cocoa: an empty display list is a failed read, not a Mac with no screens - #113
Merged
Merged
Conversation
liveDisplays answered (nil, nil) when CGGetActiveDisplayList counted zero, so "I could not tell you" and "there are none" arrived as the same value. Screens passed the second one on, and every caller believed it. A machine running a window server has at least one display: somebody is looking at something. A count of zero with no CGError means this process could not see the list, not that the hardware went away. Measured, 2026-09-08, twice in one evening on a desk somebody was wearing. go-xrkit/desk looks its display up by name and quit with desk: "VITURE Beast" is not attached any more; there is -- stopping, and putting back everything this changed Note what follows "there is": nothing. That message names every attached display and it named NONE, so the desk put the arrangement back and took a person's screens away over an answer this package never had. An earlier unexplained closure -- no crash report, no panic, exit status 0 -- has the same shape. ⭐ AND THE PACKAGE ALREADY AGREED IN THE PLACES EITHER SIDE OF IT: primaryBounds calls an empty list ErrDisplayList, and the portable layer has ErrScreensUnsupported for a back-end that cannot enumerate. This was the one path that answered with a fact instead of an error. The count is FAKED in the test rather than the machine emptied, which is the only way to reach the branch: unplugging every display from a Mac is not a test, and a runner with one display would never take it. Proved by sabotage: with (nil, nil) put back, the test names it. The 100% gate measures internal/cocoa on Linux, where these files are excluded by build tag, so nothing here changes what it counts. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
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.
liveDisplaysanswered(nil, nil)whenCGGetActiveDisplayListcounted zero, so "I could not tell you" and "there are none" arrived as the same value.Screenspassed the second one on, and every caller believed it.A machine running a window server has at least one display: somebody is looking at something. A count of zero with no CGError means this process could not see the list, not that the hardware went away.
What it cost
Measured 2026-09-08, twice in one evening, on a desk somebody was wearing.
go-xrkit/desklooks its display up by name and quit with:Note what follows
there is: nothing. That message names every attached display, and it named none. The desk put the arrangement back and took a person's screens away over an answer this package never had. An earlier unexplained closure — no crash report, no panic, exit status 0 — has the same shape.The package already agreed, either side of it
primaryBoundscalls a display list with nothing in itErrDisplayList.ErrScreensUnsupportedfor a back-end that cannot enumerate.This was the one path that answered with a fact instead of an error.
The test fakes the count, not the machine
Unplugging every display from a Mac is not a test, and a runner with one display would never take the branch — so
cgGetActiveDisplayListis replaced with one that counts zero and reports no error, which is exactly the shape that was believed. Proved by sabotage: with(nil, nil)put back, the test names it.The 100% gate measures
internal/cocoaon Linux, where these files are excluded by build tag, so nothing here changes what it counts. The live_darwin && integrationtests are unaffected — they run with displays attached.Companion: go-xrkit/desk#164 makes the desk safe whatever this answers.
🤖 Generated with Claude Code