Jeremy, Great to see these new ideas. How about the ability to ssh to a machine while it's installed (instead of telnetting?). Also, could you please give another look at : https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=119328 https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=119329 https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=119331 They have to do with re-probing hardware after %pre (as to set up some hardware raid devices in %pre) and having anaconda prompt for network information during kickstart even if you are doing a CD install. Thanks, James James S. Martin, RHCE Contractor Administrative Office of the United States Courts Washington, DC (202) 502-2394 anaconda-devel-list-bounces@xxxxxxxxxx wrote on 05/25/2004 04:46:45 PM: > For the second trick, here's a list of a variety of things that are > being looked at/worked on within anaconda with sort of a time frame of > being done for Fedora Core 3. I've put some thoughts on > implementation for each as well, although that's just from my quick > thinking on them. If you're interested in working on one of them, feel > free to ask questions here. > > Note that none of these will definitely make it in... it entirely > depends on available time, etc > > * Support for installation via a USB network adapter > * Add a pruner of modules.usbmap similar to the one for > modules.pcimap > * Add appropriate modules to module-info > * Make sure BUS_USB is being included in NIC probe > * QUESTION: How do these persist on the installed system? Do things > "just work"? > * I've gotten a lot of this so that it should work with the current > devel tree, although it needs testing. > > * Wireless support during installation > * Add a screen to the loader to get wireless settings > * Showing this depends on whether iwconfig ioctls work or not > * Need: essid, wep > * Add support for setting these via kickstart > * QUESTION: Do we expose this in the second stage? I'm thinking > not, but willing to be convinced otherwise. > * This also is well under way (as of today), it's my current project. > > * Rework langsupport screen > * Current screen is non-sensical with what we actually do based on it. > * Should just show groups with langonly in the comps file using > those names. Allows translation, clearer and provides the obvious > way to do post-install > * Should allow language class to become much simpler > * Reworking lang-table in the process would be an added bonus > * Charlie Bennett (ccb@xxxxxxxxxx) is investigating this > > * Better support for driver disks from non-floppy sources > * Allow from USB keys, etc (aka, any disk really) > * If disk is partitioned, select partition and then give filename > * First pass can use hdinstall.c stuff to select partition + enter > the filename > * Nicer would be directory browser > > * New "migrate" upgrade option similar to what exists in OS X. > * Preserve changed %config, other files in /etc and /var > * Use versioned look-aside in /var > * Then, nuke system dirs and do a fresh install > * Migrate back hardware configuration > * Paul Nasrat (pnasrat@xxxxxxxxxx) is investigating this. > > * Partition resizing > * Often requested, tricky to do and support. > * UI thought would be to do it in editing an existing partition > * Needs parted fixes to handle new ext[23] features > > * Keyboard/mouse/monitor configuration > * We really don't need to ask about anything other than keyboard on > a GUI install. > * If they can configure depth, etc all in firstboot, why not have > them configure the monitor there too? > * Could go to just _copying_ the install x config file > * Brent Fox (bfox@xxxxxxxxxx) is investigating this > > * Firewall/security screen UI improvements > * Trusted devices goes away > * Trusted ports -> dialog. Add translations > * SELinux needs to be put in here somehow better than it currently is. > > * LVM support > * A few weird failure cases sometimes, need better error handling > * Are there new useful options to add? > > * RAID support > * Add support for RAID 6 - DONE > * Possibly add support for RAID 10? > * Better handling of "unsupported" pre-existing RAID levels > * Multipathing? > > * Handling of kickstart errors > * Currently, errors in a ks.cfg result in a traceback. We should > give a nice error dialog instead. > * Finish converting parser errors to raise KickstartError > * Catch error, show dialog with exception logged > * The big remaining difficulty is the kickstart config is parsed > before the interface exists which could lead to different tracebacks. > > * Internal logging functionality improvements > * Add log levels > * Don't display all of the debuginfo on tty3, but put in file, only > showing the most critical stuff on tty3 > > * Ask about VNC on a text mode install > * If we're unable to start X, ask if the user would like to do a VNC > install. > > * zfcp config support during installation > * Newer s390 are using zfcp > > * IPv6 support > * Both for during the install and setup for post-install > * Low priority with current IPv6 penetration > > * Other miscellaneous things to consider > * Traceback fingerprinting to make dupe hunting in bugzilla easier > * Some objects are opaque in tracebacks, fix those up > * Use LVM by default in auto-partitioning? > * Make disk filtering in driveDict more modular > * Driver disks for things that aren't scsi or network cards > > Longer term, I have some more radical changes in mind, but with other > time constraints, I don't think it's really reasonable to even consider > them for FC3. > > Jeremy > > > _______________________________________________ > Anaconda-devel-list mailing list > Anaconda-devel-list@xxxxxxxxxx > https://www.redhat.com/mailman/listinfo/anaconda-devel-list
