The function signature, and return values for crypt(3c) did not change.
For salt generation, ie on password change, a new API crypt_gensalt(3c) was added. All places in Solaris (there was actually only 2 of them, one in pam_authtok_store.so the other in rpc.nispasswdd) were changed to use crypt_gensalt(3c) instead of manually generating the old to char salt.
Yes anything using PAM for authentication and password change works without change. We highly recommend the use of PAM rather than calling crypt(3c) directly.
In some of the cases it is obvious from the vendor. HP is likely HP/UX, Sun is Solaris, MS is Windows of some variant. The abiquous one is IBM it could be AIX, Solaris, Windows or something else.
Yes Solaris does have it. When I integrated it into Solaris (along with help from Alec Muffett & Casper Dik) we made it a pluggable framework by building on the data format you created - thank-you for having the foresight to use the $ char as a "magic" switch.
Details of the pluggable framework with a policy that allows for automigration on password change are available here: crypt.conf and policy.conf
Solaris source code is NOT needed to add new algorithms all code that uses crypt(3c) should continue to work. Code that changes passwords should use crypt_gensalt(3c) to create the salt.
Yes it does. As long as the hosts in the nameservice domain are running Solaris 9 12/02 or higher. It works with files, NIS(YP), NIS+ and LDAP.
I had to change rpc.ypasswdd and rpc.nispasswdd to make this work (both made assumptions about the length of the string returned fron crypt(3c)) so the NIS/NIS+ master must also be upgraded to Solaris 9 12/02 or higher.
Note that Solaris 9 12/02 also has a 3rd module that was written by Alec Muffett (of crack fame) and myself: crypt_sunmd5.so. It is believed to be more secure than the crypt_bsdmd5 module that is compatible with BSD & Linux.
It is possible to add new algorithms by creating a shared object that implements just two functions:
In some cases the GPL and patents actualy achieve similar end goals. They both allow someone to publish work and let others use it that would otherwise have been kept a trade secret.
Personally I don't believe that the GPL and patents conflict. In particular you can give a license to use your patented technology freely as long as it is used in the GPL'd code. You can then also charge licenses for the patented technology for non GPL usage cases.
There are no PORTS of Solaris. There is only one Solaris source tree and more that 95% of that source is identical between SPARC and x86. Solaris 2.5.1 also ran on PowerPC.
The 5% difference is really all about boot and other very low level stuff, or it is the few bits of hand optimized code for things like MD5, 3DES etc.
Even the Fujitsu SPARC64 systems run the same Solaris.
Linux didn't invent or invovate X Windows. X Windows existed for years and years before the first line of code of the Linux kernel was ever written.
The fact that X Windows isn't compulsary isn't an inovation in Linux either other UNIX systems hadbeen doing it for years before Linux the first line of Linux code was written.
[ In case you wonder this flame was brought to you on a Linux system with X Windows where I choose to change settings of the default window manager ]
I hope that was meant to be funny. If not... No, but if you hit someone or damage their property as a result of answering the call they have recourse against you as a result of your driving without due care and attention.
It isn't actually necessary to have a specific law that bans driving and talking on a cell (mobile) phone because existing laws cover it anyway.
You should NOT answer or make calls when driving, you can not pay attention and it is NOT the same as talking to someone else who is in the car.
Sorry but you are wrong here, and btw you are arguing with someone who knows this technology very well and has full write access to the source for it. I've built and shipped patches when I was a sustaining engineer.
Solaris OS patches are sparse SVR4 pacakges (which are NOT zip files) with additional meta data. SVR4 packages on Solaris can contain compress cpio files or in some releases zip files.
Zip is not the meta data compoent of the patch or package system on Solaris. The patch could have been distributed as a tar.Z or tar.gz or tar.bz2 or cpio.Z etc etc, infact we have done serval of these in the past.
Zip is only used because it compresses well (in many cases better than tar.Z) and it is a single step to unpack the patch from the distribution mechanism.
Solaris packages can also be in stream format which means they are a single file (a bit more like an RPM).
Note that most (all?) Linux distributions do not do patches but instead restribute entire packages for that subsystem. There are pros and cons on both sides of this (particuarly to do with risk management of the changed components and for fixes that span package boundaries). The actual patch is what you get *after* you unpack the zip file in this case. Back when we shipped the whole patch collection on a single CD the patches were not contained in zip files.
Solaris has 2 versions of the Korn shell, not 3:/usr/bin/ksh as you say is ksh88 and/usr/dt/bin/dtksh is ksh93./usr/bin/sh is the original Bourne Shell.
The Solaris package system is not based on pkzip and never has been. It is the SVR4 pacakge system that has been extended by Sun. It can do many things you can't do in RPM. It has recently been extended to support installing pacakge streams over an https connection and has support for digitally signed patches.
Solaris is years and years ahead in resource management and massive scalability than Linux and BSD kernels. However you are correct that we took a long time to embrace some of the open soruce products. There is the Solaris Companion CD that includes Python, KDE, nmap... and many other popular open source things that we don't include in the base install.
There is a move with in Solaris engineering to make the base install smaller and smaller, the reason for this is security and easy of management (including patching). This is no different to what RedHat et al have done, sure you can install everything, but remember that everything on Solaris should include the companion CD before you do your comparsion of what is and isn't there.
Sun really only has one version of Solaris. It is built for two platforms, SPARC and x86. It is one source base and it doesn't matter if it is server or "desktop" or laptop or embedded use. There is only one Solaris.
Those in Scotland, Wales and Nothern Ireland pay the TV license as well!
The quality of stuff that is put up on the internet will probably not be broadcast quality. If PBS (or anyone else) wants to broadcast it, that is commerical use (and thus not covered) and it needs broadcast quality media.
Which is why you want to run it using OpenOffice.org or StarOffice. This ensures you get the slide content without the macro's running (because OOO doesn't understand Visual Basic).
There never was an effort to re-write Netscape Navigator in Java. Some version of Netscape Naviagtor had a Java VM embedded inside them so they could run Java Applets.
Galeon is just a GUI on top of Mozilla. Nautilus is the GNOME file manager and in current releases of GNOME it nolonger tries to be a web browser as well.
Lots of people inside and outside of Sun have discussed the "technical" (as well as legal) issues of using Linux drivers on a Solaris system. Both from the view point of a 3rd party doing it for themeselves and from the view of porting Linux drivers for inclusion in Solaris.
It isn't impossible but it is pretty difficult. Linux and Solaris both present themselves as "UNIX" to userland by the kernel services that they provide. Their kernel implementations differ a huge a mount.
The Solaris driver interface is very formal and very different to the interfaces used for adding drivers in Linux. Conceptually they are similar but the programmin interfaces are very different and not easy to port.
Other than the GPL nature of most of the Linux drivers there may be other legal issues. Some Linux drivers are reverse-engineered, this is fine for Linux but might not be fine for Sun to distribute as part of Solaris (even if the GPL wasn't an issue).
Actually you can loose a lot by someone taking a copy of your data. The data doesn't have to be an invention to be worth something, it could be your companies accounts or your confidential internal phone list etc etc.
I certainly don't believe that intellecutal property rights are pointless. Neither could anyone who believes in truely believes and understands what the GPL and other software licenses mean.
Slashdot Mirror
The function signature, and return values for crypt(3c) did not change.
For salt generation, ie on password change, a new API crypt_gensalt(3c) was added. All places in Solaris (there was actually only 2 of them, one in pam_authtok_store.so the other in rpc.nispasswdd) were changed to use crypt_gensalt(3c) instead of manually generating the old to char salt.
Yes anything using PAM for authentication and password change works without change. We highly recommend the use of PAM rather than calling crypt(3c) directly.
That should work just fine.
In some of the cases it is obvious from the vendor.
HP is likely HP/UX, Sun is Solaris, MS is Windows of some variant. The abiquous one is IBM it could be AIX, Solaris, Windows or something else.
http://docs.sun.com
Yes Solaris does have it. When I integrated it into Solaris (along with help from Alec Muffett & Casper Dik) we made it a pluggable framework by building on the data format you created - thank-you for having the foresight to use the $ char as a "magic" switch.
Details of the pluggable framework with a policy that allows for automigration on password change are available here: crypt.conf and policy.conf
Solaris source code is NOT needed to add new algorithms all code that uses crypt(3c) should continue to work. Code that changes passwords should use crypt_gensalt(3c) to create the salt.
Yes it does. As long as the hosts in the nameservice domain are running Solaris 9 12/02 or higher. It works with files, NIS(YP), NIS+ and LDAP.
I had to change rpc.ypasswdd and rpc.nispasswdd to make this work (both made assumptions about the length of the string returned fron crypt(3c)) so the NIS/NIS+ master must also be upgraded to Solaris 9 12/02 or higher.
Note that Solaris 9 12/02 also has a 3rd module that was written by Alec Muffett (of crack fame) and myself: crypt_sunmd5.so. It is believed to be more secure than the crypt_bsdmd5 module that is compatible with BSD & Linux.
It is possible to add new algorithms by creating a shared object that implements just two functions:
crypt_gensalt_impl and crypt_genhash_impl
It is impossible for "most admins" to implement Dtrace like functionality. Dtrace is very tightly integrated with the Solaris Kernel.
In some cases the GPL and patents actualy achieve similar end goals. They both allow someone to publish work and let others use it that would otherwise have been kept a trade secret.
Personally I don't believe that the GPL and patents conflict. In particular you can give a license to use your patented technology freely as long as it is used in the GPL'd code. You can then also charge licenses for the patented technology for non GPL usage cases.
Solaris has per user, group and project limits on many resources including virtual memory.
There are no PORTS of Solaris. There is only one Solaris source tree and more that 95% of that source is identical between SPARC and x86. Solaris 2.5.1 also ran on PowerPC.
The 5% difference is really all about boot and other very low level stuff, or it is the few bits of hand optimized code for things like MD5, 3DES etc.
Even the Fujitsu SPARC64 systems run the same Solaris.
Linux didn't invent or invovate X Windows. X Windows existed for years and years before the first line of code of the Linux kernel was ever written.
The fact that X Windows isn't compulsary isn't an inovation in Linux either other UNIX systems hadbeen doing it for years before Linux the first line of Linux code was written.
[ In case you wonder this flame was brought to you
on a Linux system with X Windows where I choose to
change settings of the default window manager ]
I hope that was meant to be funny. If not...
No, but if you hit someone or damage their property as a result of answering the call they have recourse against you as a result of your driving without due care and attention.
It isn't actually necessary to have a specific law that bans driving and talking on a cell (mobile) phone because existing laws cover it anyway.
You should NOT answer or make calls when driving, you can not pay attention and it is NOT the same as talking to someone else who is in the car.
Sorry but you are wrong here, and btw you are arguing with someone who knows this technology very well and has full write access to the source for it. I've built and shipped patches when I was a sustaining engineer.
Solaris OS patches are sparse SVR4 pacakges (which are NOT zip files) with additional meta data. SVR4 packages on Solaris can contain compress cpio files or in some releases zip files.
Zip is not the meta data compoent of the patch or package system on Solaris. The patch could have been distributed as a tar.Z or tar.gz or tar.bz2 or cpio.Z etc etc, infact we have done serval of these in the past.
Zip is only used because it compresses well (in many cases better than tar.Z) and it is a single step to unpack the patch from the distribution mechanism.
Solaris packages can also be in stream format which means they are a single file (a bit more like an RPM).
Note that most (all?) Linux distributions do not do patches but instead restribute entire packages for that subsystem. There are pros and cons on both sides of this (particuarly to do with risk management of the changed components and for fixes that span package boundaries).
The actual patch is what you get *after* you unpack the zip file in this case. Back when we shipped the whole patch collection on a single CD the patches were not contained in zip files.
Linux - no it is a kernel and doesn't have a copy buffer ;-)
Serious though, it depends on which distribution you use and what bits you have installed.
WindowMaker comes with a command to do this.
Openwindows (the XView stuff not the OLIT stuff) on SunOS came with a util to do this.
Standard MIT X Windows also comes with a cut and paste history which lets you "go both ways" between files and the "cut buffer(s)".
Solaris has 2 versions of the Korn shell, not 3: /usr/bin/ksh as you say is ksh88 and /usr/dt/bin/dtksh is ksh93. /usr/bin/sh is the original Bourne Shell.
The Solaris package system is not based on pkzip and never has been. It is the SVR4 pacakge system that has been extended by Sun. It can do many things you can't do in RPM. It has recently been extended to support installing pacakge streams over an https connection and has support for digitally signed patches.
Solaris is years and years ahead in resource management and massive scalability than Linux and BSD kernels. However you are correct that we took a long time to embrace some of the open soruce products. There is the Solaris Companion CD that includes Python, KDE, nmap... and many other popular open source things that we don't include in the base install.
There is a move with in Solaris engineering to make the base install smaller and smaller, the reason for this is security and easy of management (including patching). This is no different to what RedHat et al have done, sure you can install everything, but remember that everything on Solaris should include the companion CD before you do your comparsion of what is and isn't there.
For those living in the San Francisco Bay Area, Coopers can be found in Beverages and More and James Squire in Trader Joes.
Sun really only has one version of Solaris. It is built for two platforms, SPARC and x86. It is one source base and it doesn't matter if it is server or "desktop" or laptop or embedded use. There is only one Solaris.
s/English/British/g
Those in Scotland, Wales and Nothern Ireland pay the TV license as well!
The quality of stuff that is put up on the internet will probably not be broadcast quality. If PBS (or anyone else) wants to broadcast it, that is commerical use (and thus not covered) and it needs broadcast quality media.
Which is why you want to run it using OpenOffice.org or StarOffice. This ensures you get the slide content without the macro's running (because OOO doesn't understand Visual Basic).
There never was an effort to re-write Netscape Navigator in Java. Some version of Netscape Naviagtor had a Java VM embedded inside them so they could run Java Applets.
Galeon is just a GUI on top of Mozilla. Nautilus is the GNOME file manager and in current releases of GNOME it nolonger tries to be a web browser as well.
Lots of people inside and outside of Sun have discussed the "technical" (as well as legal) issues of using Linux drivers on a Solaris system. Both from the view point of a 3rd party doing it for themeselves and from the view of porting Linux drivers for inclusion in Solaris.
It isn't impossible but it is pretty difficult. Linux and Solaris both present themselves as "UNIX" to userland by the kernel services that they provide. Their kernel implementations differ a huge a mount.
The Solaris driver interface is very formal and very different to the interfaces used for adding drivers in Linux. Conceptually they are similar but the programmin interfaces are very different and not easy to port.
Other than the GPL nature of most of the Linux drivers there may be other legal issues. Some Linux drivers are reverse-engineered, this is fine for Linux but might not be fine for Sun to distribute as part of Solaris (even if the GPL wasn't an issue).
The major reason is technical not legal.
Your IBM and Sun statements are back to front. Sun has an "in perpetuity" as you put it license, IBM I believe pays per license sold.
The Glasgow Haskell Compiler has a pretty complete POSIX library, bindings for X11, and sockets (the latter written by me back '94/95).
Actually you can loose a lot by someone taking a copy of your data. The data doesn't have to be an invention to be worth something, it could be your companies accounts or your confidential internal phone list etc etc.
I certainly don't believe that intellecutal property rights are pointless. Neither could anyone who believes in truely believes and understands what the GPL and other software licenses mean.