Repository navigation
fix(menu): Settings names the graphics chip again; labelled info panels on Network, Screen and This machine - #37
Merged
Merged
Conversation
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
sysfs reports every attribute file as 4096 bytes, whatever it holds. netdetail::attr read through bounded::read_text with a 256-byte cap, whose first check refuses a file by its reported size. So no GPU vendor/device id was ever read, every DRM card was skipped and the Screen tab said "No graphics device found" on every machine; on the Network tab the MAC address and the WiFi chip were dropped the same way (only the driver, a symlink, showed). Confirmed on the owner's thin client (4.6.0: card1, 8086:1616 i915, stat -L vendor = 4096). bounded::read_attr opens a regular file (FIFO guard kept), ignores the reported size and caps the bytes read. netdetail and hardware.rs (which worked only because its cap was 8 KiB) use it. The thin client's sysfs facts are in testdata/thinclient-drm/; the tests on the host's real /sys reproduce the bug. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
infopanel.rs holds the line model (rows, notes, groups), draws a band with a label column fitted to its labels, and gives the clipboard text. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
"Connections now" draws nettab::info_lines through infopanel: per interface a group with State, Network, Signal, Address, Handshake, MAC address, Gateway, Chip and Driver rows, then "DNS in use" (read by netstat::dns_in_use as before). iface_line, detail_line and dns_in_use_line are gone. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
graphics::info_lines gives GPU and Driver rows ("GPU: none found"
without one) and the codecs as a "Video processing" group
("H.264: decode, encode"), or "Video processing: not available".
The band moves above "Turn the screen off after", as "Connections
now" leads the Network tab. The monitor click tests use a 1920x1300
window: at 1080 the tab now scrolls and the Position row sat at its edge.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
…columns and copies them The tab draws four info panels (This machine, Add-ons, Network, Graphics) with the lines the Network and Screen tabs draw; two columns from 720 pt, the machine and add-ons left with Copy under them, the network and graphics right. Copy takes all four as infopanel::text. While the tab is open the network status is polled every 2 s (NetView::Facts: status only, never the files), and the once-per-run graphics job runs on the Screen or This machine tab, whichever opens first. Info rows are as high as their text (interact height 0), which makes the panels tight; the Signal row reads "strong (-55 dBm)". Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The controller's ruling: the Screen tab must fit a 1920x1080 screen with a full Graphics band, else both panels go to the bottom. Measured at 1920x1080: the page is visible to y=676; the monitors' Position select ended at about 700 under a one-line band and out of view under a full one. A band may take about 48 px there, a single line with its title; a GPU name, its driver and three codecs need about 930 px of an 811 px line, and columns need a title line plus a row (about 66 px). So "Graphics" is the last band of the Screen tab again and "Connections now" moves to the end of the Network tab. The monitor click tests are back at 1920x1080, and the_position_row_is_in_view_on_a_1080_screen checks the Position select inside the page with a full Graphics band. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
netdetail reads the chip and the driver of every interface with a device link, not only WiFi; WireGuard has none and shows neither. Test with the thin client's r8169 adapter. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
sysfs reports one page per attribute, which is 64 KiB on some hosts; the test asserted exactly 4096 and failed there. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
It reused the AX201's PCI ID and was named a Wi-Fi chip; it is now 8086:15fc, Ethernet Connection (13) I219-V. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The This machine tab's column split and the panel list spelled the title twice; a change to one would have moved the panel to the right column. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
… tabs The gap between the two columns was set as the item spacing of the whole row and reached into the columns: the label column was wider and the values stood 16 px further right than on the Network tab. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The bands' spacing was set on the column itself and stayed after the last band, so Copy stood at another distance below it than intended. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Until the machine's facts answered, the tab showed one band saying "Reading…" and hid the network and graphics it already had. Now only the machine's band waits; Copy stays off until its facts are read, as the copy without them was refused before. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
It passed only because the theme's gap differs from the bands' gap. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
oetiker
added a commit
that referenced
this pull request
Nov 20, 2026
The panels tests merged in #37 build netdetail::Details by hand; the time field this branch adds was missing there after the rebase. Co-Authored-By: Claude Opus 5.5 <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
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.
Settings info panels and the GPU detection fix (owner's 4.6.0 feedback, items E, F, G).
F: "No graphics device found" on every machine
Cause: sysfs reports one page (4096 bytes on most hosts) as the size of every attribute, and the 256-byte size check before reading made every GPU, MAC and WiFi chip read fail in 4.6.0. New
bounded::read_attrreads up to the cap without trusting the size (FIFO and non-regular-file guards kept);netdetail,graphicsand thestatus/hardware.rsDMI/hwmon reads use it. Thin-client evidence in the spec.E/G: labelled facts
infopanel::Line) for labelled facts on the Network, Screen and This machine tabs ("Driver: iwlwifi").the_position_row_is_in_view_on_a_1080_screen).Review
Whole-branch review (ready to merge), two fix waves for every minor finding and a scoped re-review.
Gates (rebased on main after #36)
cargo clippy --all-targets -D warnings,cargo fmt --check: clean.make test-unit: Python 971 tests OK; menu cargo test 1201 passed, 0 failed.make test-runtime: 68 ok, 0 failed.🤖 Generated with Claude Code