west: add --color option to control output colorization by TharakaUJ · Pull Request #930 · zephyrproject-rtos/west

TharakaUJ · 2026-03-04T20:12:49Z

Add a global --color option with values 'always', 'never', and 'auto' to override automatic terminal detection.

The command-line option takes precedence over configuration settings. When provided, it overrides the value returned by WestCommand.color_ui.

fixes #651

TharakaUJ · 2026-03-04T20:17:42Z

I initially considered adding support for a west config color always style configuration. However, since the configuration system expects section.key format (e.g., color.ui), I decided not to extend it in that direction.

Also for now 'auto' value just initialize colorama as default (existing behaviour).

Let me know if any changes to be made.
Thank you.

bencefr · 2026-03-05T10:36:25Z

This change would be very welcome with regards to environments like VS Code where executing child processes like west build as tasks we cannot cleanly provide a tty attached to get colored output. While environment variables like FORCE_COLOR=1 CLICOLOR_FORCE=1 TERM=xterm-256color does help with some underlying tools, they don't help with west's colorama.

marc-hb · 2026-03-06T16:54:49Z

Thanks @TharakaUJ ! Can you please run uv run poe format and force push? Please also look at the other CI check. Do NOT spend too much time with cosmetics before review (in case the code has to change), but uv run poe format takes approximately zero time to run! Maybe the other CI check is very quick to fix too; I haven't looked at it yet.

TharakaUJ · 2026-03-06T18:18:34Z

Thanks! I ran uv run poe format and fixed the lint issues. Force pushed the updated commit.

codecov · 2026-03-09T04:35:21Z

Codecov Report

❌ Patch coverage is 84.09091% with 7 lines in your changes missing coverage. Please review.
✅ Project coverage is 85.84%. Comparing base (5f5931f) to head (55dc524).
⚠️ Report is 6 commits behind head on main.
✅ All tests successful. No failed tests found.

Files with missing lines	Patch %	Lines
src/west/app/main.py	82.50%	7 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #930      +/-   ##
==========================================
- Coverage   85.95%   85.84%   -0.12%     
==========================================
  Files          11       11              
  Lines        3454     3511      +57     
==========================================
+ Hits         2969     3014      +45     
- Misses        485      497      +12

Files with missing lines	Coverage Δ
src/west/commands.py	`94.66% <100.00%> (-0.88%)`	⬇️
src/west/app/main.py	`79.10% <82.50%> (+0.02%)`	⬆️

... and 2 files with indirect coverage changes

pdgendt · 2026-03-13T09:00:52Z

@TharakaUJ thanks for the PR, I still need to do a full review, but I wonder if we could test this somehow. Codecov coverage shows that most of the changes isn't even executed during testing.

TharakaUJ · 2026-03-13T20:11:06Z

@TharakaUJ thanks for the PR, I still need to do a full review, but I wonder if we could test this somehow. Codecov coverage shows that most of the changes isn't even executed during testing.

ok. I will look into adding test cases. Thank you!

pdgendt · 2026-03-14T21:03:25Z

Make sure to run the formatter on the newly added tests, thanks!

src/west/app/main.py

TharakaUJ · 2026-03-15T15:26:38Z

Thanks, applied the suggestion.

marc-hb · 2026-03-16T22:39:47Z

@TharakaUJ please squash and force-push, the zephyrproject cares about git history and does not accept "fixup" commits in it (like 147920c)

TharakaUJ · 2026-03-17T05:04:12Z

Thank you. Squashed the fixup commit.

marc-hb

Sorry I really don't understand why we need this new cmd._color_override field and new override function. It should be possible to keep all the logic centralized in a single accessor like color_ui() rather than scattered in different places like this, why not?

BTW you can probably remove some of the if self.color_ui checks in the code to simplify it now that this PR uses colorama.init(strip=True) (some other places will likely still need that check).

Last but not least: make sure you test thoroughly on both Linux and Windows. I started to do some manual testing that I will summarize soon in #651 (EDIT: done) and I found some existing issues even before this PR. Fixing them is out of scope for this PR but we have to be sure this #930 is not making anything worse. Also, #651 provides a good list of manual tests.

marc-hb · 2026-03-18T16:47:03Z

src/west/app/main.py

+            config_value = self.config.get('color.ui', 'auto')
+            # Handle both string values (always/never/auto) and
+            # boolean values (true/false) for backward compatibility
+            if config_value in ('always', 'true', 'True', '1'):


I don't think we can do that because color.ui=True (or 1, etc.) had no effect since forever. Now with this PR it becomes always. That could seriously mess up some CI logs and others.

I think it's OK for True and always to have different meanings.

marc-hb · 2026-03-18T16:50:29Z

src/west/app/main.py

+        elif color_mode == 'never':
+            colorama.init(strip=True)
+        else:
+            # 'auto' or None: use colorama's default behavior


Suggested change

# 'auto' or None: use colorama's default behavior

# 'auto' or None or "true": defer to colorama's terminal detection

marc-hb · 2026-03-18T17:12:27Z

src/west/app/main.py

+        color_mode = self.color_mode
+
+        if color_mode is None and self.config:
+            config_value = self.config.get('color.ui', 'auto')


We don't want to access self.config.get('color.ui',... in two different places in the code, this seems too complicated and too high-maintenance. You can read the config value at a different time and place if there's a good reason for that, but don't add a new place while keeping the old place.

grep.color and other quirks add enough complexity already :-(

Please keep using getboolean() and catch the ValueError when it's always or never rather than duplicating the special values from configparser.

TharakaUJ · 2026-03-18T20:13:29Z

Thank you the detailed feedback. I will look into necessary changes.

Add a global --color option with values 'always', 'never', and 'auto' to override automatic terminal detection. The command-line option takes precedence over configuration settings. When provided, it overrides the value returned by WestCommand.color_ui. Signed-off-by: Tharaka Jayasena <9dmpires2k17.tuj@gmail.com>

Add tests for: - CLI override precedence in WestCommand.color_ui - behavior with and without configuration - parsing of the new --color argument in parse_early_args Signed-off-by: Tharaka Jayasena <9dmpires2k17.tuj@gmail.com>

TharakaUJ · 2026-03-20T10:53:53Z

I made some changes and force pushed. I thought a early review might helpful to stay on needed track.
also, unfortunately I won't be able to test this on windows until Monday.

TharakaUJ · 2026-03-20T11:00:31Z

Sorry I really don't understand why we need this new cmd._color_override field and new override function. It should be possible to keep all the logic centralized in a single accessor like color_ui() rather than scattered in different places like this, why not?

Got rid of most of the _color_override logic. But still could not centralize all the logic, mainly because I'm not sure whether calling calorama.init() in westCommand would initialize it more than once. I will look into it.

marc-hb · 2026-03-20T17:05:31Z

Got rid of most of the _color_override logic.

Thanks and thanks for the heads-up.

But still could not centralize all the logic, mainly because I'm not sure whether calling calorama.init() in westCommand would initialize it more than once. I will look into it.

Why would we need colorama.init() is WestCommands? Thanks to your new colorama.init(strip=True), we should be able to have LESS color-related code in WestCommands. If not in this PR, at least in the future. But for sure no additional color-related code in WestCommands, or why? Colorization is global.

Ideally, the only color-related code in WestCommands should be decisions like "how important is this log message?" and never "What does the user/configuration/terminal want?".

OK, now I just realized we haven't looked at colorization from a west API perspective... no idea what is the colorization status there :-(

TharakaUJ force-pushed the main branch from a697c48 to 8942a47 Compare March 6, 2026 18:14

TharakaUJ force-pushed the main branch from 2395571 to c7f5c29 Compare March 15, 2026 04:19

pdgendt approved these changes Mar 15, 2026

View reviewed changes

src/west/app/main.py Outdated Show resolved Hide resolved

TharakaUJ force-pushed the main branch from 255fb6f to 147920c Compare March 15, 2026 16:04

TharakaUJ force-pushed the main branch from 147920c to c2293b7 Compare March 17, 2026 05:00

marc-hb requested changes Mar 18, 2026

View reviewed changes

marc-hb mentioned this pull request Mar 18, 2026

Add west --color=always and west config color=always #651

Open

TharakaUJ added 2 commits March 20, 2026 16:09

tests: add coverage for --color option

55dc524

Add tests for: - CLI override precedence in WestCommand.color_ui - behavior with and without configuration - parsing of the new --color argument in parse_early_args Signed-off-by: Tharaka Jayasena <9dmpires2k17.tuj@gmail.com>

TharakaUJ force-pushed the main branch from c2293b7 to 55dc524 Compare March 20, 2026 10:48

	# 'auto' or None: use colorama's default behavior
	# 'auto' or None or "true": defer to colorama's terminal detection

Conversation

TharakaUJ commented Mar 4, 2026

Uh oh!

TharakaUJ commented Mar 4, 2026

Uh oh!

bencefr commented Mar 5, 2026

Uh oh!

marc-hb commented Mar 6, 2026

Uh oh!

TharakaUJ commented Mar 6, 2026

Uh oh!

codecov bot commented Mar 9, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

pdgendt commented Mar 13, 2026

Uh oh!

TharakaUJ commented Mar 13, 2026

Uh oh!

pdgendt commented Mar 14, 2026

Uh oh!

Uh oh!

TharakaUJ commented Mar 15, 2026

Uh oh!

marc-hb commented Mar 16, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

TharakaUJ commented Mar 17, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

marc-hb left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

marc-hb Mar 18, 2026

Choose a reason for hiding this comment

Uh oh!

marc-hb Mar 18, 2026

Choose a reason for hiding this comment

Uh oh!

marc-hb Mar 18, 2026

Choose a reason for hiding this comment

Uh oh!

marc-hb Mar 18, 2026

Choose a reason for hiding this comment

Uh oh!

TharakaUJ commented Mar 18, 2026

Uh oh!

TharakaUJ commented Mar 20, 2026

Uh oh!

TharakaUJ commented Mar 20, 2026

Uh oh!

marc-hb commented Mar 20, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

codecov bot commented Mar 9, 2026 •

edited

Loading

marc-hb commented Mar 16, 2026 •

edited

Loading

TharakaUJ commented Mar 17, 2026 •

edited

Loading

marc-hb left a comment •

edited

Loading

marc-hb commented Mar 20, 2026 •

edited

Loading