clippy #88
clippy #88
Conversation
|
MFW there's an xclippy command. So, clippy has something that turns all warnings into errors. See https://github.com/MegatonHammer/linkle/blob/master/ci/script.sh#L16 (-D warnings means "deny warnings"). When we add it to the ci, we should use that. |
|
Closes #68 |
map_grub_module panicmap_grub_module panics if virtual space is exhausted, but should return an error instead, as it is used in the panic handler itself. This comment was generated by todo based on a
|
VirtalSpaceLand start_table/ end_table is arch specificThese functions should be moved to `paging::arch::i386::table.rs` This comment was generated by todo based on a
|
Now start_addr and end_addr are defined from an associated constant of the VirtualSpaceLand trait, instead of calling a function.
Land::contains_region() should not panic on 0 lengthThis function should return an error, as it really is likely someone (I) will call it at some point not expecting it can panic. This comment was generated by todo based on a
|
move KernelLand Userland RTL to arch-specific pagingThey are arch dependant, we should stop trying defining them in an agnostic way, even if they are expected to be the same for 32 bits architectures. This comment was generated by todo based on a
|
We used to break alignment requirements in elf_loader::get_kacs. To avoid the issue, get_kacs now simply returns a byte array, and ProcessCapabilities::parse uses byteorder to extract the u32 kacs from it.
- Rewrite EntryOptions using bit_flags. Add some type safety around the Gate Type and Privilege Level. - Manually implement Debug on IdtEntry, treating it like an enum based on the present flag and gate type, and joining the pointer fields. - Manually implement Debug on Idt to deal with lack of Debug on large arrays.
More type safety, and better debug implementation. While we're at it, make SegmentSelector aware of the TI flag (called is_ldt here).
While a DT entry's limit is the byte length - 1, the DTP's limit is the full byte length. When converting back to a Vec<DTE>, I would probably end up losing the entry.
Missing modules are: ipc, devices and frame_allocator.
This allows clippy to proceed on the full project.
Feed the timer handler into a kernel preemption handler.This comment was generated by todo based on a
|
Missing argument slot for SVCs on i386 backendOur i386 SVC ABI is currently fairly different from the ABI used by Horizon/NX. This is for two reasons:
This comment was generated by todo based on a
|
Added documenting for * pic * pit * rs232 * frame_allocator * main * ...
|
Missing:
We're almost there \o/ |
Only thing missing: vi has some cast warnings I'm not sure how to fix.
Meme for KFS5.We cannot give KFS5 until we have a meme. It is of utmost importance that a meme is found and placed here. This comment was generated by todo based on a
|
|
We're almost there. Only a few warnings in vi are left. I'm going to fix them tomorrow when my head is fresh because they might be tricky and be pointing to actual bugs... |
Write documentation for AHCIDocumentation lints are disabled on AHCI for now, to avoid conflicts. Get @Orycterope to write some doc \o/ This comment was generated by todo based on a
|
Review Terminal::new cast_sign_loss and cast_possible_wrapPing @Orycterope. I'm pretty sure they're correct but just want to make sure it looks good with you since it's originally your code. This comment was generated by todo based on a
|
Review Terminal::display_glyph_in_box cast_sign_loss and cast_possible_wrapPing @Orycterope. I'm pretty sure they're correct but just want to make sure it looks good with you since it's originally your code. This comment was generated by todo based on a
|
Fixes #68
clippy lint