Owner override postdating PR #300 QA pass:
- .col-url 340 -> 230px; reclaim slack from history-only columns (Views 96->90,
Created 190->150, ID 190->120) so the table fits the .float pane with no
horizontal scroll at 1400x900 and the ID column is fully visible.
- URL anchors drop the shared .slug pill (background/padding/radius/mono font)
and render plain, styled via td.dim exactly like the none case; links stay
anchors with accent hover. Ellipsis kept for genuinely long slugs.
Verified by CDP probe on a local build at 1400x900 and 375x812 on /public and
/mine (/saved): column order, table fits pane (scrollWidth == clientWidth),
20-char slug fully visible, min Created-to-URL content gap 77px, sort arrows
intact 6px right of labels, no CSP/console violations. go build/test pass.
Reverts PR #297's column reorder (owner override): URL returns to position
6 in both list tables (Paste, Type, Size, Views, Created, URL, ID; /mine
without Views), data-sort attrs and td cells move back together. Keeps
#297's non-position improvements: /mine URL uses the shared .col-url class
instead of reusing col-e, and mine's ID column returns to .col-f (100px).
Widens .col-url 150px -> 340px so a ~20-char slug displays fully at
1400x900; genuinely long custom slugs keep text ellipsis. No mobile media
query needed: <=640px already scrolls the table horizontally (min-width
720px), verified at 375x812.
Verified via CDP probe at 1400x900 and 375x812 on /history and /mine:
column order, colgroup/header alignment, sort arrow right of label, no
other column clipped, no CSP changes (no inline styles added).
With wrap on, a logical line occupies several visual rows in the textarea
but the gutter showed one number per logical line, so every number after
the first wrapped line drifted off its text (the paste view fixed this in
#167; the editor gutter did not). Measure the wrapped row count per logical
line with a hidden mirror div sharing the editor's font and wrapping rules,
and render one .gutline block per visual row with the number on the first
row of its logical line. Re-measure on input, wrap toggle and resize.
Verified: gutter scrollHeight == textarea scrollHeight with zero diff at
1400x900 and 375x812, wrap on and off.
Fixed-position Top/Bottom pills appear only when content exceeds 2x
viewport height (window scroll on paste view, textarea scroll on /new).
New static/jump.js drives them; markup added to paste.html and new.html.
- image pastes render the image scaled to fit the viewer box (aspect
ratio preserved, max-height 70vh), no text/code box below it
- link pill moved under the image as a small inline chip
- /raw serves image attachment bytes as an image instead of empty text
- view details size reports the actual attachment file size
The details pill drew its own 1px border with a 10px radius inside .float, which draws the real border at 20px radius and clips with overflow:hidden. The parent's looser curve clipped the pill's tighter corner curve, making the border appear thinner/cut at the corners, in both collapsed and expanded states. Drop the pill's own border/radius so .float is the single border surface, matching the other pills on the page.
Ref #222
/mine rows render 7 cells (ID + delete button) but the colgroup and
thead only declared 6 columns, so under table-layout:fixed the delete
button cell overlapped the ID column at narrower widths. Add a
col-del column + empty header th to match the row cells.
topbar.js binds click handlers to every .dark-toggle, including the
settings one; settings.js then also wired a delegation that re-clicked
the topbar button, so each click flipped dark mode twice (net no-op).
Mark buttons as wired in topbar.js and only add a fallback handler in
settings.js when topbar.js did not run.
CSS: center the toggle horizontally under the Settings title with auto
margins, matching .settings-body padding.
Closes#197
- splitLines(): join .codeline blocks with "" not "\n" — newline text
nodes under pre-wrap rendered an extra line box per logical line, so
every gutter number drifted one row per line
- renumber(): geometry-driven placement — one gutline per visual row of
the code body, each number assigned to the row matching its line's
measured offsetTop; verified second pass re-runs if a partially-filled
last wrapped row reflows after the gutter rebuild
- app.css: mobile media query gives .code .gutter the same 13px font as
.code so gutter rows and code rows share one line box at 375px
Verified in local build with headless chromium (long unbroken token on
line 5, wrap ON): all numbers align with line starts, no misalignment,
no horizontal scroll at 1400x900 and 375x812.
Split the paste codebody into per-line block spans (paste-lines.js) and
rebuild the gutter from each line's wrapped height so numbers count
visual rows, not logical lines. Pinned the gutter width (flex 0 0 auto,
3ch) so its row count no longer feeds back into the code column width.
ResizeObserver renumbers on wrap toggle / resize; wrap-off behavior is
unchanged (one number per newline).
QA FAIL of PR #179 (verified on palette-dev 2026-09-10) found two defects:
1) Row accounting: splitLines() joined .codeline blocks with '\n' text
nodes. Since dee062d made .codeline display:block, each interleaved
newline text node rendered as its own extra visual row in the code
column under pre-wrap, but renumber() counted only .codeline span
heights, so the gutter had one fewer row per line break and numbers
drifted off alignment on wrapped lines. Fix: join('') so blocks are
adjacent with no phantom newline rows; blank source lines become
empty blocks given a line box via .codeline:empty::before (ZWJ) in
app.css so they still count as one row.
2) Inline gutter spans: .gutline spans were appended with no separators
and .code .gutter .gutline { display: block } did not exist on the
deployed build, so all numbers landed on one row. That rule is
present in origin/dev (merged via #183); kept unchanged and verified
by rendering in a local build.
Wrap OFF path unchanged: gutter still gets '1\n2\n3\n4' text and the
horizontal-scroll gutter behavior is untouched (no CSS changes to
scrolling; only a :empty line-box rule). CSP unchanged: no inline
styles/handlers, static JS only.
The can page 'Show content' dropdown had no divider between the summary
title and the body; the only line-like artifact appeared hover-dependent
(the stats header hover background edge). Per the owner follow-up, the
divider between a dropdown title and its body must be persistent (not
hover dependent) and extend the full width of the card left and right.
can page: .can-item-body[open] summary now carries a permanent 1px
var(--border) bottom rule spanning the full card width via negative
margins matching the card padding, plus matching bottom padding.
The divider between the stats dropdown head and body lived on
.stats-body's border-top, so it vanished when collapsed and only the
hover background hinted at the boundary (flicker). Move it to
.stats-head border-bottom so it is always present and spans the full
pill width.
QA of PR #179 found paste-lines.js creates .gutline spans but no CSS rule
existed, so all gutter numbers rendered inline on one visual row. Stack
them one per row under .code .gutter; blank fill spans inherit it.
- Pill now reads only 'Paste Created' (no combined 'Paste Created - Link Copied')
- Copy fallback toast now says 'Link Copied'
- Pill shrunk (smaller padding/font) and anchored tighter to the right side
Fixes#168
#139 removed inline display:block from paste.html's created banner and
compensated with a global .created-banner { display:block } override.
That override also hit the hidden placeholder banner on /new, whose
base rule is display:none, making an empty padded bar render above the
action bar. Scope the override to .created-banner.show and mark only
the paste-view banner with the show class.
- images render fitted into the editor area (object-fit contain), editor hidden
- text files load content into the editor; title placeholder convention (Python.py, Text.txt fallback)
- title auto-fills only when blank; never overwrites a typed title
- external JS + CSS classes only, CSP-safe (no inline styles/scripts)
Missing or expired paste IDs (and unknown routes) now render the main
UI chrome with a 'Paste ID not found' message in a color-coded result
card instead of a bare text/JSON 404. HTTP status stays 404. No inline
scripts or styles; new CSS uses existing --err token and pill radii.
wrap button on the new paste page looked identical whether wrap was on
or off because the btn base is already accent filled. give the toggle a
neutral off state so the accent on state is visible, matching the icon
button toggles on the paste view.
also remove the accent border on the whole language settings pill when
any child has focus. individual controls already show their own focus
rings, so the pill highlight made no sense.
fen