulates developers to exchange patches between the projects. Do you use the patches I provide at ftp://ftp.berlios.de/pub/schillix/patches/ lseek.patch looks like we could take. But somehow sunw* stuff works for us as is or with little changes. GNU/Solaris actually will bring 1000+(literally) usability patches applied for GNU software all over the place. We derive them from Ubuntu/Breezy world mostly. Some patches came from Debian directly. See, for Debian GNU/Solaris, proportion is different. We have just 120 sunw* core packages and 3500+ Ubuntu/Debian packages. So, we mostly concentrating on Debian part right now. But we have quite a few plans for sunw* stuff in the near future. Erast ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
[EMAIL PROTECTED] wrote: Jake Maciejewski [EMAIL PROTECTED] wrote: Let us face reality... Belenix has been developped by starting with SchilliX and modifying it. Um ... I'd strongly object to this statement! It is correct that 2 ideas were taken from the earlier discussions on this list: * Using the math library from FreeBSD * Using the aperture driver as a replacement for xsvc Well, Belenix has just too much in common with SchilliX that is completely different to what Sun does on Sun Solaris to believe you. - The way to do the overall split of / and /usr - The fact that the CD-ROM is mounted on /.cdrom - The fact that subdirectories from /.cdrom are mounted to e.g. /usr via lofs. and some other minor stuff. It is most unlikely that you did never have a look at SchilliX and we all know, that knowing that/how something works makes things a lot easier for people who do it the second time and are not forced to develop everything from scratch as I did. Jörg -- EMail:[EMAIL PROTECTED] (home) Jörg Schilling D-13353 Berlin [EMAIL PROTECTED](uni) [EMAIL PROTECTED](work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
Does it matter? One of the great features of open source is being able to save time and derive from other people's work, whether it be implicitly or explicitly. Our nascent community should be helping and supporting each other - not arguing silly debates. cheers, steve Joerg Schilling wrote: [EMAIL PROTECTED] wrote: Jake Maciejewski [EMAIL PROTECTED] wrote: Let us face reality... Belenix has been developped by starting with SchilliX and modifying it. Um ... I'd strongly object to this statement! It is correct that 2 ideas were taken from the earlier discussions on this list: * Using the math library from FreeBSD * Using the aperture driver as a replacement for xsvc Well, Belenix has just too much in common with SchilliX that is completely different to what Sun does on Sun Solaris to believe you. - The way to do the overall split of / and /usr - The fact that the CD-ROM is mounted on /.cdrom - The fact that subdirectories from /.cdrom are mounted to e.g. /usr via lofs. and some other minor stuff. It is most unlikely that you did never have a look at SchilliX and we all know, that knowing that/how something works makes things a lot easier for people who do it the second time and are not forced to develop everything from scratch as I did. Jörg -- stephen lau // [EMAIL PROTECTED] | 650.786.0845 | http://whacked.net opensolaris // solaris kernel development ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
Erast Benson [EMAIL PROTECTED] wrote: let me clarify a little bit on what GNU/Solaris distro is. Thank you! The idea behind it is simple: do not re-invent the wheel and try to re-use existing 17000 high quality Debian packages, Debian infrastracture(read Dpkg, APT repositories, Debootstraps, installation program, utilities, developer's policy and so on) and Debian developer's if you will. It depends on how we define re-invent the wheel I try to avoid to re-invent the wheel for SchilliX by not using a different package system than the one used by Solaris and I don't try to replace standard UNIX tools by GNU clones. As a result of the missing pkg system, I need to wait and to create a simple intermediate method for packaging. In addition, I look at the quality of the debian packages and I see that Debian publishes a version of cdrtools that is broken because of the patches that are applied by Debian. For this reason, I believe that it may be similar with other tools and the only way to have guaranteed quality is to create your own compile environment for the free software you linke or need to package together with OpenSolaris. GNU/Solaris distribution uses OpenSolaris kernel and runtime(libc). So, it runs any existing Solaris software without modifications. In addition to that(and this is what differes it from SchiliX and BeliniX), it greatly simplifies porting effort for pure Linux applications and packages, since it provides real Debian environment. And how do you include these Debian packages? If you put them into /usr/bin, you will overwrite existing standard UNIX tools and then you will not be able to benefit from e.g. smf because you are forced to use the init process that Debian uses on Linux. If you would like to get guest password and participate in further development you could send me a request e-mail. If you are willing to participate in further SchilliX development, I would be happy to help your project too. Jörg -- EMail:[EMAIL PROTECTED] (home) Jörg Schilling D-13353 Berlin [EMAIL PROTECTED](uni) [EMAIL PROTECTED](work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
Stephen Lau [EMAIL PROTECTED] wrote: Does it matter? One of the great features of open source is being able to save time and derive from other people's work, whether it be implicitly or explicitly. Our nascent community should be helping and supporting each other - not arguing silly debates. Well sayd! I am still waiting for more people who contribute to SchilliX Jörg -- EMail:[EMAIL PROTECTED] (home) Jörg Schilling D-13353 Berlin [EMAIL PROTECTED](uni) [EMAIL PROTECTED](work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
Jörg, I have the kit and Schillix-0.2 so just need time to review. If someone already has X built for you then just tar up the binaries/libs and post them on the mirrors. I have tarballs of X 6.8.2 (CVS) compiled as well as in Solaris package formats. Take your pick. if you have a natively built one already then let us get it our there so we can test it and move on. If rebuilding Schillix with X included appeals, then better do it now with the new patches and files than wait. Schillix 0.2.1 ?!?! I beleive it should really be that simple with any of these distros. Anyhow, I'm looking about 4,000 GNU packages that can work universally with either Schillix and other distros. Possible? Mainly, to evaluate Sun Studio 10 in various build situations. So, once the debating is over... can we review the autobuild systems and how we can use Schillix to meet that goal?? ;o ~Ken Mays --- Joerg Schilling [EMAIL PROTECTED] wrote: ken mays [EMAIL PROTECTED] wrote: I have X.org 6.8.2 libraries, binaries, and header packages built for Solaris 8/9/10 (x86/SPARC) over at Blastwave. Could these work for you?!? I am not sure. The Blastwave packages itself do not work at all because the pkg* tools are missing on OpenSolaris and because most */reloc/* files do not have the right permissions in the archive. In addition, symlinks are often missing. This was e.g. a big problem when I tried to install the java compiler on Schillix (which is needed to compile OpenSolaris but which is not available as tar archive for OpenSolaris). If you like to unserstand how SchilliX is created from packages, please have a look at the SchilliX ISO kit. Jörg -- EMail:[EMAIL PROTECTED] (home) Jörg Schilling D-13353 Berlin [EMAIL PROTECTED] (uni) [EMAIL PROTECTED] (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily __ Yahoo! Mail - PC Magazine Editors' Choice 2005 http://mail.yahoo.com ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
ken mays [EMAIL PROTECTED] wrote: Jörg, I have the kit and Schillix-0.2 so just need time to review. If someone already has X built for you then just tar up the binaries/libs and post them on the mirrors. I have tarballs of X 6.8.2 (CVS) compiled as well as in Solaris package formats. Take your pick. if you have a natively built one already then let us get it our there so we can test it and move on. If rebuilding Schillix with X included appeals, then better do it now with the new patches and files than wait. X should compile on SchilliX-0.2 as I did pot some effort in libm and math.h Schillix 0.2.1 ?!?! I am currently compiling Build 25 and I expect SchilliX-0.2.1 to be ready soon :-) I beleive it should really be that simple with any of these distros. Anyhow, I'm looking about 4,000 GNU packages that can work universally with either Schillix and other distros. Possible? Mainly, to evaluate Sun Studio 10 in various build situations. As Studio 10 is fully C-99, it is easier to deal with various sources that use .member = val in struct init statements. So, once the debating is over... can we review the autobuild systems and how we can use Schillix to meet that goal?? ;o I did start the sps system as a wrapper for sources and did use it already for all the packages that are currently part of SchilliX. Jörg -- EMail:[EMAIL PROTECTED] (home) Jörg Schilling D-13353 Berlin [EMAIL PROTECTED](uni) [EMAIL PROTECTED](work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
[EMAIL PROTECTED] wrote: Let us face reality... Belenix has been developped by starting with SchilliX and modifying it. Um ... I'd strongly object to this statement! It is correct that 2 ideas were taken from the earlier discussions on this list: * Using the math library from FreeBSD * Using the aperture driver as a replacement for xsvc Well, Belenix has just too much in common with SchilliX that is completely different to what Sun does on Sun Solaris to believe you. - The way to do the overall split of / and /usr - The fact that the CD-ROM is mounted on /.cdrom - The fact that subdirectories from /.cdrom are mounted to e.g. /usr via lofs. and some other minor stuff. It is most unlikely that you did never have a look at SchilliX and we all know, that knowing that/how something works makes things a lot easier for people who do it the second time and are not forced to develop everything from scratch as I did. True. But these are just the ideas that you also get by by looking at any of the Linux LiveCDs from a high level. Translating from these ideas into a working self-hosting distro is the real challenge: * Just looking at a disro does not tell you how the Solaris New Boot works. One needs to understand this in order to create an OpenSolaris distro. How does one figure out the key to getting a rootfs on the ramdisk without knowing Newboot - the fact that ?bootpath? should be absent. * How do you find out the cd device node that contains the LiveCD early in the boot when devfsadm has not yet executed ? It took me two iterations to arrive at an optimal solution. First solution was suboptimal till I wrote a small utility using libdevinfo to list the cd nodes and then use fstyp on them to get the volume id. * How do you handle the Math library requirements ? I ported and enhanced the FreeBSD math libary which was more than a week's effort for me. SchilliX did not make the Math library sources available. * What files do you include in the miniroot ? Optimise space. * What are the correct options to pass to mkisofs to get the desired bootable cd ? * How do you handle the inetd failures ? Inetd has a dependency on ksh and it took me several days of hacking to identify that the same functionality exists in /usr/xpg4/bin/sh * I had to write the missing sysidtool service manifests and renamed them to sysonfigure. * I had to preconfigure additional open-source drivers like network drivers from Masayuki. Which files should you change to add the drivers by hand ? * One can easily configure dhcp on a interface but it is quite involved to write a proper /etc/dhcp/eventhook script that auto-configures DNS and NIS based on whichever is available. * I have actually picked ideas from KnoppiX like: - Giving user option to select the keyboard layout and subsequently using it for setting XKBLayout in Xorg as well. - Scanning all block devices and mounting any supported filesystems - recovery feature - Using physical swap if a Solaris partition is present on the harddisk How to identify the swap slice ? * Knowing the boot sequence in SMF using Dan Price and Eric Schrock's bootchart. * Fixing shutdown issues * Configuring many things in /etc to get a clean boot * Vold throws out the cdrom occasionally so I had to disable the cdrom module in vold. * Which files do you need to modify in OpenSolaris source to get the BeleniX name instead of Solaris ? I had to tackle all these and many more smaller issues too numerous to mention here. Obviously I had seen the SchilliX 0.1 in action and got a few ides from there. But there is a big difference in getting an idea and it's implementation. Building an entire distro from source code is a real challenge - I built everything in BeleniX from source code. I have written build and package metadata generation scripts. Nothing apart from the SchilliX iso was available when I started. Now that you have made the ISO kit available, it will be easy to create SchilliX derivatives. People can look at your scripts and quickly know how things are done. So I object to the term: Belenix has been developped by starting with SchilliX and modifying it. I did not modify SchilliX. I just picked a few ideas from it as I picked from KnoppiX and built everything else on my own. That effort should not be underestimated. Having said that it is also pertinent to point out that your challenge was greater than mine since you are the pioneer, and started from a clean slate. Also as Stephen pointed out in another email, lets end the debate right now. This is how open-source works by sharing ideas and code. I came out with a distro because I had my own reasons and ideas. I do not think there is anything wrong with that. It will be great for OpenSolaris in
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
[EMAIL PROTECTED] wrote: Let us face reality... Belenix has been developped by starting with SchilliX and modifying it. Um ... I'd strongly object to this statement! It is correct that 2 ideas were taken from the earlier discussions on this list: * Using the math library from FreeBSD * Using the aperture driver as a replacement for xsvc Well, Belenix has just too much in common with SchilliX that is completely different to what Sun does on Sun Solaris to believe you. - The way to do the overall split of / and /usr - The fact that the CD-ROM is mounted on /.cdrom - The fact that subdirectories from /.cdrom are mounted to e.g. /usr via lofs. and some other minor stuff. It is most unlikely that you did never have a look at SchilliX and we all know, that knowing that/how something works makes things a lot easier for people who do it the second time and are not forced to develop everything from scratch as I did. True. But these are just the ideas that you also get by by looking at any of the Linux LiveCDs from a high level. Translating from these ideas into a working self-hosting distro is the real challenge: * Just looking at a disro does not tell you how the Solaris New Boot works. One needs to understand this in order to create an OpenSolaris distro. How does one figure out the key to getting a rootfs on the ramdisk without knowing Newboot - the fact that ?bootpath? should be absent. * How do you find out the cd device node that contains the LiveCD early in the boot when devfsadm has not yet executed ? It took me two iterations to arrive at an optimal solution. First solution was suboptimal till I wrote a small utility using libdevinfo to list the cd nodes and then use fstyp on them to get the volume id. * How do you handle the Math library requirements ? I ported and enhanced the FreeBSD math libary which was more than a week's effort for me. SchilliX did not make the Math library sources available. * What files do you include in the miniroot ? Optimise space. * What are the correct options to pass to mkisofs to get the desired bootable cd ? * How do you handle the inetd failures ? Inetd has a dependency on ksh and it took me several days of hacking to identify that the same functionality exists in /usr/xpg4/bin/sh * I had to write the missing sysidtool service manifests and renamed them to sysonfigure. * I had to preconfigure additional open-source drivers like network drivers from Masayuki. Which files should you change to add the drivers by hand ? * One can easily configure dhcp on a interface but it is quite involved to write a proper /etc/dhcp/eventhook script that auto-configures DNS and NIS based on whichever is available. * I have actually picked ideas from KnoppiX like: - Giving user option to select the keyboard layout and subsequently using it for setting XKBLayout in Xorg as well. - Scanning all block devices and mounting any supported filesystems - recovery feature - Using physical swap if a Solaris partition is present on the harddisk How to identify the swap slice ? * Knowing the boot sequence in SMF using Dan Price and Eric Schrock's bootchart. * Fixing shutdown issues * Configuring many things in /etc to get a clean boot * Vold throws out the cdrom occasionally so I had to disable the cdrom module in vold. * Which files do you need to modify in OpenSolaris source to get the BeleniX name instead of Solaris ? I had to tackle all these and many more smaller issues too numerous to mention here. Obviously I had seen the SchilliX 0.1 in action and got a few ides from there. But there is a big difference in getting an idea and it's implementation. Building an entire distro from source code is a real challenge - I built everything in BeleniX from source code. I have written build and package metadata generation scripts. Nothing apart from the SchilliX iso was available when I started. Now that you have made the ISO kit available, it will be easy to create SchilliX derivatives. People can look at your scripts and quickly know how things are done. So I object to the term: Belenix has been developped by starting with SchilliX and modifying it. I did not modify SchilliX. I just picked a few ideas from it as I picked from KnoppiX and built everything else on my own. That effort should not be underestimated. Having said that it is also pertinent to point out that your challenge was greater than mine since you are the pioneer, and started from a clean slate. Also as Stephen pointed out in another email, lets end the debate right now. This is how open-source works by sharing ideas and code. I came out with a distro because I had my own reasons and ideas. I do not think there is anything wrong with that. It will be great for OpenSolaris in
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
On Tue, 2005-10-25 at 15:20 +0200, Joerg Schilling wrote: Erast Benson [EMAIL PROTECTED] wrote: let me clarify a little bit on what GNU/Solaris distro is. Thank you! The idea behind it is simple: do not re-invent the wheel and try to re-use existing 17000 high quality Debian packages, Debian infrastracture(read Dpkg, APT repositories, Debootstraps, installation program, utilities, developer's policy and so on) and Debian developer's if you will. It depends on how we define re-invent the wheel I try to avoid to re-invent the wheel for SchilliX by not using a different package system than the one used by Solaris and I don't try to replace standard UNIX tools by GNU clones. Well, this is the idea behind of GNU/Solaris: to be as much GNU centric as possible, but do not break sunw* core. GNU/Solaris trying to find gold middle and trying to do it right. As a result of the missing pkg system, I need to wait and to create a simple intermediate method for packaging. In addition, I look at the quality of the debian packages and I see that Debian publishes a version of cdrtools that is broken because of the patches that are applied by Debian. For this reason, I believe that it may be similar with other tools and the only way to have guaranteed quality is to create your own compile environment for the free software you linke or need to package together with OpenSolaris. GNU/Solaris distribution uses OpenSolaris kernel and runtime(libc). So, it runs any existing Solaris software without modifications. In addition to that(and this is what differes it from SchiliX and BeliniX), it greatly simplifies porting effort for pure Linux applications and packages, since it provides real Debian environment. And how do you include these Debian packages? If you put them into /usr/bin, you will overwrite existing standard UNIX tools and then you will not be able to benefit from e.g. smf because you are forced to use the init process that Debian uses on Linux. If you would like to get guest password and participate in further development you could send me a request e-mail. If you are willing to participate in further SchilliX development, I would be happy to help your project too. virtually, all our distros based on the same core bits - sunw*. We are planning quite a bit of work there. So, all these beginnings, (i.e. SchiliX, BeliniX) are very good for overall progress. After all, thanks to GPL/CDDL. It forces/stimulates developers to exchange patches between the projects. We will launch pilot GNU/Solaris developer's program by the end of this week. So, if you didn't get invitation yet, send e-mail to me, and hold your breath a little longer. Thank you. Erast ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
1. XFCE 4.2.2 ? 2. Xorg 6.9 RC1 (CVS) - is this with the latest DRI drivers as well ? If so this, this is great! ~Ken M. --- [EMAIL PROTECTED] wrote: BTW BeleniX 0.2 is due out shortly and boots into a full Graphical XFce4 desktop on Xorg 6.9 RC1 (CVS). It has preliminary auto-configuration support that augments X -configure. Regards, Moinak. J?rg -- EMail:[EMAIL PROTECTED] (home) J?rg Schilling D-13353 Berlin [EMAIL PROTECTED] (uni) [EMAIL PROTECTED] (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org __ Start your day with Yahoo! - Make it your home page! http://www.yahoo.com/r/hs ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
Erast Benson and Joerg Schilling were discussing GNU/Solaris: GNU/Solaris distribution uses OpenSolaris kernel and runtime(libc). So, it runs any existing Solaris software without modifications. If you put them into /usr/bin, you will overwrite existing standard UNIX tools This points out some large differences in people's perceptions of what is a Solaris app? One perspective is a minimalist one, concerned with syscalls in libc; another is broader and takes into account the utilities and other commands that are part of the system. Still others focus on middle ware and libraries, Java, web services, etc... That is, runs any existing Solaris Software without modifications is more difficult to do than it is to say. It is safe to say that the ARC process at Sun spends much of its time ensuring that changes to the system don't negatively impact this area. The Solaris Binary Compatibility effort (see appcert(1)) grapples with this issue as well. Going forward with OpenSolaris distros, a simple expectation might be If it runs on Solaris AND it passes appcert(1), then it should also run on any Solaris Compatible system. (Noting that appcert focuses on shared libraries and does not address system()'d or exec()'d utilities...) -John ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
1. XFCE 4.2.2 ? XFce 4.2.0 . 2. Xorg 6.9 RC1 (CVS) - is this with the latest DRI drivers as well ? Yes. I did a cvs update of the tree several hours back. I have also included Firefox 1.5/ Thunderbird 1.5 Beta2, Gaim, XMMS, some of the libraries from Gnome 2.12 with Cairo support, Vim 6.4 etc. All of these built on BeleniX 0.2 using gcc 3.4. In fact I am sending this email from Firefox 1.5 running on BeleniX 0.2 which is running on my AMD laptop :) Regards, Moinak. If so this, this is great! ~Ken M. --- [EMAIL PROTECTED] wrote: BTW BeleniX 0.2 is due out shortly and boots into a full Graphical XFce4 desktop on Xorg 6.9 RC1 (CVS). It has preliminary auto-configuration support that augments X -configure. Regards, Moinak. J?rg -- EMail:[EMAIL PROTECTED] (home) J?rg Schilling D-13353 Berlin [EMAIL PROTECTED](uni) [EMAIL PROTECTED](work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org __ Start your day with Yahoo! - Make it your home page! http://www.yahoo.com/r/hs ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
Very valid point. Would be nice if all opensolaris-based distros could guarantee to run unmodified C binaries. There are quite a few ways to achive that. Erast On Tue, 2005-10-25 at 11:22 -0700, John Plocher wrote: Erast Benson and Joerg Schilling were discussing GNU/Solaris: GNU/Solaris distribution uses OpenSolaris kernel and runtime(libc). So, it runs any existing Solaris software without modifications. If you put them into /usr/bin, you will overwrite existing standard UNIX tools This points out some large differences in people's perceptions of what is a Solaris app? One perspective is a minimalist one, concerned with syscalls in libc; another is broader and takes into account the utilities and other commands that are part of the system. Still others focus on middle ware and libraries, Java, web services, etc... That is, runs any existing Solaris Software without modifications is more difficult to do than it is to say. It is safe to say that the ARC process at Sun spends much of its time ensuring that changes to the system don't negatively impact this area. The Solaris Binary Compatibility effort (see appcert(1)) grapples with this issue as well. Going forward with OpenSolaris distros, a simple expectation might be If it runs on Solaris AND it passes appcert(1), then it should also run on any Solaris Compatible system. (Noting that appcert focuses on shared libraries and does not address system()'d or exec()'d utilities...) -John ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
[EMAIL PROTECTED] wrote: and some other minor stuff. It is most unlikely that you did never have a look at SchilliX and we all know, that knowing that/how something works makes things a lot easier for people who do it the second time and are not forced to develop everything from scratch as I did. True. But these are just the ideas that you also get by by looking at any of the Linux LiveCDs from a high level. Translating from these ideas into a working self-hosting distro is the real challenge: From a short look at Belenix, it seems that SchilliX is much closer to a self hosting enviroment than Belenix. Everythig provided by SchilliX-0.1 (except the Sun OpenSolaris sources itself) has been compiled on SchilliX-0.1 and even Xorg compiles on SchilliX-0.2. The binaries found on Belenix have been compiled on Solaris 10 instead. * Just looking at a disro does not tell you how the Solaris New Boot works. One needs to understand this in order to create an OpenSolaris distro. How does one figure out the key to getting a rootfs on the ramdisk without knowing Newboot - the fact that ?bootpath? should be absent. The related facts have been discussed with a lot of details on the opensolaris mailing list and it was absolutely no problem for me to set up the boot environment for SchilliX. * How do you find out the cd device node that contains the LiveCD early in the boot when devfsadm has not yet executed ? It took me two iterations to arrive at an optimal solution. First solution was suboptimal till I wrote a small utility using libdevinfo to list the cd nodes and then use fstyp on them to get the volume id. SchilliX provides code to 'search' for the right CD, why didn't you ask? * How do you handle the Math library requirements ? I ported and enhanced the FreeBSD math libary which was more than a week's effort for me. SchilliX did not make the Math library sources available. Having a short look at the (probable) current state of your port of libm shows that you seem to be at approx. the same state as I have been after the first week. The libm you find on SchilliX-0.2 got one month's effort from me. Why did you start your unneeded enterprize? * What files do you include in the miniroot ? Optimise space. It is simple to find out whatyou need for a minimal version, it should not take more than 3 iterations. * What are the correct options to pass to mkisofs to get the desired bootable cd ? As the needed options are documented in detail in the mkisofs man page and the included README's and as even the GRUB documentation includes a chapter on mkisofs it should not take more than 30 minutes to find this out. * How do you handle the inetd failures ? Inetd has a dependency on ksh and it took me several days of hacking to identify that the same functionality exists in /usr/xpg4/bin/sh As I did write many mails on this problem in this mailing list in August, it was easier for you than for me. You had the advantage of not being on vacation after I did the introducing work on the bug. After I did recover from the 4600 mails in my mailbox after I did return from vacation, it took me 2 days to find the problem - you did not send any information on it meanwhile :-( * I had to write the missing sysidtool service manifests and renamed them to sysonfigure. I did the same in May, so you have not been forced to do this. * I had to preconfigure additional open-source drivers like network drivers from Masayuki. Which files should you change to add the drivers by hand ? See SchilliX ;-) SchilliX-0.1 includes all drivers except the one that has been created by Masayuki _after_ I found that SchilliX-0.1 did not work on some emulators and asked him for help. * One can easily configure dhcp on a interface but it is quite involved to write a proper /etc/dhcp/eventhook script that auto-configures DNS and NIS based on whichever is available. Schillix does not yet support this due to lack of time. SchilliX uses a statix intermediate solution that works too. * I have actually picked ideas from KnoppiX like: - Giving user option to select the keyboard layout and subsequently using it for setting XKBLayout in Xorg as well. As the Sun keyboard layout setup is broken, you only have two opportunities: - fix the Sun keyboard setup bugs (as I did) - or implement this select service. - Scanning all block devices and mounting any supported filesystems - recovery feature I am not sure if this is a good idea in special as Belenix does this without asking. I prefer to a boot method that does not touch the hard disk at all. - Using physical swap if a Solaris partition is present on the harddisk How to identify the swap slice ? See above. I don't like this being done without asking me.
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
John Plocher [EMAIL PROTECTED] wrote: If you put them into /usr/bin, you will overwrite existing standard UNIX tools This points out some large differences in people's perceptions of what is a Solaris app? One perspective is a minimalist one, concerned with syscalls in libc; another is broader and takes into account the utilities and other commands that are part of the system. Still others focus on middle ware and libraries, Java, web services, etc... The biggest compatibility problem of OpenSolaris (compared to Sun Solaris) is the fact that libm is not part of OpenSolaris. In case you don't know, it took me a full month already to work on FreeBSD's libm in order to be halfway compatible with Sun Solaris and I am not even shure about the effort that would be needed for a mostly 100% compatibility. I am definitely interested in a UNIX centric OpenSolaris distro. Do not expect people who work on Linux centric Open Solaris distros to put a similar amount of work into compatibility issues. That is, runs any existing Solaris Software without modifications is more difficult to do than it is to say. It is safe to say that the ARC process at Sun spends much of its time ensuring that changes to the system don't negatively impact this area. The Solaris Binary Compatibility effort (see appcert(1)) grapples with this issue as well. I am not sure whether appcert(1) is sufficient to check for the compatibility ussies we will see on typcal OpenSolaris dustros. Going forward with OpenSolaris distros, a simple expectation might be If it runs on Solaris AND it passes appcert(1), then it should also run on any Solaris Compatible system. (Noting that appcert focuses on shared libraries and does not address system()'d or exec()'d utilities...) So judging from above, do you consider libc to be unstable as it calls system()'d or exec()'d utilities to support basic Solaris features like smf? Jörg -- EMail:[EMAIL PROTECTED] (home) Jörg Schilling D-13353 Berlin [EMAIL PROTECTED](uni) [EMAIL PROTECTED](work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
On Tue, 2005-10-25 at 23:10 +0200, Joerg Schilling wrote: John Plocher [EMAIL PROTECTED] wrote: If you put them into /usr/bin, you will overwrite existing standard UNIX tools This points out some large differences in people's perceptions of what is a Solaris app? One perspective is a minimalist one, concerned with syscalls in libc; another is broader and takes into account the utilities and other commands that are part of the system. Still others focus on middle ware and libraries, Java, web services, etc... The biggest compatibility problem of OpenSolaris (compared to Sun Solaris) is the fact that libm is not part of OpenSolaris. In case you don't know, it took me a full month already to work on FreeBSD's libm in order to be halfway compatible with Sun Solaris and I am not even shure about the effort that would be needed for a mostly 100% compatibility. I am definitely interested in a UNIX centric OpenSolaris distro. Do not expect people who work on Linux centric Open Solaris distros to put a similar amount of work into compatibility issues. They will. Simply becase not everything in this world is open-source... Also I do not see big advantages of creating yet another copy of Solaris Express. SchiliX and BeliniX over time will offer thier own features. This will create incompatabilities anyways. We should not expect fully 100% compatability between distros. Compatability to some extent ... yes. This we could achive. We probably might need to create some sort of Solaris Distributions Foundation(SDF) which will control the spec similar to LSB. Erast ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
Erast Benson [EMAIL PROTECTED] wrote: The biggest compatibility problem of OpenSolaris (compared to Sun Solaris) is the fact that libm is not part of OpenSolaris. In case you don't know, it took me a full month already to work on FreeBSD's libm in order to be halfway compatible with Sun Solaris and I am not even shure about the effort that would be needed for a mostly 100% compatibility. I am definitely interested in a UNIX centric OpenSolaris distro. Do not expect people who work on Linux centric Open Solaris distros to put a similar amount of work into compatibility issues. They will. Simply becase not everything in this world is open-source... Then it seems that you have a different background than a typial Debian maintainer. Also I do not see big advantages of creating yet another copy of Solaris Express. SchiliX and BeliniX over time will offer thier own features. This will create incompatabilities anyways. SchilliX does not like to be another Solaris Express. Even now, there is the advantage that SchilliX is freely redistributable. Currently, SchilliX needs to add basic features, later I expect that SchilliX may even include new dirvers and software before Sun Solaris will do it. With respect to Ethernet drivers this is already true thanks to Masayuki Murayama. We should not expect fully 100% compatability between distros. Compatability to some extent ... yes. This we could achive. What we need to define is a basic compatibility set. What Sun People may not yet know is that there cannot be 100% compatibility with Sun Solaris as long as Sun Solaris is not 100% OpenSource. We probably might need to create some sort of Solaris Distributions Foundation(SDF) which will control the spec similar to LSB. We need to define a basic compatibility set and we need to define a way of dealing with incompatibilities that are caused by missing source availability. It seems that we are at the same place as we have been a year agio when I did write my postulates on OpenSolaris that did end with the creation of the CAB. Jörg -- EMail:[EMAIL PROTECTED] (home) Jörg Schilling D-13353 Berlin [EMAIL PROTECTED](uni) [EMAIL PROTECTED](work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
The biggest compatibility problem of OpenSolaris (compared to Sun Solaris) is the fact that libm is not part of OpenSolaris. I agree. And now, there are (at least) two separate attempts to develop a replacement: one by you and one by Moinak for BeleniX. Do you think it makes sense to start a project for this and pool efforts? We could host the development repository on sourceforge. It'd have potential to become the de-facto libm for any OpenSolaris distribution and cut down some of the potential incompatibilities we'll eventually have to grapple with. Venky. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
The roadmap hasn't been updated (I'll look into that), but libm is coming. Source is targeted for January, and in the interim, we hope to get binaries posted this week or next. Hope that helps. Bonnie Venky wrote On 10/25/05 16:40,: The biggest compatibility problem of OpenSolaris (compared to Sun Solaris) is the fact that libm is not part of OpenSolaris. I agree. And now, there are (at least) two separate attempts to develop a replacement: one by you and one by Moinak for BeleniX. Do you think it makes sense to start a project for this and pool efforts? We could host the development repository on sourceforge. It'd have potential to become the de-facto libm for any OpenSolaris distribution and cut down some of the potential incompatibilities we'll eventually have to grapple with. Venky. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
Bonnie Corwin [EMAIL PROTECTED] wrote: The roadmap hasn't been updated (I'll look into that), but libm is coming. Source is targeted for January, and in the interim, we hope to get binaries posted this week or next. I was planning to pubish my libm source soon also (after I had the time for some cleanup ;-) Jörg -- EMail:[EMAIL PROTECTED] (home) Jörg Schilling D-13353 Berlin [EMAIL PROTECTED](uni) [EMAIL PROTECTED](work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
Source is targeted for January, and in the interim, we hope to get binaries posted this week or next. Great! Now that's one less issue to fight about! :) Venky. ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
Jake Maciejewski [EMAIL PROTECTED] wrote: I don't understand why X is a priority. You provide fluxbox as a window manager instead of a full desktop like GNOME, KDE, or XFCE. You don't provide GUI web browsers like Firefox and Konqueror, email apps like Thunderbird and Evolution, or any of the other software that's included in (or pre-packaged for) a typical Linux or BSD distro. Other than running those programs over X from a different machine, I fail to see what users are supposed to do with your X packages. If users are skilled enough to build the applications they want from source, they ought to be able to build Xorg too. By the time you release 0.3, GNUSolaris will be providing a complete desktop with GNOME, Firefox, GAIM, Evolution, etc. Let us face reality... Belenix has been developped by starting with SchilliX and modifying it. GNUSolaris is currently no more than an annunced distro. www.gnusolaris.org is unreachable and the announcement was not clear enough to understand what GNUSolaris will be. Once we know more about GNUSolaris, we will be able to judge based on it's real features. I would asume that they take the Solaris kernel and use the same Debian userland than Debian uses for Debian/Linux. If this is true, then it is still uncler whether they use glibc or the standard libc. If they use the Debian userland, GNUSolaris will most likely not what Solaris users expect and GNUSolaris will not pass an OpenSolaris compliance test If they use the Solaris userland, and only add other free software, GNUSolaris will be nothing different than SchilliX except that is less probable that it will pass an OpenSolaris compliance test and key features from Solaris (e.g. zones) will probably never work. With these constraints, it is obvious that there is a need to create an own X package for SchilliX in the near future. Jörg -- EMail:[EMAIL PROTECTED] (home) Jörg Schilling D-13353 Berlin [EMAIL PROTECTED](uni) [EMAIL PROTECTED](work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
Joerg, I have X.org 6.8.2 libraries, binaries, and header packages built for Solaris 8/9/10 (x86/SPARC) over at Blastwave. Could these work for you?!? I have Schillix 0.2 so do I need to know something to port X.org over to Schillix that you know of or is everything just a ./configure smake ?!? ;oP I can get this done tonite or tomorrow if I know the details. Maybe get GNOME 2.12.1 on it as well (maybe during Halloween timeframe). ~Ken Mays --- Joerg Schilling [EMAIL PROTECTED] wrote: Jake Maciejewski [EMAIL PROTECTED] wrote: I don't understand why X is a priority. You provide fluxbox as a window manager instead of a full desktop like GNOME, KDE, or XFCE. You don't provide GUI web browsers like Firefox and Konqueror, email apps like Thunderbird and Evolution, or any of the other software that's included in (or pre-packaged for) a typical Linux or BSD distro. Other than running those programs over X from a different machine, I fail to see what users are supposed to do with your X packages. If users are skilled enough to build the applications they want from source, they ought to be able to build Xorg too. By the time you release 0.3, GNUSolaris will be providing a complete desktop with GNOME, Firefox, GAIM, Evolution, etc. Let us face reality... Belenix has been developped by starting with SchilliX and modifying it. GNUSolaris is currently no more than an annunced distro. www.gnusolaris.org is unreachable and the announcement was not clear enough to understand what GNUSolaris will be. Once we know more about GNUSolaris, we will be able to judge based on it's real features. I would asume that they take the Solaris kernel and use the same Debian userland than Debian uses for Debian/Linux. If this is true, then it is still uncler whether they use glibc or the standard libc. If they use the Debian userland, GNUSolaris will most likely not what Solaris users expect and GNUSolaris will not pass an OpenSolaris compliance test If they use the Solaris userland, and only add other free software, GNUSolaris will be nothing different than SchilliX except that is less probable that it will pass an OpenSolaris compliance test and key features from Solaris (e.g. zones) will probably never work. With these constraints, it is obvious that there is a need to create an own X package for SchilliX in the near future. Jörg -- EMail:[EMAIL PROTECTED] (home) Jörg Schilling D-13353 Berlin [EMAIL PROTECTED] (uni) [EMAIL PROTECTED] (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org __ Start your day with Yahoo! - Make it your home page! http://www.yahoo.com/r/hs ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
ken mays wrote: Joerg, I have X.org 6.8.2 libraries, binaries, and header packages built for Solaris 8/9/10 (x86/SPARC) over at Blastwave. Could these work for you?!? I have Schillix 0.2 so do I need to know something to port X.org over to Schillix that you know of or is everything just a ./configure smake ?!? ;oP Won't be a ./configure unless you're going to build Xorg 7.0. (Xorg 6.x all use Imake instead of configure.) -- -Alan Coopersmith- [EMAIL PROTECTED] Sun Microsystems, Inc. - X Window System Engineering ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
Jake Maciejewski [EMAIL PROTECTED] wrote: Let us face reality... Belenix has been developped by starting with SchilliX and modifying it. Um ... I'd strongly object to this statement! It is correct that 2 ideas were taken from the earlier discussions on this list: * Using the math library from FreeBSD * Using the aperture driver as a replacement for xsvc But that is all there is to it. Apart from this I have NOT touched a single bit of work coming out of the SchilliX project. I did not base my work on the SchilliX binaries. When I started there was no documentation available on how to create an OpenSolaris LiveCD. No build tools or scripts were available that can generate an ISO image from OpenSolaris binaries. I have spent countless hours of effort starting with understanding x86 boot from scratch. So please do not state misconceptions. Because both the projects have, to some extent, similar aims and started from the same base they are bound to end up with similar approaches to solve the same issues, especially where options are limited. But to say that one project has depended on the other just because of these similatrities is wrong. In fact I started my work with the official Nevada builds that go into Solaris Express and built an initial ramdisk-only boot environment. Later on I moved to OpenSolaris. If one reads my blog one will immediately see to what extent I investigated the issues and how I arrived at solutions and subsequently improved them. GNUSolaris is currently no more than an annunced distro. www.gnusolaris.org is unreachable and the announcement was not clear enough to understand what GNUSolaris will be. Once we know more about GNUSolaris, we will be able to judge based on it's real features. I would asume that they take the Solaris kernel and use the same Debian userland than Debian uses for Debian/Linux. If this is true, then it is still uncler whether they use glibc or the standard libc. If they use the Debian userland, GNUSolaris will most likely not what Solaris users expect and GNUSolaris will not pass an OpenSolaris compliance test A GNU/Solaris distro is just that. It is not intended to be a Solaris compatible distro. There is a lot of interest in this sort of an environment as well. If they use the Solaris userland, and only add other free software, GNUSolaris will be nothing different than SchilliX except that is less probable that it will pass an OpenSolaris compliance test and key features from Solaris (e.g. zones) will probably never work. Zones can be made to work. Obviously even a GNU/Solaris distro cannot throw away core OpenSolaris features like Zones or SMF. One of the aims of BeleniX is to also provide the option of a GNU userland environment or even that of a GNU Zone. Regards, Moinak. With these constraints, it is obvious that there is a need to create an own X package for SchilliX in the near future. J?rg -- EMail:[EMAIL PROTECTED] (home) J?rg Schilling D-13353 Berlin [EMAIL PROTECTED] (uni) [EMAIL PROTECTED] (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
Re: [osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
Guys, let me clarify a little bit on what GNU/Solaris distro is. The idea behind it is simple: do not re-invent the wheel and try to re-use existing 17000 high quality Debian packages, Debian infrastracture(read Dpkg, APT repositories, Debootstraps, installation program, utilities, developer's policy and so on) and Debian developer's if you will. GNU/Solaris distribution uses OpenSolaris kernel and runtime(libc). So, it runs any existing Solaris software without modifications. In addition to that(and this is what differes it from SchiliX and BeliniX), it greatly simplifies porting effort for pure Linux applications and packages, since it provides real Debian environment. As you might know, Debian(as of today) is the engine for 30+ child projects. GNU/Solaris is practically based on latest Ubuntu/Breezy bits (except few imporant packages derived directly from Debian) and basically offers similar options for novice users. GNU/Solaris has been a private project for the last 8 months and will be publically available(including source code for everything) in early November, once we are done with more/less functional web-portal. We will probably enable early access for developers next week. If you would like to get guest password and participate in further development you could send me a request e-mail. Erast On Sun, 2005-10-23 at 14:07 -0600, [EMAIL PROTECTED] wrote: Jake Maciejewski [EMAIL PROTECTED] wrote: Let us face reality... Belenix has been developped by starting with SchilliX and modifying it. Um ... I'd strongly object to this statement! It is correct that 2 ideas were taken from the earlier discussions on this list: * Using the math library from FreeBSD * Using the aperture driver as a replacement for xsvc But that is all there is to it. Apart from this I have NOT touched a single bit of work coming out of the SchilliX project. I did not base my work on the SchilliX binaries. When I started there was no documentation available on how to create an OpenSolaris LiveCD. No build tools or scripts were available that can generate an ISO image from OpenSolaris binaries. I have spent countless hours of effort starting with understanding x86 boot from scratch. So please do not state misconceptions. Because both the projects have, to some extent, similar aims and started from the same base they are bound to end up with similar approaches to solve the same issues, especially where options are limited. But to say that one project has depended on the other just because of these similatrities is wrong. In fact I started my work with the official Nevada builds that go into Solaris Express and built an initial ramdisk-only boot environment. Later on I moved to OpenSolaris. If one reads my blog one will immediately see to what extent I investigated the issues and how I arrived at solutions and subsequently improved them. GNUSolaris is currently no more than an annunced distro. www.gnusolaris.org is unreachable and the announcement was not clear enough to understand what GNUSolaris will be. Once we know more about GNUSolaris, we will be able to judge based on it's real features. I would asume that they take the Solaris kernel and use the same Debian userland than Debian uses for Debian/Linux. If this is true, then it is still uncler whether they use glibc or the standard libc. If they use the Debian userland, GNUSolaris will most likely not what Solaris users expect and GNUSolaris will not pass an OpenSolaris compliance test A GNU/Solaris distro is just that. It is not intended to be a Solaris compatible distro. There is a lot of interest in this sort of an environment as well. If they use the Solaris userland, and only add other free software, GNUSolaris will be nothing different than SchilliX except that is less probable that it will pass an OpenSolaris compliance test and key features from Solaris (e.g. zones) will probably never work. Zones can be made to work. Obviously even a GNU/Solaris distro cannot throw away core OpenSolaris features like Zones or SMF. One of the aims of BeleniX is to also provide the option of a GNU userland environment or even that of a GNU Zone. Regards, Moinak. With these constraints, it is obvious that there is a need to create an own X package for SchilliX in the near future. J?rg -- EMail:[EMAIL PROTECTED] (home) J?rg Schilling D-13353 Berlin [EMAIL PROTECTED] (uni) [EMAIL PROTECTED] (work) Blog: http://schily.blogspot.com/ URL: http://cdrecord.berlios.de/old/private/ ftp://ftp.berlios.de/pub/schily ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
[osol-discuss] Re: Re: SchilliX-0.2 ready, but SchilliX (the project) needs help
I don't understand why X is a priority. You provide fluxbox as a window manager instead of a full desktop like GNOME, KDE, or XFCE. You don't provide GUI web browsers like Firefox and Konqueror, email apps like Thunderbird and Evolution, or any of the other software that's included in (or pre-packaged for) a typical Linux or BSD distro. Other than running those programs over X from a different machine, I fail to see what users are supposed to do with your X packages. If users are skilled enough to build the applications they want from source, they ought to be able to build Xorg too. By the time you release 0.3, GNUSolaris will be providing a complete desktop with GNOME, Firefox, GAIM, Evolution, etc. The problem I was referring to with that poorly phrased sentence is that the version number in /boot/grub/menu.lst.sample hasn't been updated. It won't prevent anything from working, but it should be fixed nevertheless. This message posted from opensolaris.org ___ opensolaris-discuss mailing list opensolaris-discuss@opensolaris.org
