Re: [RFC] Rewriting sade(8)

On Wed, Apr 07, 2010 at 10:49:35AM -0700, Randi Harper wrote:
> On Wed, Apr 7, 2010 at 9:35 AM, Andrey V. Elsukov <bu7cher_at_yandex.ru> wrote:
> > Hi, All.
> >
> > Some days ago i begun rewriting sade(8) to libgeom(3). Just for fun :-)
> > Today i have progress and you can see some screenshoots here:
> > http://butcher.heavennet.ru/sade/
> >
> > What is done:
> > 1. It's fully rewritten, but yes, i reuse some code from old sade.
> > 2. I wrote small framework "customdlg" for creating custom dialogs (not yet fully finished, but works).
> > 3. Now sade use libgeom for searching providers and it can use:
> >        {"DISK",        "disk device"},
> >        {"MD",          "memory backed virtual disk"},
> >        {"ZFS::ZVOL",   "ZFS volume"},
> >        {"MIRROR",      "GEOM based mirror (RAID1)"},
> >        {"STRIPE",      "GEOM based stripe (RAID0)"},
> >        {"RAID3",       "GEOM based RAID3 array"},
> >        {"CONCAT",      "GEOM based concatenated disk"},
> >        {"ELI",         "GEOM based encrypted disk"},
> >        {"JOURNAL",     "GEOM based journalled disk"},
> >        {"MULTIPATH",   "GEOM based disk with multiple path"}
> > 4. Access to partitions based on geom_part.so library. I made some changes in geom_part.c and
> > sent it to Marcel, but currently didn't receive answer. Any way if these changes is not acceptable
> > i can rewrite it for using libgeom(3) only.
> > 5. Undo/Commit features are based on internal implementation of geom class PART. I found there
> > two bugs:
> > http://lists.freebsd.org/pipermail/freebsd-geom/2010-April/004018.html
> > http://lists.freebsd.org/pipermail/freebsd-bugs/2010-April/039390.html
> > 6. All "device" related code can be placed in shared library and in future can be reused in another
> > frontend...
> > 7. New Partition Editor:
> > * partitions list now scrollable;
> > * all operations based on geom_part.so and similar to gpart(8);
> > * opening partition editor on empty provider opens new "Selech Partitioning Scheme" dialog
> > and create selected scheme;
> > * new add slice dialog;
> > * deleting slice from empty table destroys scheme;
> >
> > Todo:
> > 1. Refine customdlg.
> > 2. Waiting for fixes in geom_part_mbr, geom_part and geom_part.so.
> > 3. "Set Type and Label" dialog.
> > 4. Better wording needed for messages and help. My english is bad :(
> > ------
> > Missing functions:
> > Creating filesystem and save them into /etc/fstab.
> >
> > What new can be added:
> > 1. Creating ZFS pools and datasets.
> > 2. Creating GEOM-based raids.
> >
> > Any comments are welcome.
> > --
> > WBR, Andrey V. Elsukov
> > _______________________________________________
> > freebsd-current_at_freebsd.org mailing list
> > http://lists.freebsd.org/mailman/listinfo/freebsd-current
> > To unsubscribe, send any mail to "freebsd-current-unsubscribe_at_freebsd.org"
> >
> 
> Wow. This is awesome. patches? :D
> 
> I've been working on moving sysinstall from libdisk to libgeom, but
> unfortunately it's a bit more complicated (redoing the way we detect
> devices while I'm at it). I've done a lot of the heavy lifting code,
> but I haven't even started on the GUI parts yet. I'd love to see how
> someone else tackled doing this. I'm particularly interested in #5 &
> #7. :)
> 
> Generally, we try to keep sysinstall's disk tools and sade in sync, so
> I would like to work with you on this and see what we can come up
> with. I'm not entirely sure if #2 is a viable option since we already
> have functions in sysinstall that handle generating dialog boxes with
> libdialog, but if it's an improvement on what's existing, bring it on.
I have half-finished code that uses /sbin/gpart instead of libdisk or libgeom.
Half-finished, because it initialy used /sbin/{fidsk,bsdlabel,gpt}
So now I plan to finish the conversion to gpart and add work with gpt to
mbr and bsdlable handling. 
 
> CC:ing Devin Teske on this thread, as he's displayed interest in
> helping as well. My goal is to get all this in for 8.2, so there's
> quite a bit to work out in a short period of time.
> 
> -- randi


--
Adios