feat: derive published browser counts, and add Naver Whale - #7
Merged
Merged
Conversation
Answering "what does three lines actually mean" by doing it. Adding Naver Whale took three hand-written lines across two files: the reference table row, the family path, and the process name. It worked, and it also broke three tests, which is the useful part. The browser count published in the README and on the site was suddenly wrong, and nothing but those tests would have noticed. That made the claim false: adding a browser really cost three lines plus remembering to update two numbers and a list by hand. So the published numbers are derived now. sync-skills.py reads platforms.py and rewrites the count and the browser list in both the README and the site, and CI fails if they drift. Adding a browser is three lines and one command again, this time accurately. Whale is kept rather than reverted. It is a real browser with real users in Korea, and all three skills now cover it.
The coverage section still said "three lines and no new skill", which omits the sync command and is the same overstatement just corrected everywhere else. It now says three lines and one command, names the three places, and cites Naver Whale as the browser added exactly that way.
The coverage copy said "Adding Vivaldi took three lines", which only lands if you already know Vivaldi is a browser. Most readers do not, and swapping in Naver Whale made it worse. The prose now says every browser listed above was added that way, which needs no prior knowledge and is equally true. The niche names stay in the supported-browser list, which is exactly where an unfamiliar name is useful rather than confusing.
vamgan
force-pushed
the
test/add-a-browser
branch
from
August 25, 2026 03:43
a9bc7bc to
b15010a
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Someone asked what "three lines and no new skill" actually means, so I measured it by adding a real browser instead of asserting it.
The measurement
Adding Naver Whale took exactly three hand-written lines across two files:
references/app-data-locations.mdscripts/platforms.py→CHROMIUMscripts/platforms.py→PROCESS_NAMESlocate whaleandrunning whaleboth resolved, and all three skills covered it immediately. No new skill, no new code.But it broke three tests, which is the useful part
The browser count published in the README and on the site was instantly wrong. Nothing except those drift tests would have caught it.
So the claim was still false. Adding a browser really cost three lines plus remembering to update two numbers and a list by hand — and "remember to update it" is how the earlier "one line in a table" claim went stale in the first place.
The fix: derive the numbers
sync-skills.pynow readsplatforms.pyand rewrites, in both the README and the site:Eleven browsers, one markdown file)CI fails if they drift. Adding a browser is three lines and one
synccommand again, this time accurately.Whale stays
It is a real browser with real users, not a test fixture. All three skills cover it.
109 tests green.