diff options
author | Erovia <Erovia@users.noreply.github.com> | 2020-09-06 13:06:12 +0100 |
---|---|---|
committer | GitHub <noreply@github.com> | 2020-09-06 13:06:12 +0100 |
commit | ac24f203cc4141d6d27f27dc173e04fc5edde741 (patch) | |
tree | 10ffa9431363417a6a8aeb1aa64ff555bf9fd3b0 | |
parent | d0eabd083ebe13b864dd7628b43096c8362d0f29 (diff) |
Docs/PR_checklist: Reorder, unify and pet-peeves (#10253)
Co-authored-by: James Young <18669334+noroadsleft@users.noreply.github.com>
-rw-r--r-- | docs/pr_checklist.md | 69 |
1 files changed, 36 insertions, 33 deletions
diff --git a/docs/pr_checklist.md b/docs/pr_checklist.md index 8755552b9d..22e8a3fe1a 100644 --- a/docs/pr_checklist.md +++ b/docs/pr_checklist.md @@ -1,39 +1,42 @@ # PR checklists -This is a non-exhaustive checklist of what the QMK collaborators will be checking when reviewing submitted PRs. +This is a non-exhaustive checklist of what the QMK Collaborators will be checking when reviewing submitted PRs. -If there are any inconsistencies with these recommendations, you're best off [creating an issue](https://github.com/qmk/qmk_firmware/issues/new) against this document, or getting in touch with a QMK Collaborator on Discord. +If there are any inconsistencies with these recommendations, you're best off [creating an issue](https://github.com/qmk/qmk_firmware/issues/new) against this document, or getting in touch with a QMK Collaborator on [Discord](https://discord.gg/Uq7gcHh). ## General PRs - PR should be submitted using a non-`master` branch on the source repository - - This does not mean you target a different branch for your PR, rather that you're not working out of your own master branch - - If submitter _does_ use their own `master` branch, they'll be given a link to the ["how to git"](https://docs.qmk.fm/#/newbs_git_using_your_master_branch) page after merging -- (end of this document will contain the contents of the message) -- Newly-added directories and filenames must be lowercase - - This rule may be relaxed if upstream sources originally had uppercase characters (e.g. ChibiOS, or imported files from other repositories etc.) - - If there is enough justification (i.e. consistency with existing core files etc.) this can be relaxed + - this does not mean you target a different branch for your PR, rather that you're not working out of your own master branch + - if submitter _does_ use their own `master` branch, they'll be given a link to the ["how to git"](https://docs.qmk.fm/#/newbs_git_using_your_master_branch) page after merging -- (end of this document will contain the contents of the message) +- newly-added directories and filenames must be lowercase + - this rule may be relaxed if upstream sources originally had uppercase characters (e.g. ChibiOS, or imported files from other repositories etc.) + - if there is enough justification (i.e. consistency with existing core files etc.) this can be relaxed - a board designer naming their keyboard with uppercase letters is not enough justification -- Valid license headers on all `*.c` and `*.h` source files +- valid license headers on all `*.c` and `*.h` source files - GPL2/GPL3 recommended for consistency - - Other licenses are permitted, however they must be GPL-compatible and must allow for redistribution. Using a different license will almost certainly delay a PR getting merged. -- QMK codebase "best practices" followed - - This is not an exhaustive list, and will likely get amended as time goes by + - other licenses are permitted, however they must be GPL-compatible and must allow for redistribution. Using a different license will almost certainly delay a PR getting merged. +- QMK Codebase "best practices" followed + - this is not an exhaustive list, and will likely get amended as time goes by - `#pragma once` instead of `#ifndef` include guards in header files - - No "old-school" GPIO/I2C/SPI functions used -- must use QMK abstractions unless justifiable (and laziness is not valid justification) - - Timing abstractions should be followed too: + - no "old-school" GPIO/I2C/SPI functions used -- must use QMK abstractions unless justifiable (and laziness is not valid justification) + - timing abstractions should be followed too: - `wait_ms()` instead of `_delay_ms()` (remove `#include <util/delay.h>` too) - `timer_read()` and `timer_read32()` etc. -- see [timer.h](https://github.com/qmk/qmk_firmware/blob/master/tmk_core/common/timer.h) for the timing APIs - - If you think a new abstraction is useful, you're encouraged to: + - if you think a new abstraction is useful, you're encouraged to: - prototype it in your own keyboard until it's feature-complete - discuss it with QMK Collaborators on Discord - refactor it as a separate core change - remove your specific copy in your board +- rebase and fix all merge conflicts before opening the PR (in case you need help or advice, reach out to QMK Collaborators on Discord) -## Core PRs +## Keymap PRs -- Must now target `develop` branch, which will subsequently be merged back to `master` on the breaking changes timeline -- Other notes TBD - - Core is a lot more subjective given the breadth of posted changes +- `#include QMK_KEYBOARD_H` preferred to including specific board files +- prefer layer `enum`s to `#define`s +- require custom keycode `enum`s to `#define`s, first entry must have ` = SAFE_RANGE` +- terminating backslash (`\`) in lines of LAYOUT macro parameters is superfluous +- some care with spacing (e.g., alignment on commas or first char of keycodes) makes for a much nicer-looking keymap ## Keyboard PRs @@ -48,12 +51,14 @@ https://github.com/qmk/qmk_firmware/pulls?q=is%3Apr+is%3Aclosed+label%3Akeyboard - standard template should be present - flash command has `:flash` at end - valid hardware availability link (unless handwired) -- private groupbuys are okay, but one-off prototypes will be questioned. If open-source, a link to files should be provided. + - clear instructions on how to reset the board into bootloader mode + - a picture about the keyboard and preferably about the PCB, too - `rules.mk` - removed `MIDI_ENABLE`, `FAUXCLICKY_ENABLE` and `HD44780_ENABLE` - modified `# Enable Bluetooth with the Adafruit EZ-Key HID` -> `# Enable Bluetooth` - - No `(-/+size)` comments related to enabling features - - Remove the list of alternate bootloaders if one has been specified - - No re-definitions of the default MCU parameters if same value, when compared to the equivalent MCU in [mcu_selection.mk](https://github.com/qmk/qmk_firmware/blob/master/quantum/mcu_selection.mk) + - no `(-/+size)` comments related to enabling features + - remove the list of alternate bootloaders if one has been specified + - no re-definitions of the default MCU parameters if same value, when compared to the equivalent MCU in [mcu_selection.mk](https://github.com/qmk/qmk_firmware/blob/master/quantum/mcu_selection.mk) - keyboard `config.h` - don't repeat `MANUFACTURER` in the `PRODUCT` value - no `#define DESCRIPTION` @@ -71,12 +76,12 @@ https://github.com/qmk/qmk_firmware/pulls?q=is%3Apr+is%3Aclosed+label%3Akeyboard - `keyboard.h` - `#include "quantum.h"` appears at the top - `LAYOUT` macros should use standard definitions if applicable - - Use the Community Layout macro names where they apply (preferred above `LAYOUT`/`LAYOUT_all`) + - use the Community Layout macro names where they apply (preferred above `LAYOUT`/`LAYOUT_all`) - keymap `config.h` - no duplication of `rules.mk` or `config.h` from keyboard - `keymaps/default/keymap.c` - `QMKBEST`/`QMKURL` removed (sheesh) - - If using `MO(_LOWER)` and `MO(_RAISE)` keycodes or equivalent, and the keymap has an adjust layer when holding both keys -- if the keymap has no "direct-to-adjust" keycode (such as `MO(_ADJUST)`) then you should prefer to write... + - if using `MO(_LOWER)` and `MO(_RAISE)` keycodes or equivalent, and the keymap has an adjust layer when holding both keys -- if the keymap has no "direct-to-adjust" keycode (such as `MO(_ADJUST)`) then you should prefer to write... ``` layer_state_t layer_state_set_user(layer_state_t state) { return update_tri_layer_state(state, _LOWER, _RAISE, _ADJUST); @@ -90,22 +95,20 @@ https://github.com/qmk/qmk_firmware/pulls?q=is%3Apr+is%3Aclosed+label%3Akeyboard - submitters can also have a "manufacturer-matching" keymap that mirrors existing functionality of the commercial product, if porting an existing board Also, specific to ChibiOS: -- **Strong** preference to using existing ChibiOS board definitions. - - A lot of the time, an equivalent Nucleo board can be used with a different flash size or slightly different model in the same family - - Example: For an STM32L082KZ, given the similarity to an STM32L073RZ, you can use `BOARD = ST_NUCLEO64_L073RZ` in rules.mk +- **strong** preference to using existing ChibiOS board definitions. + - a lot of the time, an equivalent Nucleo board can be used with a different flash size or slightly different model in the same family + - example: For an STM32L082KZ, given the similarity to an STM32L073RZ, you can use `BOARD = ST_NUCLEO64_L073RZ` in rules.mk - QMK is migrating to not having custom board definitions if at all possible, due to the ongoing maintenance burden when upgrading ChibiOS -- If a board definition is unavoidable, `board.c` must have a standard `__early_init()` (as per normal ChibiOS board defs) and an empty `boardInit()`: +- if a board definition is unavoidable, `board.c` must have a standard `__early_init()` (as per normal ChibiOS board defs) and an empty `boardInit()`: - see Arm/ChibiOS [early initialization](https://docs.qmk.fm/#/platformdev_chibios_earlyinit?id=board-init) - `__early_init()` should be replaced by either `early_hardware_init_pre()` or `early_hardware_init_post()` as appropriate - `boardInit()` should be migrated to `board_init()` -## Keymap PRs +## Core PRs -- `#include QMK_KEYBOARD_H` preferred to including specific board files -- Prefer layer `enum`s to `#define`s -- Require custom keycode `enum`s to `#define`s, first entry must have ` = SAFE_RANGE` -- Terminating backslash (`\`) in lines of LAYOUT macro parameters is superfluous -- Some care with spacing (e.g., alignment on commas or first char of keycodes) makes for a much nicer-looking keymap +- must now target `develop` branch, which will subsequently be merged back to `master` on the breaking changes timeline +- other notes TBD + - core is a lot more subjective given the breadth of posted changes --- |