In the past, Samba 3 was available for Mandrake, but support for it was flaky.
As maintainer for Mandrake's samba packages, I take exception to that, considering 9.2 had samba-3.0.0 (granted, 3.0.0 had some isses) available in contrib, and parallel-installable, compiled against MIT kerberos-1.3.x, with mostly integrated smbldap-tools etc etc etc.
Anyway, packages that are 99% like those in 10.0 are also available on the samba mirrors, like: http://www.samba.org/samba/ftp/Binary_Packa ges/Man drake/RPMS/9.2/ http://www.samba.org/samba/ftp/Bi nary_Packages/Man drake/RPMS/9.1/
Note, they are also compiled to install in parallel so as not to mess up installations for people who might use the urpmi media for 2.2.8a packages and by accident get 3.0.x...
Anyway, you can install via urpmi (if you have 9.1/9.2 boxen):
# urpmi.addmedia samba-9.2 \ http://www.samba.org/samba/ftp/Binary_Packages/ Man drake/RPMS/9.2/ with hdlist-ldap.cz # urpmi samba3-server samba3-winbind
Hopefully I will get around to follow-ups of somedocumentation I did for samba-2.2.x which I think helped bring some cool features to Mandrake users.
Finally, there are also some nice additions (IMHO) to openldap (but one or two minor bugs that need to be fixed still...). Makes the whole LDAP+Samba PDC and/or NT domain migration almost painless...
BTW, this post seems to insist on putting a space between the n and d in Mandrake in the URLs... remove it if it makes it to the page...
So, by publically asserting that the GPL is invalid, SCO leaves themselves wide open to copyright infringement lawsuits.
No, nothing in the GPL states that anyone has to assert that the GPL is valid, or agree with it. All they have to do is abide by the conditions. If they abide by the conditions for distributing nmap, their right to do so is provided by the GPL.
If they violate those conditions with respect to nmap, *then* they may no longer distribute it.
There's a strong argument that SCO, by their actions of offering a paid license for their alleged intellectual property in Linux
I assume here you mean the Linux kernel, not other software which may run on Linux and they may have provided.
has violated the GPL.
No, they have violated the license agreement for the Linux kernel (hint: If I violate some proprietary software, nothing affects the license on the competitors software licensing, even if they use identical wording..).
That legally, ethically and morally terminates their right to the software under the GPL.
Ethical and moral aren't involved here, and legally, they only lose the right to distribute the software in question.
Satisfying the license agreement for the kernel (or not) has nothing to do with them satisfying the license agreement of other software.
If this were the case, it might be that people distributing nmap with this additional license restriction would lose their license to distribute the Linux kernel!
Sorry, but for the nmap author to do this, he has to show that SCO distributed nmap with additional restrictions.
All you have to do is select "printer sharing" on the sharing preference pane and the printer shows up on other people's machines. You may not like Macs, I really don't care if you do or not. But there's a lesson here.
And, all Eric had to do was select the already-visible printer on the remote machine, and make it the default.
CUPS provides this functionality (trivial sharing of printers, especially compared to lpd), but Eric isn't complaining about CUPS, he's complaining about something like redhat-config-printer, which is a totally different thing.
I've never had this problem with printerdrake...
Now, the reason he saw all the confusing options for types of printers, is precisely because he *thought* he had to configure something, when all along he didn't. I think Eric may just be too used to *having* to configure something, instead of just expecting it to work.
If he had just clicked the print button, or done lpstat -t, I am sure he would have found an available printer, without *any* configuration.
The fact that he saw confusing options should have prompted him to realise this, and that this dialog is only for people whose printers aren't detected/browsed automatically (and there's not much you can do about removing those options, considering CUPS has support for them all, and there's absolultely no way for CUPS or redhat-config-printer to know whether Eric has a printer with a JetDirect card, an embedded print server with an lpd daemon, a multi-function network device with smb support etc etc etc).
While I don't contest the stats in the article, I just wonder if web server stats are valid for "Fastest Growing Linux Distro", even if they are valid for "Fastest Growing Apache platform".
Unless you assume every corporate/SME file/print/authentication server and all desktops run publicly accessible webservers, this is a really bad metric...
IDC stats aren't much better either...
Of the > 30 machines running Mandrake that I have installed (ranging from corporate authentication servers to firewalls to laptops), only one has a publicly accessible web server.
I just wish now that *someone* would release a version of fedora core that includes support for mp3 and various popular video formats so that it would make a usable desktop for most people out of the box.
Of course, if you used Mandrake, the first (and NTFS ro support - for now) wouldn't be a problem, and the 2nd could be solved by the PLF with a:
# urpmi xine-win32
(been running 'urpmi --auto-select' via cron for over 18 months, most of the time with my most-used personal box running cooker..)
The thing that makes dpkg/deb 'easier' than rpms is that rpms have this horrible addiction to breaking and having horrible dependencies. Then it gets to be difficult/a pain in the ass.
And what precisely is different about deb's that prevent this? How does a deb package on it's own avoid problems RPMs have??
It doesn't.
There are a few things you can't do with RPMS that you can do with deb's (like suggested packages), and there are a few things you can't do with DEBs that you can with RPMS (AFAIK you can't do triggers with DEBs).
The differences you see between "DEBs" and "RPMs" may largely be due to the fact that you are comparing Redhat RPMs to DEBs. Redhat doesn't do library packages (like Mandrake and Debian do), so upgrading a library package can break all other software that uses that library (whereas on Mandrake you can have say libsasl.so.7 - in libsasl7 - and libsasl2.so.2 - in libsasl2 - installed simultaneously).
So, please don't make generic comparisons.
There is no fundamental difference between DEBs and RPMs regarding library packaging, just differences in the way different distros do it.
There will however always still be some library problems (like the notorious libpng2 vs libpng3 problem), but neither Debain nor Mandrake are better than the other in that respect... since it's a problem that could even affect source-based distros....
package management is much easier with debian than with an RPM system.
So I'm guessing you use dpkg to install all packages on your Debian box? What, you don't???
Just as I don't use rpm to install all RPM packages on my box, I use urpmi for 99.9% of them, I only use rpm when I want to revert a package to test scripts in an upgrade scenario for the packages I maintain.
Nope sorry Fedora is NOT the same as Cooker. Ever heard of Rawhide? Who is copying who again?
OK, so show me the list of packages in Rawhide that have been maintained by the community.
Show me the collaborative website that tracked the development of Rawhide.
Show me the public CVS for packaging.
Show me the public CVS (with community members who have commit rights) in the software specific to the distro.
All of this has been in place for Mandrake cooker for more than a year, many have been in place for more than 3 years.
Sorry, Rawhide wasn't anywhere near as community oriented as Cooker was, and Fedora still has a long way to go before they are as open.
The only distro that may be as close to being open to the common user is Debian... but not quite...
And yes, I maintain a number of packages in Mandrake main, and lots in contrib...
Second off Fedora releases go through a LOT of public testing unlike Rawhide and Mandrake Cooker.
There are hundreds of users who run Cooker as a desktop distro, and some even run it on (non-production) servers, plus there is an official 2-month testing cycle. But, you could find all of that out on the (aforementioned) wiki. The only issue is it's still not wide enough testing to catch bugs in software no-one else is shipping by default (or where the distro doesn't have enough resources to debug it, like Gentoo), such as the packet-writing patch, which is now safe for the rest of us *because* Mandrake shipped it in a disto many users used.
Contrast that with "The goal of the Fedora Project is to work with the Linux community to build a complete, general purpose operating system exclusively from free software.". Pretty dam big difference.
If you're comparing Cooker to Fedora releases, what do you compare the Mandrake Community release to? RHEL? At least Mandrake Community is better packaged... and then there will still be Mandrake Official, and the Corporate Server products for those who need longer support.
Sorry, but you're not comparing apples with apples here...
The ONLY difference between Fedora and Mandrake's new "community" product is the respective QA of each company and how long the releases are supported.
Yes, with Fedora you get 9 months, with Mandrake Community you get 3 months + whatever the official release gets (at least another 18 months for most releases).
Good Troll, but *Red Hat* is the one innovating here.
Gentoo is simple, one install per machine for life.
On Mandrake also, except it doesn't take a week to get a functional system.
Put this in your daily cron to keep the whole system up to date: emerge sync
urpmi.update
emerge -pvu world
urpmi --auto-select --auto
Then every morning you can see what new stuff you may want to update that day.
On Mandrake it's already updated for you, you don't have to wait the rest of the day for it to compile...
Look for new software with: emerge -s whatever
urpmq -y whatever (not quite the same, but there are other tools)
Remove software with: emerge -pvC whatever
urpme whatever
Unless you have and run exactly what chipset and compiler flags your "distro" based binarys are compiled for, your system will never be as fast as it can be.
Mandrake already ships with the optimisations that make the biggest difference... some optimisations for your "chipset" (I assume you mean CPU, there aren't many compile-time optimisations you can make for the motherboard chipset besides in the kernel...) cause instability (OpenOffice.org for instance doens't do -O3), and most software doesn't have support for MMX etc (only multimedia stuff), and in those cases there are conditionals in the packages, so it's a simple 'rpm --rebuild --with mmx mplayer*.src.rpm'.
This isn't so much a change in the way Mandrake operates, only in the way they publish the releases.
Mandrake already maintains internal trees of some of the releases. For instance, HP has a version of Mandrake 9.1, 9.1.2, which has all the fixes for 9.1 plus some customisations for HP.
So, now we're just seeing this externally....
You may also be old enough to remember Mandrake 7.0.1...
"For traditional file-based data, such as text documents, audio tracks, and video clips, WinFS is the new Windows file system. Typically, you will store the main data of a file, the file stream, as a file on an NTFS volume. However, whenever you call an API that changes or adds items with NTFS file stream parts, WinFS extracts the metadata from the stream and adds the metadata to the WinFS store. This metadata describes information about the stream, such as its path, plus any information that WinFS can extract from the stream. Depending on file contents, this metadata can be the author (of a document), the genre (of an audio file), keywords (from a PDF file), and more. WinFS synchronizes the NTFS-resident file stream and the WinFS-resident metadata. New Longhorn applications can also choose to store their file streams directly in WinFS. File streams can be accessed using the existing Win32 file system API or the new WinFS API."
So, it seems that the files themselves are still stored in a filesytem, it's only the metadata that is stored in "WinFS".
The data itself does *not* live as a blob in a giant database.
There are so many words for you, but Idiot seems to sum you up the best.
Nice, call people idiots based on your (most likely) incorrect interpretation of something you haven't seen, or (it seems) researched.
Re:PCLinuxOS - Mandrake done right
on
MandrakeSoft Roundup
·
· Score: 2, Insightful
1)pclinuxos is Mandrake 9.2 + updates + some PLF packages, + some updates from contrib + some custom packages by Texstar, generated using mklivecd, which is in Mandrake 9.2 contribs (and easy enough to use that there are already about 5 other Mandrake 9.2-based live CDs made with it).
2)Quicktime playback is only possible with the win32 codec, which: -is not open source (thus can not go in the Mandrake download version which must only consist of open-source software) -probably not commercially distributeable without a license -has other potential legal problems -trivially installable from PLF.
I am running Mandrake 9.2 with some packages from PLF, and I just double click on the Quicktime movies on my digital camera, and they play, no mess, no fuss.
3)The slowness of your machine is likely due to a miscofigured hostname, to test, try: $ time getent hosts `hostname`
If that takes more than about 2 seconds, that's your problem, you can probably fix it by running: # echo -e "127.0.0.1\t`hostname`" >>/etc/hosts
4)Never had printing problems with 9.2, quite a few printers, and a live CD I made based on 9.2 worked out-the-box with all printers I tested with. Your CUPS problem could be related to your problem above if it is indeed name resolution issues.
I wonder if you think about all the other people who contribute the thousands of packages available in Mandrake...
I have ordered the 9.0 DVD-only and 9.1 and 9.2 DVD+CDs-only packs from the store, and all 3 orders arrived within the quoted shipping time (which at 10 days is faster than Amazon for the same shipping cost), and this to South Africa...
Considering typical road vehicles have transmission systems with efficiencies of the order of 85%, I think the gains of this old idea (has been published in a number of journals, and is actually present in a prototype military transport vehicle in South Africa - but there you have a *lot* more space to use...) are overstated. You could possible achieve 15% max over a similar hybrid-electric, at the cost of higher purchase and maintenance cost (high efficiency/power dense electric motors are quite expensive...).
Most drivers simply push data to the right place and fiddle with registers in the right way. There is nothing the competitor wouldn't have already thought off. If youre competitor has to learn from your driver they are atleast two generations behind you and you have nothing to fear.
So why don't we have competitors (Linux kernel DRI developers etc) that have written drivers with similar 3d performance?
Why isn't there a competitor providing open-source drivers for cards in the same league?
Linux will get is name far faster by being accepted on the corporate desktop. There you don't need gaming performance, you don't need 3d performance.
So you're saying that Linux should never be used for (say) CAD/CAM work (for instance Pro\Engineer), 3d-modelling, 3d GIS visualisation, and many other sceintific uses of OpenGL?
Maybe you haven't noticed that Pro\Engineer + NVidia cards actually make one high-end business case for Linux on the desktop (especially with the deals HP is pushing for their Linux Workstations, which AFAIK come with NVidia Quadro cards).
But with MDK, you have not only to pay them money, but buy the retail pack to get this capability.
You aren't very clear here, but as far as I know, there will be two versions of MandrakeMove, one freely downloadable version without USB support, then the retail version with USB support, Realplayer, NVidia drivers, etc etc.
The retail version can be bought on it's own for $20 or with a 128MB USB storage device (that Mandrakesoft has tested with the release, and the only on they support, since the have had issues with some hardware) for $60.
The whole point of network account management is to easily allow the network administrators to have administrative rights on the machines on their network.
Most any LDAP authentication setup on a unix system will allow a uid=0 user defined only in LDAP to log in (and essentially be root). That's the whole point.
The problem that needs to be solved is making this both secure (preventing rogue DHCP/LDAP servers causing exploits) yet easy to set up.
One possible solution would be to require TLS and SSL certs signed by a (manually-installed) CA cert.
What was your proposal?/me wonders about Kerberos and DNS SRV records...
Why would you send a command to flush the write cache of a read-only device?
This was apparently to test if it was read-only or read-write (although there are other options, this was probably easier and should be safe according to the standards). So, maybe they should have tested if it was read-only first... oh wait...
Why were mandrake using a development driver in the "stable" kernel? Something wrong with the stable ATA/ATAPI drivers?
It's not a development driver, it's a patch which enables packet-writing, a feature that has been available on Windows (via software like Adaptec DirectCD) for a few years, and is standard in Windows XP. The patch has been in use for quite some time by a lot of users, but hasn't gone into a kernel in a large distro (ie one of the big 4).
And, BTW, one of the aspects of getting a new patch like this into the stock kernel is showing that it has been in a distribution kernel. So, someone had to ship it, unfortunately Mandrake happened to ship it before any of the other large distros.
But, maybe instead we should never let distros ship with features that aren't in the official kernel tree? Then we could stop RH shipping with NTPL, SuSE and Mandrake shipping with XFS, LVM etc. Then we could all run Debian stable, and all be missing the features we need.
Slashdot Mirror
# urpmi mklivecd
... and I haven't tested it on 10.0 ... it will most likely need an update for a 2.6 kernel ...).
...).
# mklivecd livecd.iso
(you should also read the --help of mklivecd
See http://livecd.berlios.de to see some examples of live CDs people have made (many more have been made than those listed
In the past, Samba 3 was available for Mandrake, but support for it was flaky.
a ges/Man drake/RPMS/9.2/i nary_Packages/Man drake/RPMS/9.1/
...
/ Man drake/RPMS/9.2/ with hdlist-ldap.cz
...). Makes the whole LDAP+Samba PDC and/or NT domain migration almost painless ...
... remove it if it makes it to the page ...
As maintainer for Mandrake's samba packages, I take exception to that, considering 9.2 had samba-3.0.0 (granted, 3.0.0 had some isses) available in contrib, and parallel-installable, compiled against MIT kerberos-1.3.x, with mostly integrated smbldap-tools etc etc etc.
Anyway, packages that are 99% like those in 10.0 are also available on the samba mirrors, like:
http://www.samba.org/samba/ftp/Binary_Pack
http://www.samba.org/samba/ftp/B
Note, they are also compiled to install in parallel so as not to mess up installations for people who might use the urpmi media for 2.2.8a packages and by accident get 3.0.x
Anyway, you can install via urpmi (if you have 9.1/9.2 boxen):
# urpmi.addmedia samba-9.2 \
http://www.samba.org/samba/ftp/Binary_Packages
# urpmi samba3-server samba3-winbind
Hopefully I will get around to follow-ups of some documentation I did for samba-2.2.x which I think helped bring some cool features to Mandrake users.
Finally, there are also some nice additions (IMHO) to openldap (but one or two minor bugs that need to be fixed still
BTW, this post seems to insist on putting a space between the n and d in Mandrake in the URLs
Unless you call Solaris useable without half of the software at sunfreeware.com
So, by publically asserting that the GPL is invalid, SCO leaves themselves wide open to copyright infringement lawsuits.
No, nothing in the GPL states that anyone has to assert that the GPL is valid, or agree with it. All they have to do is abide by the conditions. If they abide by the conditions for distributing nmap, their right to do so is provided by the GPL.
If they violate those conditions with respect to nmap, *then* they may no longer distribute it.
There's a strong argument that SCO, by their actions of offering a paid license for their alleged intellectual property in Linux
I assume here you mean the Linux kernel, not other software which may run on Linux and they may have provided.
has violated the GPL.
No, they have violated the license agreement for the Linux kernel (hint: If I violate some proprietary software, nothing affects the license on the competitors software licensing, even if they use identical wording
That legally, ethically and morally terminates their right to the software under the GPL.
Ethical and moral aren't involved here, and legally, they only lose the right to distribute the software in question.
Satisfying the license agreement for the kernel (or not) has nothing to do with them satisfying the license agreement of other software.
If this were the case, it might be that people distributing nmap with this additional license restriction would lose their license to distribute the Linux kernel!
Sorry, but for the nmap author to do this, he has to show that SCO distributed nmap with additional restrictions.
All you have to do is select "printer sharing" on the sharing preference pane and the printer shows up on other people's machines. You may not like Macs, I really don't care if you do or not. But there's a lesson here.
...
And, all Eric had to do was select the already-visible printer on the remote machine, and make it the default.
CUPS provides this functionality (trivial sharing of printers, especially compared to lpd), but Eric isn't complaining about CUPS, he's complaining about something like redhat-config-printer, which is a totally different thing.
I've never had this problem with printerdrake
Now, the reason he saw all the confusing options for types of printers, is precisely because he *thought* he had to configure something, when all along he didn't. I think Eric may just be too used to *having* to configure something, instead of just expecting it to work.
If he had just clicked the print button, or done lpstat -t, I am sure he would have found an available printer, without *any* configuration.
The fact that he saw confusing options should have prompted him to realise this, and that this dialog is only for people whose printers aren't detected/browsed automatically (and there's not much you can do about removing those options, considering CUPS has support for them all, and there's absolultely no way for CUPS or redhat-config-printer to know whether Eric has a printer with a JetDirect card, an embedded print server with an lpd daemon, a multi-function network device with smb support etc etc etc).
This was just a useless rant.
While I don't contest the stats in the article, I just wonder if web server stats are valid for "Fastest Growing Linux Distro", even if they are valid for "Fastest Growing Apache platform".
...
...
Unless you assume every corporate/SME file/print/authentication server and all desktops run publicly accessible webservers, this is a really bad metric
IDC stats aren't much better either
Of the > 30 machines running Mandrake that I have installed (ranging from corporate authentication servers to firewalls to laptops), only one has a publicly accessible web server.
Mandrake is a lot like RedHat, especially if you're used to downloading compiled RPMs and such.
;-).
And haven't heard of urpmi yet
I just wish now that *someone* would release a version of fedora core that includes support for mp3 and various popular video formats so that it would make a usable desktop for most people out of the box.
..)
Of course, if you used Mandrake, the first (and NTFS ro support - for now) wouldn't be a problem, and the 2nd could be solved by the PLF with a:
# urpmi xine-win32
(been running 'urpmi --auto-select' via cron for over 18 months, most of the time with my most-used personal box running cooker
The thing that makes dpkg/deb 'easier' than rpms is that rpms have this horrible addiction to breaking and having horrible dependencies. Then it gets to be difficult/a pain in the ass.
... since it's a problem that could even affect source-based distros ....
And what precisely is different about deb's that prevent this? How does a deb package on it's own avoid problems RPMs have??
It doesn't.
There are a few things you can't do with RPMS that you can do with deb's (like suggested packages), and there are a few things you can't do with DEBs that you can with RPMS (AFAIK you can't do triggers with DEBs).
The differences you see between "DEBs" and "RPMs" may largely be due to the fact that you are comparing Redhat RPMs to DEBs. Redhat doesn't do library packages (like Mandrake and Debian do), so upgrading a library package can break all other software that uses that library (whereas on Mandrake you can have say libsasl.so.7 - in libsasl7 - and libsasl2.so.2 - in libsasl2 - installed simultaneously).
So, please don't make generic comparisons.
There is no fundamental difference between DEBs and RPMs regarding library packaging, just differences in the way different distros do it.
There will however always still be some library problems (like the notorious libpng2 vs libpng3 problem), but neither Debain nor Mandrake are better than the other in that respect
... comparing apt to rpm?
package management is much easier with debian than with an RPM system.
So I'm guessing you use dpkg to install all packages on your Debian box? What, you don't???
Just as I don't use rpm to install all RPM packages on my box, I use urpmi for 99.9% of them, I only use rpm when I want to revert a package to test scripts in an upgrade scenario for the packages I maintain.
Nope sorry Fedora is NOT the same as Cooker. Ever heard of Rawhide? Who is copying who again?
... but not quite ...
...
... and then there will still be Mandrake Official, and the Corporate Server products for those who need longer support.
...
...
OK, so show me the list of packages in Rawhide that have been maintained by the community.
Show me the collaborative website that tracked the development of Rawhide.
Show me the public CVS for packaging.
Show me the public CVS (with community members who have commit rights) in the software specific to the distro.
All of this has been in place for Mandrake cooker for more than a year, many have been in place for more than 3 years.
Sorry, Rawhide wasn't anywhere near as community oriented as Cooker was, and Fedora still has a long way to go before they are as open.
The only distro that may be as close to being open to the common user is Debian
And yes, I maintain a number of packages in Mandrake main, and lots in contrib
Second off Fedora releases go through a LOT of public testing unlike Rawhide and Mandrake Cooker.
There are hundreds of users who run Cooker as a desktop distro, and some even run it on (non-production) servers, plus there is an official 2-month testing cycle. But, you could find all of that out on the (aforementioned) wiki. The only issue is it's still not wide enough testing to catch bugs in software no-one else is shipping by default (or where the distro doesn't have enough resources to debug it, like Gentoo), such as the packet-writing patch, which is now safe for the rest of us *because* Mandrake shipped it in a disto many users used.
Contrast that with "The goal of the Fedora Project is to work with the Linux community to build a complete, general purpose operating system exclusively from free software.". Pretty dam big difference.
If you're comparing Cooker to Fedora releases, what do you compare the Mandrake Community release to? RHEL? At least Mandrake Community is better packaged
Sorry, but you're not comparing apples with apples here
The ONLY difference between Fedora and Mandrake's new "community" product is the respective QA of each company and how long the releases are supported.
Yes, with Fedora you get 9 months, with Mandrake Community you get 3 months + whatever the official release gets (at least another 18 months for most releases).
Good Troll, but *Red Hat* is the one innovating here.
Right
Gentoo is simple, one install per machine for life.
...
... some optimisations for your "chipset" (I assume you mean CPU, there aren't many compile-time optimisations you can make for the motherboard chipset besides in the kernel ...) cause instability (OpenOffice.org for instance doens't do -O3), and most software doesn't have support for MMX etc (only multimedia stuff), and in those cases there are conditionals in the packages, so it's a simple 'rpm --rebuild --with mmx mplayer*.src.rpm'.
On Mandrake also, except it doesn't take a week to get a functional system.
Put this in your daily cron to keep the whole system up to date:
emerge sync
urpmi.update
emerge -pvu world
urpmi --auto-select --auto
Then every morning you can see what new stuff you may want to update that day.
On Mandrake it's already updated for you, you don't have to wait the rest of the day for it to compile
Look for new software with:
emerge -s whatever
urpmq -y whatever
(not quite the same, but there are other tools)
Remove software with:
emerge -pvC whatever
urpme whatever
Unless you have and run exactly what chipset and compiler flags your "distro" based binarys are compiled for, your system will never be as fast as it can be.
Mandrake already ships with the optimisations that make the biggest difference
This isn't so much a change in the way Mandrake operates, only in the way they publish the releases.
....
...
Mandrake already maintains internal trees of some of the releases. For instance, HP has a version of Mandrake 9.1, 9.1.2, which has all the fixes for 9.1 plus some customisations for HP.
So, now we're just seeing this externally
You may also be old enough to remember Mandrake 7.0.1
"For traditional file-based data, such as text documents, audio tracks, and video clips, WinFS is the new Windows file system. Typically, you will store the main data of a file, the file stream, as a file on an NTFS volume. However, whenever you call an API that changes or adds items with NTFS file stream parts, WinFS extracts the metadata from the stream and adds the metadata to the WinFS store. This metadata describes information about the stream, such as its path, plus any information that WinFS can extract from the stream. Depending on file contents, this metadata can be the author (of a document), the genre (of an audio file), keywords (from a PDF file), and more. WinFS synchronizes the NTFS-resident file stream and the WinFS-resident metadata. New Longhorn applications can also choose to store their file streams directly in WinFS. File streams can be accessed using the existing Win32 file system API or the new WinFS API."
So, it seems that the files themselves are still stored in a filesytem, it's only the metadata that is stored in "WinFS".
The data itself does *not* live as a blob in a giant database.
Nice, call people idiots based on your (most likely) incorrect interpretation of something you haven't seen, or (it seems) researched.
1)pclinuxos is Mandrake 9.2 + updates + some PLF packages, + some updates from contrib + some custom packages by Texstar, generated using mklivecd, which is in Mandrake 9.2 contribs (and easy enough to use that there are already about 5 other Mandrake 9.2-based live CDs made with it).
/etc/hosts
...
2)Quicktime playback is only possible with the win32 codec, which:
-is not open source (thus can not go in the Mandrake download version which must only consist of open-source software)
-probably not commercially distributeable without a license
-has other potential legal problems
-trivially installable from PLF.
I am running Mandrake 9.2 with some packages from PLF, and I just double click on the Quicktime movies on my digital camera, and they play, no mess, no fuss.
3)The slowness of your machine is likely due to a miscofigured hostname, to test, try:
$ time getent hosts `hostname`
If that takes more than about 2 seconds, that's your problem, you can probably fix it by running:
# echo -e "127.0.0.1\t`hostname`" >>
4)Never had printing problems with 9.2, quite a few printers, and a live CD I made based on 9.2 worked out-the-box with all printers I tested with. Your CUPS problem could be related to your problem above if it is indeed name resolution issues.
I wonder if you think about all the other people who contribute the thousands of packages available in Mandrake
I have ordered the 9.0 DVD-only and 9.1 and 9.2 DVD+CDs-only packs from the store, and all 3 orders arrived within the quoted shipping time (which at 10 days is faster than Amazon for the same shipping cost), and this to South Africa ...
Considering typical road vehicles have transmission systems with efficiencies of the order of 85%, I think the gains of this old idea (has been published in a number of journals, and is actually present in a prototype military transport vehicle in South Africa - but there you have a *lot* more space to use ...) are overstated. You could possible achieve 15% max over a similar hybrid-electric, at the cost of higher purchase and maintenance cost (high efficiency/power dense electric motors are quite expensive ...).
The Utah GLX nvidia driver wasn't the greatest but it did work with both 3D and 2D.
The 3d performance isn't acceptable for any real 3d work.
And you would sacrifice all that for slightly faster 3D graphics?
Sacrifice what? We're not. We're compromising until the open source drivers are up to scratch.
In our case, we either run the NVidia drivers on Linux, or we run Windows.
Your priorities are completely foreign to me.
So, you would prefer to run Windows then, a totally proprietary system?
Most drivers simply push data to the right place and fiddle with registers in the right way. There is nothing the competitor wouldn't have already thought off.
If youre competitor has to learn from your driver they are atleast two generations behind you and you have nothing to fear.
So why don't we have competitors (Linux kernel DRI developers etc) that have written drivers with similar 3d performance?
Why isn't there a competitor providing open-source drivers for cards in the same league?
Linux will get is name far faster by being accepted on the corporate desktop. There you don't need gaming performance, you don't need 3d performance.
So you're saying that Linux should never be used for (say) CAD/CAM work (for instance Pro\Engineer), 3d-modelling, 3d GIS visualisation, and many other sceintific uses of OpenGL?
Maybe you haven't noticed that Pro\Engineer + NVidia cards actually make one high-end business case for Linux on the desktop (especially with the deals HP is pushing for their Linux Workstations, which AFAIK come with NVidia Quadro cards).
But with MDK, you have not only to pay them money, but buy the retail pack to get this capability.
You aren't very clear here, but as far as I know, there will be two versions of MandrakeMove, one freely downloadable version without USB support, then the retail version with USB support, Realplayer, NVidia drivers, etc etc.
The retail version can be bought on it's own for $20 or with a 128MB USB storage device (that Mandrakesoft has tested with the release, and the only on they support, since the have had issues with some hardware) for $60.
Of course.
A seperate project related to Mandrake and Live CDs is livecd, which has some tools available for making Mandrake-based Live CDs.
mklivecd is a simple tool to generate a live CD, and it is included in Mandrake 9.2 contribs. Basically, you can do something like this to try it out:
# urpmi mklivecd
# mklivecd livecd.iso
There are some issues, which have been addressed in the CVS version.
Also new in CVS is a seperate minimal CD-to-HD installer.
I have made about 3 Live CDs based on Mandrake 9.2 using mklivecd.
The whole point of network account management is to easily allow the network administrators to have administrative rights on the machines on their network.
/me wonders about Kerberos and DNS SRV records ...
Most any LDAP authentication setup on a unix system will allow a uid=0 user defined only in LDAP to log in (and essentially be root). That's the whole point.
The problem that needs to be solved is making this both secure (preventing rogue DHCP/LDAP servers causing exploits) yet easy to set up.
One possible solution would be to require TLS and SSL certs signed by a (manually-installed) CA cert.
What was your proposal?
Why would you send a command to flush the write cache of a read-only device?
... oh wait ...
This was apparently to test if it was read-only or read-write (although there are other options, this was probably easier and should be safe according to the standards). So, maybe they should have tested if it was read-only first
Why were mandrake using a development driver in the "stable" kernel? Something wrong with the stable ATA/ATAPI drivers?
It's not a development driver, it's a patch which enables packet-writing, a feature that has been available on Windows (via software like Adaptec DirectCD) for a few years, and is standard in Windows XP. The patch has been in use for quite some time by a lot of users, but hasn't gone into a kernel in a large distro (ie one of the big 4).
And, BTW, one of the aspects of getting a new patch like this into the stock kernel is showing that it has been in a distribution kernel. So, someone had to ship it, unfortunately Mandrake happened to ship it before any of the other large distros.
But, maybe instead we should never let distros ship with features that aren't in the official kernel tree? Then we could stop RH shipping with NTPL, SuSE and Mandrake shipping with XFS, LVM etc. Then we could all run Debian stable, and all be missing the features we need.