summaryrefslogtreecommitdiffstats
path: root/TODO
diff options
context:
space:
mode:
Diffstat (limited to 'TODO')
-rw-r--r--TODO294
1 files changed, 294 insertions, 0 deletions
diff --git a/TODO b/TODO
new file mode 100644
index 0000000..4124684
--- /dev/null
+++ b/TODO
@@ -0,0 +1,294 @@
+TODO
+====
+
+BUG: Mountconfig has no "enable/disable" in the context menu.
+
+BUG: Mountconfig: "enable/disable" can sometime be available for normal users.
+
+BUG: mountconfig: The disk details dialog (sizeview.py) doesn't show *unpartioned* free space.
+
+BUG: Live gamma changes didn't seem to work on one of the S3s.
+
+BUG: mountconfig. Authentication details in the SMSShareSelectDialog are sometimes not
+correctly used when browsing.
+
+
+
+TODO: Some S3 cards need to have the video ram specified, and some later model don't.
+ It would be good if we could tell the difference and only offer the ram pulldown
+ when strictly needed.
+
+
+TODO: Monitor type selection (CRT, lcd etc). needed for clone mode on ATI at least.
+
+TODO: When using the proprietary nVidia driver, choose between the nVidia AGP
+and kernel agpgart based on what is best the of the machine's chipset.
+
+TODO: The proprietary ATI drivers have a 'Option "BusType" "PCI"' thing which may
+of may not need to be set for PCI based ATI cards.
+
+TODO: 3D accel on the 9250 with open source drivers.
+
+
+Future TODO
+===========
+
+setup.py
+~~~~~~~~
+.
+
+userconfig
+~~~~~~~~~~
+.
+
+unixauthdb.py
+~~~~~~~~~~~~~
+* LDAP (post-1.0) in-progress
+* (others?) (post-1.0)
+
+serviceconfig
+~~~~~~~~~~~~~
+* Change os.system() to the calls Simon uses in some cases.
+* Some services are running, but not in /var/run, implement special treatment. :>
+* Remove commented line from /etc/shells in userconfig -> Modify -> Shell.
+
+mountconfig
+~~~~~~~~~~~
+* Add 'proper' GUIs for editing some of the more common FS types.
+ - NFS.
+* AttributeError when mounting Samba volume
+* Handle the 'managed' mount entry options. See http://www.die.net/doc/linux/man/man8/fstab-sync.8.html
+ This is stand on Mandriva 2005.
+* Implement "real" HAL backend.
+
+displayconfig
+~~~~~~~~~~~~~
+* Use HAL for fetching PCI and card info? alongside existing systems (ldetect)?
+
+
+
+Extra?
+~~~~~~
+* Swap/kernel config:
+ http://kerneltrap.org/node/view/3000
+ "To tune, simply echo a value from 0 to 100 onto /proc/sys/vm/swappiness."
+
+* Hardware detection info:
+
+ http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&threadm=lyit5wc14g.fsf%40leia.mandrakesoft.com&rnum=1&prev=/groups%3Fq%3Ddebian%2520ldetect%2520hardware%2520detection%26hl%3Den%26lr%3D%26ie%3DUTF-8%26sa%3DN%26tab%3Dwg
+
+----------------------------------------------------------------------------
+From: Pixel (pixel@mandrakesoft.com)
+Subject: Re: Why so many HW detection packages?
+Newsgroups: linux.debian.devel
+Date: 2002-05-09 17:20:07 PST
+
+On Fre, 26 Apr 2002, Petter Reinholdtsen wrote:
+
+[...]
+
+> The reson is that there are 3 hardware detection system:
+> - Mandrake (libdetect, old)
+
+truly libdetect is old and deprecated
+(harddrake (was lothar) used to use it)
+
+we (mandrake) are now mostly using ldetect & ldetect-lst
+
+AFAIK here are the various free software hardware databases:
+
+--------------------------------------------------------------------------------
+- pci ----------
+
+ - pci.ids (in pciutils)
+ maps vendor+device -> description
+ and vendor+device+subvendor+subdevice -> description
+ also has device classes names
+
+ - modules.pcimap (in kernel /lib/modules/2.4*/)
+ maps vendor+device -> module
+ and vendor+device+subvendor+subdevice -> module
+
+ - XFree's xf86PciInfo.h (in XFree's source: xc/programs/Xserver/hw/xfree86/common/xf86PciInfo.h)
+ maps vendor+device -> description
+
+ - RedHat's pcitable (in hwdata)
+ maps vendor+device -> module+description
+ and a few vendor+device+subvendor+subdevice -> module+description
+ when needed
+ module can also be "Card:xxxx" for XFree (using Cards, see below)
+
+ - Mandrake's pcitable (in ldetect-lst)
+ same format as RedHat's (except for a few syntactical changes since mandrake kept old RedHat's format)
+ module can also be
+ "Card:xxxx" for XFree (using Cards+, see below)
+ "Server:xxxx" for XFree3
+ "ISDN:xxxx" for hisax special parameters
+ "Bad:xxxx" for warning about unhandled devices (mainly winmodems)
+
+ - Mandrake old detect's pci.lst (in detect-lst)
+ maps vendor+device -> class+module+description
+
+ - Progeny's pci.lst (in discover-data)
+ same format as detect
+ maps vendor+device -> class+module+description
+ module can also be
+ "Server:XFree86(module)" for XFree4
+ "Server:XF86_xxx" for XFree3
+
+some comments:
+
+ - the "class" in pci.lst is not useful when "module" is given since
+ from the module name, one can have the "class"
+
+ - the subvendor+subdevice distinction is sometimes useful
+ (not very often though)
+
+ - hopefully one day modules.pcimap will be the reference :)
+ (except for XFree of course)
+
+tools using those databases:
+
+ - kudzu and anaconda are using pcitable from hwdata
+
+ - library ldetect accesses pcitable from ldetect-lst,
+ this library is used by DrakX and drakxtools.
+ Mandrake's patched kudzu uses pcitable from ldetect-lst
+
+ - discover uses pci.ids from discover-data
+
+ - i don't know if tools are using modules.pcimap
+
+- XFree --------------------
+
+ - XFree comes with Cards
+
+ - Redhat has its own version (in hwdata)
+
+ - Mandrake has its own version Cards+ & CardsNames (in ldetect-lst)
+ (mainly a merge of XF3 Cards and XF4 Cards)
+
+ - discover doesn't need it since it's precising the server name
+ (XF3) or the module name (XF4) (?)
+ this is usually enough (except if you want to propose the choice,
+ but who wants XF3 nowadays :)
+
+
+- usb --------------------
+
+ - usbutils's usb.ids
+ maps vendor+device -> description
+ also has device classes names and some more stuff
+
+ - modules.usbmap (in kernel /lib/modules/2.4*/)
+ maps vendor+device -> module (?)
+
+ - Mandrake's usbtable (in ldetect-lst)
+ maps vendor+device -> module+description
+ module can also be
+ "Mouse:xxxx" for mouse configuration (fed to mousedrake)
+ "Tablet:wacom" for wacom tablet configuration
+ "Flag:xxxx" for DrakX package choosing
+ "Floppy:normal"
+
+ - Progeny's usb.lst (in discover-data)
+ maps vendor+device -> class+module+description
+ (but current's version only have module=unknown, so what's it
+ for, why not usb.ids?)
+
+
+tools using those databases:
+
+ - library ldetect accesses usbtable from ldetect-lst
+ this library is used by DrakX and drakxtools
+
+ - i don't know if tools are using modules.usbmap
+
+- scanner --------------------
+
+ - ScannerDB (in ldetect-lst)
+ maps name -> driver+kind(usb,scsi,serial,parallel)+options+various
+ (i don't know much about it, i don't know if yves made it from
+ scratch or what. ask yduret@mandrakesoft.com for more)
+
+- isdn --------------------
+
+ - isdn.db (in ldetect-lst)
+ list of internet providers by country
+ -> phone number + domainname + dns1 (ip) + dns2
+
+- old or small databases ----------
+
+ - isa.lst (detect), isatable (ldetect-lst), modules.isapnpmap (kernel)
+ - pcmcia.lst (detect, discover-data), pcmciatable (ldetect-lst)
+ - modules.parportmap (kernel but empty?)
+ - modules.ieee1394map (kernel but empty?)
+
+--------------------------------------------------------------------------------
+
+There may be some errors, or some missing stuff, please correct me!
+
+I've written a tool to keep in sync with as many databases as
+possible. see merge2pcitable.pl in
+http://www.mandrakelinux.com/cgi-bin/cvsweb.cgi/soft/ldetect-lst/convert/
+
+Maybe some common mailing list could be set up to deal with this?
+
+*but* note that the database is quite kernel dependent.
+- our pcitable doesn't handle this nicely
+- redhat has "upgradelist" in hwdata to partly handle this
+- i know we handle some pbs via /lib/modutils/macros with things like
+"if `kernelversion` = 2.4", debian seems to have it in
+/etc/modutils/arch
+
+
+Once again, hopefully one day modules.pcimap and modules.usbmap will
+be the reference! :)
+
+
+
+[...]
+
+> Mandrake switched from libdetect to kudzu, afaik
+> (latest mandrake (8.2) version is using kudzu for HW detection).
+
+well, mandrake has many tools doing more or less the same thing (and
+alas, not exactly always the same thing): DrakX (during install),
+drakxtools (when called, after install), kudzu (at boot, usually
+calling a drakxtools)
+
+----------------------------------------------------------------------------
+
+
+List of hardware probing tools to use for displayconfig:
+--------------------------------------------------------
+
+
+[1] xvinfo - Print out X-Video extension adaptor information
+
+ xvinfo prints out the capabilities of any video adaptors
+ associated with the display
+ that are accesible through the X-Video extension.
+
+[2] xresprobe - Prints out resolutions, frequency and displaytype.
+ Doesn't work in all cases. Works via ddc, I guess.
+
+[3] ddcprobe - Uses VESA BIOS Extension
+ Detects VGA + OEM, modes (only set up modes?), vid mem (kudzu)
+
+[3] read-edid
+ get-edid|parse-edid prints out a "good-looking" Monitor Section for
+ xorg.conf, not reliable (failed on notebook)
+
+[4] ddcxinfo - prints out modelines, hsync and vsync (kudzu)
+
+[5] svgamodes - prints out supported video modes (kudzu)
+
+
+
+
+--
+Pixel
+programming languages addict http://merd.net/pixel/language-study/
+
+----------------------------------------------------------------------------