summaryrefslogtreecommitdiffstats
path: root/iw
Commit message (Collapse)AuthorAgeFilesLines
* Structure the repo layout so it matches final structure better and make isys ↵Martin Sivak2010-05-3135-11562/+0
| | | | | | a real Python package. Also updates the build and autotools stuff to work with the new structure
* Do not allow editing of extended partitions (#593754)Hans de Goede2010-05-211-3/+7
| | | | | | Editing extended partitions and setting mountpoints and the like for them is not a good idea. The only useful edit of extended partitions would be resizing of pre-existing ones and we don't support that.
* Check for sane mountpoint in raid dialog (#592185)Brian C. Lane2010-05-201-3/+9
| | | | | | | Resolves: rhbz#592185 Display a message to the user when the mountpoint name they have chosen is invalid. Skip checking when the format type has no mountpoint.
* Check for sane mountpoint in lvm dialog (#592185)Brian C. Lane2010-05-201-2/+9
| | | | | | | Resolves: rhbz#592185 Display a message to the user when the mountpoint name they have chosen is invalid. Skip checking when the format type has no mountpoint.
* Check for sane mountpoint in partition dialog (#592185)Brian C. Lane2010-05-201-3/+9
| | | | | | | Resolves: rhbz#592185 Display a message to the user when the mountpoint name they have chosen is invalid. Skip checking when the format type has no mountpoint.
* Don't drop encryption when re-editing new encrypted partitions. (#582888)David Lehman2010-05-191-1/+11
|
* Display unpartitioned disks in main partitioning gui. (#588637)David Lehman2010-05-191-0/+5
| | | | | Disks that are components of a biosraid array or multipath device will not be displayed.
* Pick up mountpoint for existing formats on encrypted LVs. (#587002)David Lehman2010-05-191-0/+6
|
* Fix of typo.Radek Vykydal2010-05-141-1/+1
|
* Fix two minor errors found by checkbot.Chris Lumens2010-05-141-1/+1
|
* gui: "_use anyway" mnemonic.Ales Kozumplik2010-05-141-1/+1
| | | | | I know Hans wants to remove the screen, but let's just make it less annoying until then.
* logging: remove references to the 'bootloaderadvanced' step.Ales Kozumplik2010-05-142-4/+0
| | | | originally removed in 2008 by 7ad65d9e30fd21d8c0356dc998ea97c579a73d53.
* Add the best package for this arch to the optional package selector (#591653).Chris Lumens2010-05-131-4/+13
| | | | | | | | Before, we were just having ayum.searchNevra return a list of packages and grab whatever happened to be first on that list. Sometimes it ends up being the x86-64 package, and sometimes it ends up being the i686 one instead. To be more deterministic, we should return the best package for the architecture we're running on always.
* Make sure we write out multipath.conf before discovery (#563242)Peter Jones2010-05-131-0/+13
| | | | | We need to make sure the config file is there every time, or else /sbin/multipath helpfully does nothing useful at all.
* And add a menu to the right hand side so you can see the new column.Chris Lumens2010-05-131-0/+1
| | | | Related: rhbz#560666
* Add support for wireless configuration using nm-c-e in stage 2 (#473803)Radek Vykydal2010-05-131-0/+56
|
* Disable [Configure Network] button if there are no net devs (#520146)Radek Vykydal2010-05-131-1/+4
|
* Add net device description into selection dialog (#520146)Radek Vykydal2010-05-131-4/+8
|
* Check preselected install network device as nm-controlled (#520146)Radek Vykydal2010-05-131-10/+7
|
* Enable networking in stage 2 using nm-c-e (#520146)Radek Vykydal2010-05-133-352/+149
| | | | | | | | | | | | | | | | | | | | Key patch of series. enableNetwork method is the root of all other changes in the patchset. The method is used both for network enabling in stage 2 triggered by need of network (repos, network storage, etc..) and [Configure Network] button What we select in stage 2 network UI is: - device to be activated (only in case of network enabling), this is controlled by ONBOOT ifcfg setting, note that devices actually activated depend on editing in nm-c-e - devices to be controlled by NM (and thus visible in nm-c-e), this is set by NM_CONTROLLED ifcfg setting A dialog for that is part of the patch, it needs some polishing. bringUp method had to be split into write and waitForConnection because in case of just config, without actually activating any device in nm-c-e, we don't want to wait for activation.
* Use ifcfg files via NetworkDevice in Network class (#520146)Radek Vykydal2010-05-131-1/+1
| | | | | | | | | | | | | | | | | Remove reading of some keys from NM via dbus, count on ifcfg file contents (written by stage 1, ks, or nm-c-e). This should also ensure support for ipv6 in stage2 using nm-c-e. Remove available() method - it is useless - HWADDR is read from ifcfg file, description is set in constructor, and Network class will be kept up-to-date (e.g. after configuring with nm-c-e) by explicit call to added update() method. readIfcfgContets became useless, we can use (now ifcfg-backed) NetworkDevices' method loadIfcfgFile instead. Hunk starting with if len(available_devices) > 0: is a noop, probably some leftover - removing.
* Use proper attribute instead of NetworkDevice 'DESC' hack (#520146)Radek Vykydal2010-05-131-1/+1
| | | | | Note: other occurences of 'DESC' usage are replaced in following patches.
* logging: remove all references to the "installtype" step.Ales Kozumplik2010-05-131-3/+1
| | | | | | | | | it has been virtually nonexistent for three years, see 9e1a444c46abcfc29c2e44ffd102fcc1eda6289e, 2d90bc12dcc682a7f0ff68e599cbb81a34a0b4a3 and 076c853f50c10203b45a552735f692b13a191bcb. I found out about this when auditing the rawhide logs today.
* Add "Serial Number" column to the right side of the cleardisks UI.Chris Lumens2010-05-121-0/+1
| | | | Related: rhbz#560666
* Remove raid clone option and code (#587036)Hans de Goede2010-05-112-328/+1
| | | | | | | | | | Remove the raid clone option from the create-storage menu, and also remove the underlying code. It is broken, and this has not been noticed until very recently as no-one uses it. So as discussed on irc it is better to just remove it, otherwise the next time we accidentally break it, it will likely once more remain broken for a long time. This also "fixes" RHEL-6 bug 589957.
* Make Format and Resize checkboxes mutually exclusive (#589977)David Cantrell2010-05-071-5/+11
| | | | | | | | | | In the Edit Partition dialog, disable Resize if Format is checked and disable Format if Resize is checked. Users cannot check both Format and Resize and have it perform a useful operation. Technically, it's possible, but our UI workflow in anaconda does not really allow an easy way to accomplish this. For now, prevent users from selecting both Format and Resize. If we want to let people shrink a filesystem and reformat it, we can get creative with that later.
* Use gtk consts instead of magic ints.Radek Vykydal2010-05-075-14/+14
|
* Don't clear bootloader radio selection on double click (#588771).Chris Lumens2010-05-052-1/+2
| | | | | | | | The problem here was that ClearDisksWindow._remove_clicked wasn't even getting called on double click. Instead, DeviceSelector._row_activated was getting called because it was the first registered signal handler in the chain. Adding a second handler appends to the chain. It does not remove previously added handlers.
* Revert commit 27a4c7df871744454d1ca8979a576f9f45c67189Hans de Goede2010-05-041-0/+4
| | | | | We want to have a sane default for the UTC checkbox so we need to fixup hasWindows() rather then removing it.
* Clean up wording for oversized LVs (#587459)David Cantrell2010-05-041-7/+9
|
* Remove broken hasWindows function from bootloader.py and its callersHans de Goede2010-05-041-4/+0
| | | | | | | | | | | | | | | | As pylint rightfully points out the hasWindows function from bootloader.py is broken, it tries to access bootloaderInfo.dosFilesystems, but bootloaderInfo refers to the bootloaderInfo class, not the module. This does not cause any tracebacks because it tries to access bootloaderInfo.dosFilesystems while iterating over bootloader.images, and bootloader.images has not been set up at this time (so its an empty dict). Thus hasWindows() will always return false, as there is no way to determine hasWindows at this point during the installation (we have not yet paritioned, so any exsting windows partitions may end up being removed), this patches fixes things by simply removing the hasWindows function and fixing up any callers of it.
* gui: gray out OK button while adding raid set (#587161).Ales Kozumplik2010-05-031-2/+34
| | | | | | | This has a tricky part: the checkbox toggle callback in our implementation (WideCheckList.toggled_item) is called *before* the actual toggle occurs. This has to be considered in the button-enabling code. An alternative would be to change/extend the toggle callback hook mechanism.
* Offer to format unformatted DASD devices (#560702)David Cantrell2010-04-281-1/+3
| | | | | | | | | | | | | | | | | | | | | | On s390, if you had a single DASD that needed dasdfmt run, you would get a ZeroDivisionError traceback because the self._totalCylinders value was zero. The problem was caused by the DASD.totalCylinders property and how it initialized. It would initialize itself on the first call, but since we'd already fired off a dasdfmt process for the device, the one running to capture the cylinder count couldn't get access to the device. Users with more than one DASD needing a dasdfmt would not see the traceback, but would have an incorrect total cylinder count, so 100% would be reached when n-1 devices had been formatted. Honor the exclusiveDisks list in the storage object and log the /dev/disk/by-path alias for the device when logging the info message about what device was found that was unformatted. [Originally worked up in November 2009, I've updated the patch a bit to work with our current code. Tested in text mode and gui mode. If you have any unformatted DASD devices, anaconda prompts you and asks if you want to format them.]
* Make it more clear what the purpose of the "Boot" column is (#584811).Chris Lumens2010-04-231-1/+1
| | | | | The purpose is to select which device gets the bootloader installed, not which device contains /boot.
* lvm: check resizing against format's targetSize (#580171).Ales Kozumplik2010-04-221-4/+4
| | | | | | | | lv.format.currentSize remains on the original value all the time. This means that when the user resizes the LV once, currentSize stays the same and the test for change in lvm_dialog_gui.py always succeeds every time the user reopens the LV edit dialog. This causes a traceback just below, in ActionResizeFormat.
* Restore storage.clearPartType after reset when backing out of GUI. (#559233)David Lehman2010-04-211-0/+4
| | | | | | | We unset storage.clearPartType so that all devices will be found during reset of storage, but we need to restore it afterwards since the user is headed to the cleardisks screen and not necessarily all the way back to the parttype screen where clearPartType would get set again.
* Make Cancel button the default for 'Weak Password' dialog (#582660)Brian C. Lane2010-04-211-0/+1
| | | | Resolves: rhbz#582660
* Set Create Storage focus to first active radio button (#582676)Brian C. Lane2010-04-211-0/+4
| | | | Resolves: rhbz#582676
* Don't make all devices on the boot device selector immutable (#583028).Chris Lumens2010-04-191-11/+16
| | | | | | I introduced the immutable column in DeviceSelector without making space for it in the cleardisks store. So, add a new column to the store and make all values in it False to indicate nothing's immutable.
* Fix various syntax errorsHans de Goede2010-04-151-1/+1
| | | | | Courtesey of pylint, this patch also fixes some things which are not syntax errors but still make pylint unhappy.
* Check for and offer to format unformatted DASD devices (#560702).David Cantrell2010-04-141-2/+2
| | | | | | In the storage filtering UI, we need to call the startup() method of storage.dasd.DASD so it will check for unformatted DASD devices and offer the user the ability to format them.
* filter_gui.py: fixup isProtected changes for biosraid and mpathHans de Goede2010-04-131-7/+10
| | | | | | | | | | | | | | | | | | | | | commit 8e62952a0e52d9ac5df781eb2f4ff5915208602b causes a number of issues for biosraid / mpath: 1) For biosraid it passes the raidset name to _active, but _active expects a udev info dict now 2) _active now takes the name to compare to exclusivedisks from the passed udev info dict, but for biosraid and mpath the name stored in exclusiveDisks is not the member udev name, but the name of the higher level device (Note the code before the above commit already got this wrong for the mpath case) This patch fixes this by: 1) Passing the fake udev info data for biosraid sets to _active and to _isProtected (it adds a sysfs_path entry to the fake data for this). This fixes both above problems for biosraid 2) Creating fake udev info data for mpath, setting the name of that to the name of the higher level device and using that as _Active and _isProtected argument. By also storing this info dict into to store, we can can also drop the special mpath case in getNext()
* Apply yet another translation patch (#573870).Chris Lumens2010-04-121-1/+1
|
* Fix some previously difficult-to-translate strings (#573870).Chris Lumens2010-04-121-1/+1
|
* Adjust the Makefiles to install anaconda to /usr/lib{,64}/python?.?.Chris Lumens2010-04-061-1/+2
| | | | You know, like anaconda's any other python module.
* Automatically select devices added via the "Add Advanced" button (#579051).Chris Lumens2010-04-051-3/+3
|
* Enable network if it is needed when repo is added in UI (#577803).Radek Vykydal2010-03-311-6/+13
| | | | | | | | | | | | | | Network has to be enabled before setting the repo up so that repo metadata can be checked (http/ftp case) or even before nfs filesystem is mounted (nfs case). The bug is regarding the cases where network has not been anabled yet and user is adding network-requiring repo, or modifying non-net-requiring repo to a requiring one (e.g. DVD install -> http install). It can be worked around (apart from restarting install with ks, or something that would bring up network earlier) by canceling the dialog, and checking or modifying existing (*if there is any*) network requiring repo which would bring up network enablement dialog.
* Show protected devices in the filter UI, but make them immutable (#568343).Chris Lumens2010-03-262-22/+47
| | | | | | | We don't want people to be able to deselect these devices, but we still want to show them. Not showing them leads to a variety of annoying hacks and has the unpleasant side effect of having previously hidden devices suddenly show up in later UI screens.
* Add originalFormat handling to editLVMLogicalVolume. (#576529)David Lehman2010-03-241-1/+3
| | | | Adds 'originalFormat' to lv dict, missing from fcd6820064d.
* Fix a cut&paste error that caused a traceback (#574743)David Lehman2010-03-241-1/+0
|
generated by cgit (git 2.34.1) at 2026-03-29 00:08:54 +0000