Send patches - preferably formatted by git format-patch - to patches at archlinux32 dot org.
summaryrefslogtreecommitdiff
path: root/archinstall
AgeCommit message (Collapse)Author
2021-03-29Added some debugging.Anton Hvornum
2021-03-29Reworking cache logic to not loose .encrypted flag on partitions after ↵Anton Hvornum
flushing cache.
2021-03-29Rolling back change to check parent blockdevice for safe formatting, it ↵Anton Hvornum
would wipe the boot partition if the boot partition for multiple drives was shared on the specific drive.
2021-03-29Removing a flush of the cache that was excessive.Anton Hvornum
2021-03-29Experimenting with cache, carrying over .allow_formattingAnton Hvornum
2021-03-29Experimenting with carrying over flags across cache flush. This to solve ↵Anton Hvornum
issues when flush is cleared and target mountpoints gets lost, making it impossible to do .find_mountpoint('/') later on for instance.
2021-03-29Allowing individual partitions safety checks to rely on the parent ↵Anton Hvornum
blockdevice keep_partitions status.
2021-03-29Added support for flushing a BlockDevice() cache after a partition has been ↵Anton Hvornum
formatted. This is to avoid internal confusion of which partitions are encrypted or not when going from Encrypted -> Un-encrypted setups.
2021-03-29Toned down the debugging a bit.Anton Hvornum
2021-03-29Moved some debugging to catch more variable settings.Anton Hvornum
2021-03-29Adding callstack to debug output for easier debugging. Removed hardcoded ↵Anton Hvornum
debug crash
2021-03-29DebuggingAnton Hvornum
2021-03-29Merge pull request #123 from Torxed/torxed-v2.2.3Anton Hvornum
Reverting .strip() logic and moving the check into the get_password()…
2021-03-29Reverting .strip() logic and moving the check into the get_password() function.Anton Hvornum
2021-03-29Fix spelling error in filesystem choiceDidr
Fixes a simple spelling error when the user is asked to select a filesystem. "Select which filesystem your main partition should use (by number **of** name)" should be **or**.
2021-03-29Fixing rare occation where partitions dissapear just before .format().Anton Hvornum
2021-03-29Fixing un-encrypted drives matching on block encrypted.Anton Hvornum
2021-03-28Added some more useful debug data to the log.Anton Hvornum
2021-03-28Adding debug to correct for a potential issue where partitions get marked as ↵Anton Hvornum
encrypted even tho \!disk-password was neven set.
2021-03-28Fixed a path-check issue with Time Zones.Anton Hvornum
2021-03-25New feature: Profile() now supports .packages which returns any defined ↵Anton Hvornum
packages for that specific profile, as well as archinstall.__packages__ contain any packages that Installer() is responsible for. This can be used to quickly gather any required packages and dependencies by archinstall. Not all profiles have it yet, so .packages might return None.
2021-03-24Fixes an issue with completely empty hard drivesAnton Hvornum
If empty hard drives are being set up for the first time, then `.has_content()`'s `mount` call will fail with `wrong fstype` since there's no filesystem yet. First step in this case is to check for that scenario, then check for content.
2021-03-24Removed vfat as an option for the root partitionAnton Hvornum
2021-03-23Also setting cwd (bad choice of name) to the workdir if one was specified. ↵Anton Hvornum
Which makes the trace.log to end up in the workdir as well.
2021-03-23Enabling archinstall.sys_command() to get a working directory when ↵Anton Hvornum
executing. The sys_command() is working pretty well for this very specific need, but this is an attempt to making it a bit more generic. A more general overhaul of the command should be done at some point.
2021-03-23Corrected recursion loop in log() calling log() before setting the new path ↵Anton Hvornum
for the log file on errors.
2021-03-23Partially corrects for #116. We still don't detect if we're running as root ↵Anton Hvornum
(need to investigate if we need to run as root first). But this should at least hot-swap the log-file to the current working directory and place the logfile there.
2021-03-21Improved english grammar and dialogues. Some 'questions' were formatted in a ↵Anton Hvornum
way where the user would enter 'yes' instead of the expected input. For instance, 'Any additional users to install:' which is a question where 'yes' is a appropriate response, but the expected input was the username to be created. Rephrased it to 'Enter a username to create a additional user:' instead for instance.
2021-03-21Added a mkinitcpio configuration for non-btrfs volumes but where encryption ↵Anton Hvornum
is still used. Also added keymap to the HOOKS so that we can get non-us layouts depending on the locale chosen during installation.
2021-03-21Removed some debugging output.Anton Hvornum
2021-03-21Added an option to force-flush a blockdevice cache. creating a new ↵Anton Hvornum
Filesystem() now triggers a flush of the cache since all partitions are removed. There shouldn't be any harm in even bypassing the cache, but it's there for a few reasons.
2021-03-21Debugging a partioning case.Anton Hvornum
2021-03-21Centralized a partition call to use the logging feature raw_parted().Anton Hvornum
2021-03-21Added more debug level information to parted in case it goes heywire.Anton Hvornum
2021-03-21Moved some info logs into debug instead to minimize user splash.Anton Hvornum
2021-03-21Added some debugging to help identify why archinstall cannot find ↵Anton Hvornum
unencrypted devices when setting up the boot loader.
2021-03-21Cleaned up after some debugging.Anton Hvornum
2021-03-21Spelling error on variable.Anton Hvornum
2021-03-21Moved the globalization of 'installation' variable. Since Profile() might be ↵Anton Hvornum
created before the installation instance, we need to move the global variable declaration after installation has begun and there's an installation instance.
2021-03-21Storing original namespace of profiles as they were during initation. ↵Anton Hvornum
Namespaces now get reverted back to the original state just before .install() is called. This ensures any temporary namespace changes made during prep-checks etc doesn't stick around when we try to install.
2021-03-21Reverted reset_namespace idea. Didn't turn out as clean as I wanted it to.Anton Hvornum
2021-03-21Enabling load_instructions() to set a temporary namespace, and then ↵Anton Hvornum
reverting it after the instructions are loaded. This is to temporarly override the namespace during import - enabling avoidance of triggering __name__ checks, and at the same time reverting back the namespace automatically to enable .execute() on the script (reusability of classes)
2021-03-21Enabling load_instructions() to set a temporary namespace, and then ↵Anton Hvornum
reverting it after the instructions are loaded. This is to temporarly override the namespace during import - enabling avoidance of triggering __name__ checks, and at the same time reverting back the namespace automatically to enable .execute() on the script (reusability of classes)
2021-03-21Debugging tainted namespace on profile modules.Anton Hvornum
2021-03-21Debugging tainted namespace on profile modules.Anton Hvornum
2021-03-21Debugging tainted namespace on profile modules.Anton Hvornum
2021-03-21Debugging tainted namespace on profile modules.Anton Hvornum
2021-03-21Debugging tainted namespace on profile modules.Anton Hvornum
2021-03-21Debugging tainted namespace on profile modules.Anton Hvornum
2021-03-21Debugging tainted namespace on profile modules.Anton Hvornum