summaryrefslogtreecommitdiffstats
path: root/pyanaconda
Commit message (Collapse)AuthorAgeFilesLines
...
* Move mountpoint validation out of the add mountpoint dialog.David Lehman2012-09-261-45/+50
|
* Only run the storage sanity check if we've run autopart.David Lehman2012-09-261-1/+3
|
* Add a method to reset a device to its original state.David Lehman2012-09-261-0/+9
|
* Make a copy of the original format instead of just storing another ref.David Lehman2012-09-262-3/+4
| | | | | Changed attributes of device.format should not also be reflected in device.originalFormat. That's the whole point of saving the original.
* Reformatting effectively removes a device from an existing Root.David Lehman2012-09-261-1/+1
|
* Fix test for whether to create biosboot during autopart. (#853628)David Lehman2012-09-261-1/+16
|
* Close AddLayout dialog on double-clickVratislav Podzimek2012-09-262-1/+14
|
* Remove useless handler of Cancel button (AddLayout dialog)Vratislav Podzimek2012-09-262-4/+0
|
* Don't rely on having some month and year selected (#859185)Vratislav Podzimek2012-09-261-2/+16
| | | | | Especially the year set in HW clock (and thus as system date) may be out of our range of years, that can be selected in the combobox.
* Raise an error if bootDrive is invalidJesse Keating2012-09-251-1/+5
|
* Handle automated installsJesse Keating2012-09-251-0/+28
| | | | | | | | | | | | | This follows the same design as GUI. If the install is automated, and all the spokes are complete, proceed into the installation phase. However if there are incomplete spokes, let the user finish those spokes. As soon as all spokes are complete jump straight to the install phase. Note that any new spoke that has a background process that could make it not be "ready" immediately is going to have to be accounted for here, which is ugly. Perhaps a object attribute that is a list of potential threads to join would work?
* Handle errors from text storage executeJesse Keating2012-09-251-11/+18
| | | | | | We were just eating all errors, which is not correct. This will accept errors that might be fixable, but lets other errors bubble up and end the install. It also manipulates the ready status correctly.
* Fix ready and completed properties for text storageJesse Keating2012-09-251-7/+10
| | | | | This brings things more in-line with the GUI spoke. Also it imports the thread stuff only once rather than multiple times.
* Use ksdata to determine text password completenessJesse Keating2012-09-251-4/+7
| | | | | | | This is necessary for doing installs via kickstart where we've got the data even though we haven't gone through the spoke. Also handle the case where the account could be locked.
* Select a default environment (#858180).Chris Lumens2012-09-251-1/+12
|
* Remove compssort.py.Chris Lumens2012-09-252-70/+0
| | | | It's unused anymore and clutters my grep results.
* Don't attempt to catch and re-raise a SystemError from AnacondaThread.run.Chris Lumens2012-09-251-1/+1
| | | | | Python will detect a thread has raised a SystemError and throw it away, so what I was trying to do here doesn't actually work.
* Add a progress message for quitting the installer.Chris Lumens2012-09-253-0/+13
| | | | | We need this to quit when handling errors, since doing a sys.exit() from a thread only kills the thread and not the main program.
* GUI error handling dialogs need to be protected from threading deadlocks.Chris Lumens2012-09-251-19/+21
|
* Initialize gdk threading as well.Chris Lumens2012-09-251-0/+3
| | | | | | | This is required if we actually want to do UI stuff from more than one thread. We've avoided this in large part up to now by using the communication queues, but it's not going to be possible to avoid forever. Error reporting in particular requires this.
* Handle --ignoremissing in _applyYumSelections (#859021).Chris Lumens2012-09-253-60/+26
| | | | | | kickstart.selectPackages can go away entirely, since all it does is add stuff to ksdata and we've got that anyway. Having done that, the handling for --ignoremissing then needs to move into _applyYumSelections.
* Hide the ISO install source if you've nuked all possible drives (#858088).Chris Lumens2012-09-241-2/+5
| | | | | | | | This is a little touchy, but you first have to go to the source spoke. This will cause all widgets to be displayed. Then go to the destination spoke and choose to autopart all your drives. Then go back to the source spoke. We need to both set the ISO-related widgets to no_show_all and hide them due to the previous visit.
* Don't write out /etc/sysconfig/clock anymore (#859217).Chris Lumens2012-09-241-8/+0
|
* Index the exn mapping by string, not by object.Chris Lumens2012-09-241-21/+17
| | | | | | | When testing for membership of the NoSuchPackage/NoSuchGroup exceptions, they appear to be first converted with str(), resulting in the name of the group or package in question. There are bound to be more instances of this in the future. Thus, let's just go by the names of things.
* Don't write HOSTNAME=HOSTNAME=myhostnamehere (#859141).Chris Lumens2012-09-241-2/+0
|
* Close temp file before moving it (#858681)Vratislav Podzimek2012-09-211-3/+3
| | | | | When writing new NTP servers configuration the temporary file has to be closed before moving it to the right place.
* UEFI paths must include a leading backslash on some machines. (#856938)Peter Jones2012-09-191-1/+1
| | | | | Some machines require a leading backslash, some don't require it. None seem to break with it there.
* Read cmdline files from /run/installJesse Keating2012-09-191-2/+2
| | | | | We don't have full access to the initrd anymore, so we've had to move where we put the cmdline files. Read from the new location.
* Grab the proxy username from the correct text entry (#858536).Chris Lumens2012-09-191-1/+1
|
* Remove our use of scsi_wait_scan (#858393).Chris Lumens2012-09-191-5/+0
| | | | Turns out it's been broken for a while anyway.
* Change the keyboard shortcut for the updates checkbox.Chris Lumens2012-09-181-1/+1
| | | | Now that _D is in use for the Done button, this one has to change.
* Don't explicitly start the progress spinner in python code.Chris Lumens2012-09-181-3/+0
| | | | | This is handled by glade, and the fact that the spinner doesn't spin right now is due to a GTK bug.
* Move the progress bar back down to the bottom of the progress hub.Chris Lumens2012-09-181-1/+0
| | | | | This provides room to display the personalization on the same screen as the progress bar, which is what the mockups call for.
* Remove a bunch of stuff from the TODO list for a change.Chris Lumens2012-09-171-8/+0
|
* Move check of new partition size against format limits.David Lehman2012-09-171-14/+15
| | | | | | | Since the newui custom spoke creates all partitions with a base size of 1MB this check had to be moved to after allocation is complete since the base size is nearly meaningless in the context of these checks.
* Improve growth check when deciding where to allocate new partitions.David Lehman2012-09-171-5/+31
| | | | | | | | | | | In the case where the new partition is going to be the first logical partition on the disk (and therefore require creation of an extended partition) we were failing to correctly set things up to perform the growth calculation. This led to sub-optimal placement of partitions. This also fixes a small bug that caused us to jump to the new candidate free region even when it offered the same growth potential as the old/current one.
* Keep btrfs selectors' sizes in sync as volume size changes.David Lehman2012-09-171-1/+17
| | | | | | | Subvolume size is the same as the volume size. Adding a new subvolume means growing the volume, which requires us to update the other subvolumes' selectors' sizes to keep them all in sync (and prevent spurious size adjustment operations).
* Allow specification of a label for new swap space via custom ui.David Lehman2012-09-171-1/+2
|
* Don't allow stage2 as stage1 unless specified via location.David Lehman2012-09-171-1/+2
| | | | | This prevents automatic selection of stage2 as stage1, which is something we never want to happen.
* Remove reference to PartitioningWarning, which was removed last week (#875931).Chris Lumens2012-09-171-9/+0
|
* Add a way to test exception handlingVratislav Podzimek2012-09-171-0/+34
| | | | | | | | Instead of adding command line option throwing exception in one place of the code, let's use SIGURS1 as signal for raising exception in anaconda. This way, it is possible to raise exception in different phases of the installation and we can add code to prepare worst case scenario testing as much resolved bugs in exception handling as possible.
* Fix dumpState to work with the new python-meh (#856235)Vratislav Podzimek2012-09-171-9/+8
|
* Make sure the InstallOptionsNDialogs get the correct space labels.Chris Lumens2012-09-141-3/+0
| | | | | | The #2 and #3 dialogs already grab the correct labels so by trying to do so again in _set_free_space_labels, we're making sure to get the wrong thing and end up with a traceback.
* Get rid of the big pause going from the storage spoke back to the hub.Chris Lumens2012-09-141-10/+10
| | | | | It's all because doKickstartStorage and StorageChecker.run can take a long time, and they were being run synchronously.
* Make sure to set the default TZ in ksdata so the completed method works.Chris Lumens2012-09-141-2/+5
|
* Allow creation of biosboot and prepboot partitions in the custom spoke.David Lehman2012-09-133-9/+21
| | | | | Also switches the reformat check to use format names instead of types since the names are what we populate the fstype combo with.
* Hide removable disks containing install media from the custom spoke.David Lehman2012-09-131-0/+6
|
* Make the minimum size for custom spoke partitions 1MB.David Lehman2012-09-131-0/+4
|
* The return value of execWithRedirect is an integer.David Lehman2012-09-131-2/+2
|
* Only include following free space in partitions' max size.David Lehman2012-09-131-6/+14
| | | | Leading space isn't useful for growing a filesystem.
generated by cgit (git 2.34.1) at 2025-09-30 23:03:16 +0000