| Age | Commit message (Collapse) | Author |
|
|
|
With markers at e.g. 350 deg and 10 deg the raw azimuth difference is
340 deg instead of 20 deg. Two places choked on that:
* ProjArgs::from_params fed the raw difference into the projection
formulas. The cylindrical projection uses it LINEARLY
(x = scale * (m_view - c_view)), so the 350 deg peak was projected
350 degrees of scale to the right of center instead of 10 degrees to
the left - no parameter set can satisfy both markers and the
panorama-mode fit fails. The rectilinear projection is periodic and
unaffected. Fix: represent m_view as c_view + wrapped difference, so
every consumer (fit residuals, Jacobian, project_point) sees the
shortest arc. get_coordinates already normalized for exactly this
reason; the fit path bypassed it.
* comp_params derived the initial scale from |d_alph| = 340 deg, an
initial scale off by a factor of 17 - this hit rectilinear mode too.
Use the wrapped difference.
Regression tests: two-marker fits across north for both projections
(the cylindrical one fails without the fix), plus a direct check that
a 350 deg peak projects to -10 deg * scale.
|
|
|
|
It drifted to 0.5.5 after the bump and silently produced mislabeled
bundles. Omitting it makes tauri fall back to the crate version, which
inherits the workspace version — one source of truth again.
|
|
|
|
Bump workspace version, add ChangeLog entry (drift guard test passes),
rebuild the release web bundle into dist/gipfel (wasm rebuilt against
the merged core including the is_visible fix).
|
|
Was still titled wipfel with W1-era status and trunk-only instructions;
now documents serve-here.sh/build-release.sh, the URL parameters and
the current module layout.
|
|
Directory, Kotlin package (org.flpsed.gipfel), applicationId/namespace,
app label and gradle project name — the wipfel working name is gone
from anything that ships.
|
|
Gipfel.lean verifies the structural invariants (mark_duplicates keeps
hill indices stable). GipfelFind.lean hunts flaws: exact Int model for
the general fact, Float grid search for concrete IEEE witnesses, every
witness re-checked against the real code by rust-check/.
Findings: 1 fixed on master (is_visible azimuth wrap), 2 proven benign
via project_point (all witnesses NaN via 0*inf; residual near-zero
denominators documented), 3 open but cosmetic, 4 already handled by
the stitch caller. GipfelMathlib.lean.txt stays an unchecked sketch,
kept as .txt so it cannot pass for a verified artifact.
Verified from this location: lean checks both files clean, rust-check
runs against the workspace gipfel-core via path dependency.
|
|
# Conflicts:
# crates/gipfel-core/src/metadata.rs
|
|
GipfelFind.lean Finding 1: with |alph - a_center| > 2pi (user enters
azimuth 400deg, or a fit leaves a_center unnormalized) the single fold
produced a NEGATIVE 'distance' that passed every comparison, making
hills behind the camera visible. NaN or infinite azimuths — and a NaN
center, which made everything visible — passed as well.
Fix at the point of use: wrap with % 2pi before the fold, and bail out
on non-finite alph/a_center. Regression test pins the Lean witnesses
(out_of_range_center_angle_stays_safe).
|
|
- ?file= now also accepts absolute http(s) image URLs (CORS applies);
README documents all URL parameters
- Projection dropdown (rectilinear / cylindrical-panoramic), synced
from the gipfel comment on load; cylindrical lifts the lens-FOV
visibility limit (unlimited view angle) so 360 deg panoramas show
peaks in every direction
- force visibility rebuild on every image load (the change detection
alone could keep a stale list when the limit value was unchanged)
- Google-Maps-style +/- zoom buttons at the lower right of the canvas
- save/load roundtrip test for the projection type in the gipfel
comment; build stamp in the sidebar footer and ?debug=1 console
diagnostics to spot stale cached bundles and inspect the effective
visibility state
|
|
- rename crates/wipfel -> crates/gipfel (crate, wasm artifact, dist,
scripts, page titles); slogan back to "Photogrammetry for Mountain
Images"; crates/wipfel-tauri -> crates/gipfel-desktop
- new features in the wasm app: Save image (gipfel JPEG comment, same
format as before, via browser download) and lens distortion profiles
stored in localStorage in FLTK prefs format with camera auto-match
on image load; profile name field, save/load and dropdown
- gipfel-core: read GPSImgDirection from EXIF (magnetic -> true) as
azimuth fallback; expose parse_prefs/serialize_prefs for the wasm
build; ?file=/?az=/?pitch=/?roll=/?focal=/?vis= URL parameters
- Tauri v2 wrapper crate building .deb, AppImage and (cross-compiled)
Windows NSIS installers; safe stage-and-swap build-release.sh
- README: web-centric Installation, concise Build section for the
official targets (Web, Windows, Linux); drop native/Android mentions
- release site moved dist/wipfel -> dist/gipfel
|
|
- gipfel-cli export-hills gains -V <ratio> (the visibility/prominence
threshold, C default 0.07); default unchanged
- export_hill_rows: restricted mode (-e, curated lists like the
gipfelweb photo index) no longer drops FLAG_HIDDEN entries. The
occlusion heuristic is for thinning the full peak database; a
curated photo list should show every entry inside the field of
view. Unrestricted export (golden-pinned) unchanged.
- gipfelweb Makefile: V ?= 0.05 knob for the peak view, -V 0 for the
photo index; per-image %.opts now appended after the standard flags
(they were placed before the subcommand and silently ignored)
|
|
The historical 20000x500 slide rule was strip-specific: a normal
4:3 photo was squeezed to 500 px height. Cap the long edge at
6000 px instead (full size for typical photos, sane for stitched
images); thumbs fit 480x480 so gallery cards and hover popups look
right for both aspect ratios.
|
|
Any gipfel-annotated photo works, plain or stitched; the user-facing
wording and comments should not imply panoramas specifically.
|
|
- markers positioned in percent of the slide (identify reads the
dimensions), so the panorama scales with the viewport; a CSS-only
'Full resolution' checkbox switches to natural size with horizontal
scrolling
- peaks and photos now share one page; chip-style checkboxes toggle
Peaks / Photos / Labels / Full resolution (pure CSS sibling
selectors, no JavaScript)
- interaction for all input types: hover (mouse), :focus-within
(keyboard, markers are anchors), :target (touch, tap to open a
popup); popups get below/left/right classes from the generator so
they never leave the image
- labels show distance (the data always had it): NAME · 2849 m ·
12.3 km
- new gallery index.html (thumbnail grid), dark/light via
prefers-color-scheme, system font stack, lazy thumbnails, aria
labels, viewport meta; stylesheet is a single external style.css
- backlink updated to flpsed.org/gipfel.html
- gipfel2html: LC_ALL=C (awk decimal separator), no pipe-subshell
marker loops (counter must survive), exact cross-reference matching
The old two-page (images/hills) split is gone; the hills view is the
Peaks toggle now.
|
|
Port of the pre-Rust static web gallery pipeline (recovered from
history): Makefile drives gipfel-cli export-position/export-hills
instead of the C binary, ImageMagick replaces netpbm + jhead
(convert preserves the gipfel comment and EXIF natively). Output
structure and the CSS-only hover interaction are unchanged for now;
modernization follows separately.
gipfel2html: cross-image references now match the name column exactly
(the old grep -l also matched substrings, e.g. test.jpg inside
contest.jpg).
Verified end to end with two annotated test panoramas: slides,
thumbs, positions, cross-links and hill markers all generate.
|
|
Self-contained hardening, no server config required:
default-src 'self'; script-src 'self' 'wasm-unsafe-eval';
img-src 'self' blob: data:; connect-src 'self' blob: data:;
style-src 'self' 'unsafe-inline'
- script-src needs 'wasm-unsafe-eval' for WebAssembly compilation
(the scoped exception; plain 'unsafe-eval' stays forbidden) and no
inline script is allowed, so the module loader moved from an inline
<script> to loader.js
- style-src keeps 'unsafe-inline' because the Yew views position the
find-peak popup and size the compass strip via style attributes
- blob: in img-src/connect-src for createImageBitmap and blob reads;
data: for inline assets
Size budget note: unchanged (~550 KB gzipped total, well under the
1.5 MB plan budget).
|
|
A UseStateHandle cloned into a long-lived closure derefs to its
render-time snapshot (same root cause as the pinch bug), so
tick.set(*tick + 1) wrote the same value on every resize after the
first: no state change, no redraw, stale stretched canvas backing
store — the distortion thought fixed by d465654 returned on the second
resize (the window-resize listener, created once at startup, never
worked at all). Yew 0.21 has no functional-update setter, so the tick
now has a use_mut_ref mirror as the live counter.
|
|
The */* workaround did not circumvent the MediaProvider EXIF
redaction, so no reason to show non-image files in the picker.
|
|
Renamed to 'Load alternative hill database…' and placed after 'Load
GPX Track…' — the default database loads automatically, so the button
must not sit where it could look like a required first step.
|
|
The load result was matched on while the scrutinee's borrow_mut() was
still alive; the Ok arm called pan.borrow() again, which panics (and
traps the whole wasm module, hanging the app). End the borrow before
the match.
|
|
The longest-numeric-prefix fallback sliced s[..i] at every byte index,
which panics when i lands inside a multi-byte character (reachable via
lossy-decoded Latin-1 database fields, e.g. a trailing degree sign).
Skip non-boundary indices; behavior on ASCII numeric fields is
unchanged. Includes regression test atof("12.5 deg").
|
|
The 35mm film frame is 36 mm wide, so a lens with 35mm-equivalent
focal length f projects with focal length f*w/36 in pixels. The
original C++ code divided by 35, making the EXIF-derived initial
scale ~2.9% too large (and the fitted focal readout ~2.9% low).
- EXIF init and the focal slider/readout now use /36
- the gipfel JPEG comment KEEPS its historical scale*35/w encoding
(focal_from_comment flag), so annotated JPEGs written by any older
gipfel version still reproduce the identical view, and files saved
by this version remain readable by them
- new regression test focal_scale_exif_vs_comment; golden tests
untouched (they pin the comment path)
Verified against real photos: optical shots now agree with EXIF
within ~1% (was ~3.6%).
|
|
do_refit only updated the fit-info text; the Azimuth/Pitch/Roll/Focal
sliders kept showing their pre-fit values because the slider state
(vp) was never synced from the panorama after a fit. Long-standing
wipfel bug, independent of the recent focal-constant change. The
compass strip was unaffected (it reads the panorama at render time).
|
|
Secondary button under the database status line: picks a .dat (or
gzipped) peak list at runtime — e.g. a regional file like everest.dat
instead of the full 21k gipfel.dat. Markers and track reference hill
indices, so they are dropped when the database is swapped.
|
|
The longest-numeric-prefix fallback sliced s[..i] at every byte index,
which panics when i lands inside a multi-byte character (reachable via
lossy-decoded Latin-1 database fields, e.g. a trailing degree sign).
Skip non-boundary indices; behavior on ASCII numeric fields is
unchanged. Includes regression test atof("12.5 deg").
|
|
The 35mm film frame is 36 mm wide, so a lens with 35mm-equivalent
focal length f projects with focal length f*w/36 in pixels. The
original C++ code divided by 35, making the EXIF-derived initial
scale ~2.9% too large (and the fitted focal readout ~2.9% low).
- EXIF init and the focal slider/readout in both apps now use /36
- the gipfel JPEG comment KEEPS its historical scale*35/w encoding
(focal_from_comment flag), so annotated JPEGs written by any older
gipfel version still reproduce the identical view, and files saved
by this version remain readable by them
- new regression test focal_scale_exif_vs_comment; golden tests
untouched (they pin the comment path)
|
|
All colors moved into CSS custom properties on :root (dark default)
with a prefers-color-scheme: light override block. The two canvas-
drawn surfaces follow too: the viewport now clears transparent so the
CSS canvas background shows through, and the compass strip reads its
colors from the same variables via a new bridge::css_var helper.
theme-color metas (dark + light variants) tint mobile browser chrome
to match.
|
|
Appearing between Display and Track it pushed the whole Track block
down on every hover. As the last item before the bottom-pinned info
texts the flexible spacer absorbs it — nothing jumps anymore.
|
|
The 'web gipfel photogrammetry' tag under the title now links to
flpsed.org/gipfel.html (grey like plain text, accent blue on hover);
the separate 'More information…' row is gone. Hover readout uses the
desktop phrasing: 'Name · 2849 m · distance 12.34 km'.
|
|
hover distance
- control panel starts open on all screen sizes (so the image picker is
reachable on phones); it auto-collapses on narrow screens once an
image is loaded
- 'Fit to window' and the whole Track section only appear when an image
is loaded
- Track section moved directly below Display, above the pinned info
texts (which now carry the margin-top: auto)
- hover readout shows the distance to the peak in km, matching the
desktop format 'name · height m · dist km'
|
|
Superseded by wipfel; its feasibility question (pure-Rust core in the
browser) was answered and documented.
|
|
One text field accepting '47.676,13.32042' (optionally ',alt' as a
third component; commas, semicolons or spaces as separators) — the
format gallery apps show, so the coordinates can simply be copied and
pasted. This replaces the PWA/share-target detour for getting a
viewpoint onto mobile: EXIF GPS may be redacted by the OS, but the
gallery displays the true coordinates for manual transfer.
|
|
This reverts commit 2df51e89039b9b160aafb52b700e874411eb0555.
|
|
- manifest.webmanifest: standalone, dark theme, mountain icons
(regular + maskable)
- sw.js: network-first with cache fallback -> full offline after first
visit (wasm + hill DB included); versioned cache, stale caches pruned
- share_target: images shared to the installed app are POSTed to
__share; the SW stashes the bytes at __shared and redirects with
?shared=1; the app picks them up on startup and cleans the URL
- index.html: manifest link, theme-color, icons, SW registration
- build/serve scripts ship the new files; dist refreshed
Note: sharing does not bypass AOSP GPS redaction when the share source
passes a MediaStore URI (the browser is still the reader) — the
.nomedia/rename trick remains the way to get GPS EXIF into a browser.
|
|
The default Activity theme added a framework action bar that covered
the app's own chrome; the WebView now fills the window and matches the
site background to avoid a white flash.
|
|
Android returns EPERM on ServerSocket creation without it, even for
127.0.0.1 binds. The app still never contacts any other host.
|
|
Thin wrapper so wipfel can read original photo bytes: browsers cannot
declare ACCESS_MEDIA_LOCATION, so the OS always redacts GPS EXIF from
images they read. Our own app declares and requests it (plus
READ_MEDIA_IMAGES) and gets the intact files.
- WebView served from a loopback AssetServer (fetch/WASM need a real
HTTP origin; file:// is blocked)
- file picker wired through WebChromeClient (accepts any file type)
- zero androidx dependencies, plain Activity, minSdk 29
- build.sh: wasm release site -> assets -> assembleDebug (~3MB APK)
- loader: fall back to plain gipfel.dat (aapt gunzips .gz assets)
|
|
MediaProvider only redacts GPS EXIF in files it indexes as images.
A copy renamed to a non-image extension (or under a .nomedia folder)
reaches the browser untouched — an all-files picker makes that route
usable on GrapheneOS/AOSP where browsers cannot hold
ACCESS_MEDIA_LOCATION.
|
|
A UseStateHandle cloned into a callback derefs to its render-time
snapshot (documented Yew behaviour). Pinch emitted zoom and pan as two
separate read-modify-writes per event: both read the same stale
snapshot, so the pan write reverted the zoom write on every event and
the scale never moved. Wheel zoom survived because it emits alone with
a render between events.
Fix: keep the authoritative ViewState in a use_mut_ref mirror with an
apply_view helper (read live, mutate, write mirror + use_state), and
fold pinch zoom+pan into a single combined update. Removes all the
pinch diagnostic probes (debug callback, event counters, build stamp).
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The overlay was suppressed whenever the panel was open — but Firefox
on Android defaults to desktop-site mode, which reports a 980px
viewport and thus keeps the panel open on phones. The overlay now
always shows (it doubles as the pinch-diagnostic readout), and the
initial status carries a build stamp so a stale cached page is
recognizable at a glance.
|