diff --git a/src-tauri/src/app_auto_updater.rs b/src-tauri/src/app_auto_updater.rs
index f66463e..3c3bf37 100644
--- a/src-tauri/src/app_auto_updater.rs
+++ b/src-tauri/src/app_auto_updater.rs
@@ -825,7 +825,10 @@ impl AppAutoUpdater {
// Handle compound extensions like .tar.gz
if file_name.ends_with(".tar.gz") {
- return self.extractor.extract_tar_gz(archive_path, dest_dir).await;
+ return self
+ .extractor
+ .extract_tar_gz(archive_path, dest_dir, None)
+ .await;
}
let extension = archive_path
@@ -837,7 +840,10 @@ impl AppAutoUpdater {
"dmg" => {
#[cfg(target_os = "macos")]
{
- self.extractor.extract_dmg(archive_path, dest_dir).await
+ self
+ .extractor
+ .extract_dmg(archive_path, dest_dir, None)
+ .await
}
#[cfg(not(target_os = "macos"))]
{
@@ -901,7 +907,12 @@ impl AppAutoUpdater {
Err("AppImage installation is only supported on Linux".into())
}
}
- "zip" => self.extractor.extract_zip(archive_path, dest_dir).await,
+ "zip" => {
+ self
+ .extractor
+ .extract_zip(archive_path, dest_dir, None)
+ .await
+ }
_ => Err(format!("Unsupported archive format: {extension}").into()),
}
}
@@ -1110,7 +1121,7 @@ impl AppAutoUpdater {
// Extract ZIP file
let extracted_path = self
.extractor
- .extract_zip(installer_path, &temp_extract_dir)
+ .extract_zip(installer_path, &temp_extract_dir, None)
.await?;
// Find the executable in the extracted files
@@ -1385,7 +1396,7 @@ impl AppAutoUpdater {
// Extract tarball
let extracted_path = self
.extractor
- .extract_tar_gz(tarball_path, &temp_extract_dir)
+ .extract_tar_gz(tarball_path, &temp_extract_dir, None)
.await?;
// Find the executable in the extracted files
diff --git a/src-tauri/src/extraction.rs b/src-tauri/src/extraction.rs
index 8536f92..f21a11c 100644
--- a/src-tauri/src/extraction.rs
+++ b/src-tauri/src/extraction.rs
@@ -45,6 +45,137 @@ fn has_quarantine_attr(path: &Path) -> bool {
result >= 0
}
+/// Best-effort recursive size of a file tree. Uses `symlink_metadata` so
+/// symlinks inside .app bundles are not followed (`cp -R` copies them as
+/// links, so following them would overcount and could loop).
+#[cfg(target_os = "macos")]
+fn dir_size(path: &Path) -> u64 {
+ let Ok(meta) = fs::symlink_metadata(path) else {
+ return 0;
+ };
+ if meta.is_file() {
+ return meta.len();
+ }
+ if !meta.is_dir() {
+ return 0;
+ }
+ let Ok(entries) = fs::read_dir(path) else {
+ return 0;
+ };
+ entries.flatten().map(|entry| dir_size(&entry.path())).sum()
+}
+
+const PROGRESS_REPORT_INTERVAL: std::time::Duration = std::time::Duration::from_millis(150);
+
+/// Emits throttled "extracting" progress on the `download-progress` channel so
+/// the UI can render a moving bar during long extractions.
+pub struct ExtractionReporter {
+ browser: String,
+ version: String,
+ last_emit: std::sync::Mutex,
+}
+
+impl ExtractionReporter {
+ pub fn new(browser: String, version: String) -> Self {
+ // Backdate so the first report goes out immediately.
+ let backdated = std::time::Instant::now()
+ .checked_sub(PROGRESS_REPORT_INTERVAL)
+ .unwrap_or_else(std::time::Instant::now);
+ Self {
+ browser,
+ version,
+ last_emit: std::sync::Mutex::new(backdated),
+ }
+ }
+
+ /// Report byte-level progress where bytes map linearly onto the whole job.
+ pub fn report_bytes(&self, done: u64, total: u64) {
+ if total == 0 {
+ return;
+ }
+ let done = done.min(total);
+ self.report((done as f64 / total as f64) * 100.0, done, Some(total));
+ }
+
+ /// Report a pre-computed percentage (multi-phase extractions where byte
+ /// counts don't map linearly onto overall progress).
+ pub fn report_percentage(&self, percentage: f64) {
+ self.report(percentage, 0, None);
+ }
+
+ /// Force a final 100% event so the bar lands full before the next stage.
+ pub fn finish(&self) {
+ self.emit(100.0, 0, None);
+ }
+
+ fn report(&self, percentage: f64, downloaded_bytes: u64, total_bytes: Option) {
+ {
+ let mut last = self.last_emit.lock().unwrap();
+ if last.elapsed() < PROGRESS_REPORT_INTERVAL {
+ return;
+ }
+ *last = std::time::Instant::now();
+ }
+ self.emit(percentage, downloaded_bytes, total_bytes);
+ }
+
+ fn emit(&self, percentage: f64, downloaded_bytes: u64, total_bytes: Option) {
+ let progress = DownloadProgress {
+ browser: self.browser.clone(),
+ version: self.version.clone(),
+ downloaded_bytes,
+ total_bytes,
+ percentage: percentage.clamp(0.0, 100.0),
+ speed_bytes_per_sec: 0.0,
+ eta_seconds: None,
+ stage: "extracting".to_string(),
+ };
+ let _ = events::emit("download-progress", &progress);
+ }
+}
+
+/// A reader that passes cumulative bytes read to a callback on every read.
+/// Throttling is the reporter's job, so the callback can fire freely.
+struct ProgressReader {
+ inner: R,
+ bytes_read: u64,
+ on_read: F,
+}
+
+impl ProgressReader {
+ fn new(inner: R, on_read: F) -> Self {
+ Self {
+ inner,
+ bytes_read: 0,
+ on_read,
+ }
+ }
+}
+
+impl Read for ProgressReader {
+ fn read(&mut self, buf: &mut [u8]) -> io::Result {
+ let n = self.inner.read(buf)?;
+ self.bytes_read += n as u64;
+ (self.on_read)(self.bytes_read);
+ Ok(n)
+ }
+}
+
+/// Wrap an archive file so compressed bytes consumed report linearly against
+/// its on-disk size — shared by the streaming tar decoders, where stream
+/// position maps monotonically onto overall extraction progress.
+fn progress_file_reader(
+ file: File,
+ progress: Option<&ExtractionReporter>,
+) -> io::Result {
+ let compressed_total = file.metadata()?.len();
+ Ok(ProgressReader::new(file, move |read| {
+ if let Some(reporter) = progress {
+ reporter.report_bytes(read, compressed_total);
+ }
+ }))
+}
+
pub struct Extractor;
impl Extractor {
@@ -145,6 +276,11 @@ impl Extractor {
};
let _ = events::emit("download-progress", &progress);
+ // Reports incremental extraction progress to the UI. Formats without a
+ // measurable byte stream (MSI, plain EXE/AppImage copies) simply never
+ // report, and the frontend falls back to an indeterminate bar.
+ let reporter = ExtractionReporter::new(browser_type.as_str().to_string(), version.to_string());
+
log::info!(
"Starting extraction of {} for browser {} version {}",
archive_path.display(),
@@ -166,7 +302,7 @@ impl Extractor {
"dmg" => {
#[cfg(target_os = "macos")]
{
- self.extract_dmg(archive_path, dest_dir).await.map_err(|e| {
+ self.extract_dmg(archive_path, dest_dir, Some(&reporter)).await.map_err(|e| {
format!("DMG extraction failed for {} {}: {}", browser_type.as_str(), version, e).into()
})
}
@@ -177,22 +313,22 @@ impl Extractor {
}
}
"zip" => {
- self.extract_zip(archive_path, dest_dir).await.map_err(|e| {
+ self.extract_zip(archive_path, dest_dir, Some(&reporter)).await.map_err(|e| {
format!("ZIP extraction failed for {} {}: {}", browser_type.as_str(), version, e).into()
})
}
"tar.xz" => {
- self.extract_tar_xz(archive_path, dest_dir).await.map_err(|e| {
+ self.extract_tar_xz(archive_path, dest_dir, Some(&reporter)).await.map_err(|e| {
format!("TAR.XZ extraction failed for {} {}: {}", browser_type.as_str(), version, e).into()
})
}
"tar.bz2" => {
- self.extract_tar_bz2(archive_path, dest_dir).await.map_err(|e| {
+ self.extract_tar_bz2(archive_path, dest_dir, Some(&reporter)).await.map_err(|e| {
format!("TAR.BZ2 extraction failed for {} {}: {}", browser_type.as_str(), version, e).into()
})
}
"tar.gz" => {
- self.extract_tar_gz(archive_path, dest_dir).await.map_err(|e| {
+ self.extract_tar_gz(archive_path, dest_dir, Some(&reporter)).await.map_err(|e| {
format!("TAR.GZ extraction failed for {} {}: {}", browser_type.as_str(), version, e).into()
})
}
@@ -237,6 +373,8 @@ impl Extractor {
match extraction_result {
Ok(path) => {
+ reporter.finish();
+
// Remove quarantine attributes on macOS to prevent Gatekeeper prompts —
// but only if there's actually something to remove. Calling the
// modify-class `removexattr` syscall on a file without quarantine still
@@ -381,6 +519,7 @@ impl Extractor {
&self,
dmg_path: &Path,
dest_dir: &Path,
+ progress: Option<&ExtractionReporter>,
) -> Result> {
log::info!(
"Extracting DMG: {} to {}",
@@ -454,20 +593,42 @@ impl Extractor {
log::info!("Copying .app to: {}", app_path.display());
+ // The copy is the long pole of DMG extraction; size up the source once so
+ // we can report real progress by polling the destination while cp runs.
+ // report_bytes no-ops on a 0 total, so the None path skips both walks.
+ let total_size = if progress.is_some() {
+ dir_size(&app_entry)
+ } else {
+ 0
+ };
+
// `-X` strips extended attributes (notably com.apple.quarantine) during
// the copy itself. Without it, `cp -R` preserves quarantine from the
// mounted DMG, which then has to be removed with `xattr -dr` — and that
// removexattr syscall on a signed .app bundle trips macOS Sequoia's App
// Management TCC notification ("Donut.app was prevented from modifying
// apps on your Mac"). Stripping at copy time is silent.
- let output = Command::new("cp")
- .args([
- "-RX",
- app_entry.to_str().unwrap(),
- app_path.to_str().unwrap(),
- ])
- .output()
- .await?;
+ let copy_src = app_entry.to_str().unwrap().to_string();
+ let copy_dst = app_path.to_str().unwrap().to_string();
+ let mut copy_task = tokio::spawn(async move {
+ Command::new("cp")
+ .args(["-RX", ©_src, ©_dst])
+ .output()
+ .await
+ });
+
+ let output = loop {
+ tokio::select! {
+ result = &mut copy_task => {
+ break result.map_err(|e| format!("Copy task failed: {e}"))??;
+ }
+ () = tokio::time::sleep(std::time::Duration::from_millis(500)) => {
+ if let Some(reporter) = progress {
+ reporter.report_bytes(dir_size(&app_path), total_size);
+ }
+ }
+ }
+ };
if !output.status.success() {
let stderr = String::from_utf8_lossy(&output.stderr);
@@ -598,6 +759,7 @@ impl Extractor {
&self,
zip_path: &Path,
dest_dir: &Path,
+ progress: Option<&ExtractionReporter>,
) -> Result> {
log::info!("Extracting ZIP archive: {}", zip_path.display());
std::fs::create_dir_all(dest_dir)?;
@@ -610,6 +772,15 @@ impl Extractor {
log::info!("ZIP archive contains {} files", archive.len());
+ // Total uncompressed size, known from the central directory without any
+ // decompression. None for archives using data descriptors — those get no
+ // byte progress (the UI falls back to an indeterminate bar).
+ let total_uncompressed: Option = archive
+ .decompressed_size()
+ .and_then(|total| u64::try_from(total).ok())
+ .filter(|total| *total > 0);
+ let mut extracted_bytes: u64 = 0;
+
for i in 0..archive.len() {
let mut entry = archive
.by_index(i)
@@ -639,8 +810,16 @@ impl Extractor {
let mut outfile = File::create(&outpath)
.map_err(|e| format!("Failed to create file {}: {}", outpath.display(), e))?;
- io::copy(&mut entry, &mut outfile)
+ let entry_size = entry.size();
+ let already_extracted = extracted_bytes;
+ let mut reader = ProgressReader::new(&mut entry, |read| {
+ if let (Some(reporter), Some(total)) = (progress, total_uncompressed) {
+ reporter.report_bytes(already_extracted + read, total);
+ }
+ });
+ io::copy(&mut reader, &mut outfile)
.map_err(|e| format!("Failed to extract file {}: {}", outpath.display(), e))?;
+ extracted_bytes = extracted_bytes.saturating_add(entry_size);
// Set executable permissions on Unix-like systems based on stored mode
#[cfg(unix)]
@@ -670,12 +849,14 @@ impl Extractor {
&self,
tar_path: &Path,
dest_dir: &Path,
+ progress: Option<&ExtractionReporter>,
) -> Result> {
log::info!("Extracting tar.gz archive: {}", tar_path.display());
std::fs::create_dir_all(dest_dir)?;
let file = File::open(tar_path)?;
- let gz_decoder = flate2::read::GzDecoder::new(BufReader::new(file));
+ let counted = progress_file_reader(file, progress)?;
+ let gz_decoder = flate2::read::GzDecoder::new(BufReader::new(counted));
let mut archive = tar::Archive::new(gz_decoder);
archive.unpack(dest_dir)?;
@@ -693,12 +874,14 @@ impl Extractor {
&self,
tar_path: &Path,
dest_dir: &Path,
+ progress: Option<&ExtractionReporter>,
) -> Result> {
log::info!("Extracting tar.bz2 archive: {}", tar_path.display());
std::fs::create_dir_all(dest_dir)?;
let file = File::open(tar_path)?;
- let bz2_decoder = bzip2::read::BzDecoder::new(BufReader::new(file));
+ let counted = progress_file_reader(file, progress)?;
+ let bz2_decoder = bzip2::read::BzDecoder::new(BufReader::new(counted));
let mut archive = tar::Archive::new(bz2_decoder);
archive.unpack(dest_dir)?;
@@ -716,6 +899,7 @@ impl Extractor {
&self,
tar_path: &Path,
dest_dir: &Path,
+ progress: Option<&ExtractionReporter>,
) -> Result> {
log::info!("Extracting tar.xz archive: {}", tar_path.display());
std::fs::create_dir_all(dest_dir)?;
@@ -726,17 +910,34 @@ impl Extractor {
// Read the entire file into memory for lzma-rs
let mut compressed_data = Vec::new();
buf_reader.read_to_end(&mut compressed_data)?;
+ let compressed_total = compressed_data.len() as u64;
- // Decompress using lzma-rs
+ // Two phases with no shared byte scale: CPU-bound LZMA decompression
+ // dominates, so compressed bytes consumed map to 0–80%, and the tar
+ // unpack of the decompressed data to 80–100%.
let mut decompressed_data = Vec::new();
- lzma_rs::xz_decompress(
- &mut std::io::Cursor::new(compressed_data),
- &mut decompressed_data,
- )?;
+ let counted_input = ProgressReader::new(std::io::Cursor::new(compressed_data), |read| {
+ if let Some(reporter) = progress {
+ if compressed_total > 0 {
+ reporter
+ .report_percentage(80.0 * read.min(compressed_total) as f64 / compressed_total as f64);
+ }
+ }
+ });
+ lzma_rs::xz_decompress(&mut BufReader::new(counted_input), &mut decompressed_data)?;
// Create tar archive from decompressed data
- let cursor = std::io::Cursor::new(decompressed_data);
- let mut archive = tar::Archive::new(cursor);
+ let decompressed_total = decompressed_data.len() as u64;
+ let counted_tar = ProgressReader::new(std::io::Cursor::new(decompressed_data), |read| {
+ if let Some(reporter) = progress {
+ if decompressed_total > 0 {
+ reporter.report_percentage(
+ 80.0 + 20.0 * read.min(decompressed_total) as f64 / decompressed_total as f64,
+ );
+ }
+ }
+ });
+ let mut archive = tar::Archive::new(counted_tar);
archive.unpack(dest_dir)?;
@@ -1472,6 +1673,86 @@ mod tests {
assert_eq!(result.unwrap(), "msi");
}
+ #[tokio::test]
+ async fn test_extract_zip_reports_progress() {
+ use std::sync::{Arc, Mutex};
+
+ #[derive(Default)]
+ struct CapturingEmitter(Mutex>);
+
+ impl crate::events::EventEmitter for CapturingEmitter {
+ fn emit_value(&self, event: &str, payload: serde_json::Value) -> Result<(), String> {
+ if event == "download-progress" {
+ self.0.lock().unwrap().push(payload);
+ }
+ Ok(())
+ }
+ }
+
+ let captured = Arc::new(CapturingEmitter::default());
+ // The global emitter can only be set once per process; if another test ever
+ // claims it first we lose observability, so only assert on captured events
+ // when this test's emitter actually won.
+ let emitter_installed = crate::events::set_global_emitter(captured.clone()).is_ok();
+
+ let extractor = Extractor::instance();
+ let temp_dir = TempDir::new().expect("Failed to create temp directory");
+ let dest_dir = temp_dir.path().join("extracted");
+
+ // A payload comfortably larger than io::copy's 8KB chunks so the counting
+ // reader fires multiple times.
+ let payload = vec![0x42u8; 256 * 1024];
+ let zip_path = temp_dir.path().join("test.zip");
+ {
+ let file = std::fs::File::create(&zip_path).expect("Failed to create test zip file");
+ let mut zip = zip::ZipWriter::new(file);
+ let options =
+ zip::write::FileOptions::<()>::default().compression_method(zip::CompressionMethod::Stored);
+ zip
+ .start_file("data.bin", options)
+ .expect("Failed to start zip file");
+ zip.write_all(&payload).expect("Failed to write to zip");
+ zip.finish().expect("Failed to finish zip");
+ }
+
+ let reporter =
+ ExtractionReporter::new("test-browser-zip-progress".to_string(), "1.0.0".to_string());
+ let result = extractor
+ .extract_zip(&zip_path, &dest_dir, Some(&reporter))
+ .await;
+
+ // Extraction itself must have worked even if no executable is found.
+ assert!(dest_dir.join("data.bin").exists(), "payload not extracted");
+ if let Err(e) = result {
+ assert!(
+ e.to_string().contains("executable"),
+ "unexpected extraction error: {e}"
+ );
+ }
+
+ if emitter_installed {
+ let events = captured.0.lock().unwrap();
+ let ours: Vec<_> = events
+ .iter()
+ .filter(|p| p["browser"] == "test-browser-zip-progress")
+ .collect();
+ assert!(
+ !ours.is_empty(),
+ "expected at least one extracting progress event"
+ );
+ for p in &ours {
+ assert_eq!(p["stage"], "extracting");
+ assert_eq!(p["version"], "1.0.0");
+ let pct = p["percentage"].as_f64().unwrap();
+ assert!(
+ (0.0..=100.0).contains(&pct),
+ "percentage out of range: {pct}"
+ );
+ assert_eq!(p["total_bytes"].as_u64(), Some(payload.len() as u64));
+ }
+ }
+ }
+
#[tokio::test]
async fn test_extract_zip_with_test_archive() {
let extractor = Extractor::instance();
@@ -1496,7 +1777,7 @@ mod tests {
zip.finish().expect("Failed to finish zip");
}
- let result = extractor.extract_zip(&zip_path, &dest_dir).await;
+ let result = extractor.extract_zip(&zip_path, &dest_dir, None).await;
// The result might fail because we're looking for executables, but the extraction should work
// Let's check if the file was extracted regardless of the result
@@ -1545,7 +1826,9 @@ mod tests {
tar.finish().expect("Failed to finish tar");
}
- let result = extractor.extract_tar_gz(&tar_gz_path, &dest_dir).await;
+ let result = extractor
+ .extract_tar_gz(&tar_gz_path, &dest_dir, None)
+ .await;
// Check if the file was extracted
let extracted_file = dest_dir.join("test.txt");
@@ -1596,7 +1879,9 @@ mod tests {
tar.finish().expect("Failed to finish tar");
}
- let result = extractor.extract_tar_bz2(&tar_bz2_path, &dest_dir).await;
+ let result = extractor
+ .extract_tar_bz2(&tar_bz2_path, &dest_dir, None)
+ .await;
// Check if the file was extracted
let extracted_file = dest_dir.join("test.txt");
@@ -1657,7 +1942,9 @@ mod tests {
.expect("Failed to write compressed data");
}
- let result = extractor.extract_tar_xz(&tar_xz_path, &dest_dir).await;
+ let result = extractor
+ .extract_tar_xz(&tar_xz_path, &dest_dir, None)
+ .await;
// Check if the file was extracted
let extracted_file = dest_dir.join("test.txt");
diff --git a/src/components/custom-toast.tsx b/src/components/custom-toast.tsx
index a165abd..97702ac 100644
--- a/src/components/custom-toast.tsx
+++ b/src/components/custom-toast.tsx
@@ -159,6 +159,23 @@ function formatEtaCompact(seconds: number): string {
return `${Math.round(seconds)}s`;
}
+function ProgressBar({
+ percentage,
+ className = "w-full",
+}: {
+ percentage: number;
+ className?: string;
+}) {
+ return (
+
+ );
+}
+
function getToastIcon(type: ToastProps["type"], stage?: string) {
switch (type) {
case "success":
@@ -232,12 +249,31 @@ export function UnifiedToast(props: ToastProps) {
` • ${t("toasts.progress.remaining", { time: progress.eta })}`}
-
+
+
+ )}
+
+ {/* Extraction / verification progress. Extraction reports a real
+ percentage for most archive formats; when none is available yet
+ (or the format can't measure progress) show an indeterminate bar. */}
+ {type === "download" &&
+ (stage === "extracting" || stage === "verifying") && (
+
+ {stage === "extracting" &&
+ progress &&
+ "percentage" in progress &&
+ progress.percentage > 0 ? (
+ <>
+
+ {progress.percentage.toFixed(1)}%
+
+
+ >
+ ) : (
+
+ )}
)}
@@ -253,14 +289,10 @@ export function UnifiedToast(props: ToastProps) {
})}
-
+
{progress.current}/{progress.total}
diff --git a/src/hooks/use-browser-download.ts b/src/hooks/use-browser-download.ts
index 32bb104..90496d0 100644
--- a/src/hooks/use-browser-download.ts
+++ b/src/hooks/use-browser-download.ts
@@ -354,7 +354,14 @@ export function useBrowserDownload() {
}
} else if (progress.stage === "extracting") {
if (!isOnboardingActive()) {
- showDownloadToast(browserName, progress.version, "extracting");
+ showDownloadToast(
+ browserName,
+ progress.version,
+ "extracting",
+ progress.percentage > 0
+ ? { percentage: progress.percentage }
+ : undefined,
+ );
}
} else if (progress.stage === "verifying") {
if (!isOnboardingActive()) {
diff --git a/src/hooks/use-browser-setup.ts b/src/hooks/use-browser-setup.ts
index e2dd839..987d999 100644
--- a/src/hooks/use-browser-setup.ts
+++ b/src/hooks/use-browser-setup.ts
@@ -27,10 +27,11 @@ export interface SetupError {
stage: SetupErrorStage;
}
-// The backend emits a real percentage only while downloading; extraction sends
-// a single "extracting" event with no incremental progress (it takes ~2 min).
-// So we estimate extraction progress from elapsed time vs. a learned average,
-// seeded at 2 minutes and refined with the real durations we record.
+// The backend reports real extraction percentages for most archive formats
+// (zip, tar.*, dmg). For formats that can't measure progress (e.g. MSI) the
+// "extracting" events carry percentage 0, so we fall back to estimating from
+// elapsed time vs. a learned average, seeded at 2 minutes and refined with
+// the real durations we record.
const DEFAULT_EXTRACT_MS = 2 * 60 * 1000;
const MAX_SAMPLES = 5; // the 2-min seed + up to 4 most recent real durations
@@ -89,8 +90,9 @@ function toErrorStage(stage: string): SetupErrorStage {
}
/**
- * Tracks first-launch setup of a browser: real download progress plus an
- * estimated extraction progress (no countdown timer, percentages only).
+ * Tracks first-launch setup of a browser: real download progress plus
+ * extraction progress — real backend percentages when the archive format
+ * supports them, otherwise a time-based estimate (percentages only).
* `active` should be true while the owning dialog is open.
*/
export function useBrowserSetup(browser: string, active: boolean) {
@@ -108,8 +110,12 @@ export function useBrowserSetup(browser: string, active: boolean) {
const extractStartRef = useRef
(null);
const estimateRef = useRef(DEFAULT_EXTRACT_MS);
- // Fallback bookkeeping so a listener that mounts mid-flight (and therefore
- // misses the single "extracting" event) can still show extraction progress.
+ // True once an "extracting" event carried a real percentage — from then on
+ // the backend drives the bar and the time-based estimate stays out of it.
+ const sawRealExtractionRef = useRef(false);
+ // Fallback bookkeeping so a listener that mounts mid-flight, or that only
+ // ever receives percentage-0 "extracting" events (formats that can't
+ // measure progress), can still show extraction progress.
const sawDownloadingRef = useRef(false);
const lastProgressAtRef = useRef(null);
const lastDownloadPercentRef = useRef(0);
@@ -133,6 +139,7 @@ export function useBrowserSetup(browser: string, active: boolean) {
setExtractionOvertime(false);
setError(null);
extractStartRef.current = null;
+ sawRealExtractionRef.current = false;
sawDownloadingRef.current = false;
lastProgressAtRef.current = null;
lastDownloadPercentRef.current = 0;
@@ -143,6 +150,7 @@ export function useBrowserSetup(browser: string, active: boolean) {
let alive = true;
estimateRef.current = average(readDurations(browser));
extractStartRef.current = null;
+ sawRealExtractionRef.current = false;
sawDownloadingRef.current = false;
lastProgressAtRef.current = null;
lastDownloadPercentRef.current = 0;
@@ -182,6 +190,11 @@ export function useBrowserSetup(browser: string, active: boolean) {
}
lastProgressAtRef.current = Date.now();
setPhase("extracting");
+ if (p.percentage > 0) {
+ sawRealExtractionRef.current = true;
+ setExtractionPercent(Math.min(99, Math.round(p.percentage)));
+ setExtractionOvertime(false);
+ }
break;
case "verifying":
lastStageRef.current = "verifying";
@@ -257,9 +270,9 @@ export function useBrowserSetup(browser: string, active: boolean) {
// Drive the estimated extraction percentage while extracting.
const tick = setInterval(() => {
if (!alive || doneRef.current) return;
- // If the download visibly finished but we never saw the (single)
- // "extracting" event, start estimating extraction anyway — anchored to
- // the last download event, which is roughly when extraction began.
+ // If the download visibly finished but we never saw any "extracting"
+ // event, start estimating extraction anyway — anchored to the last
+ // download event, which is roughly when extraction began.
if (
extractStartRef.current == null &&
sawDownloadingRef.current &&
@@ -272,6 +285,9 @@ export function useBrowserSetup(browser: string, active: boolean) {
setPhase("extracting");
}
if (extractStartRef.current == null) return;
+ // Real backend percentages drive the bar; the estimate would only
+ // fight them (and flag bogus "overtime" on a healthy extraction).
+ if (sawRealExtractionRef.current) return;
const elapsed = Date.now() - extractStartRef.current;
const est = estimateRef.current || DEFAULT_EXTRACT_MS;
if (elapsed >= est) {
@@ -310,6 +326,7 @@ export function useBrowserSetup(browser: string, active: boolean) {
setExtractionOvertime(false);
setError(null);
extractStartRef.current = null;
+ sawRealExtractionRef.current = false;
sawDownloadingRef.current = false;
lastProgressAtRef.current = null;
lastDownloadPercentRef.current = 0;
diff --git a/src/styles/globals.css b/src/styles/globals.css
index bd57799..84dd59b 100644
--- a/src/styles/globals.css
+++ b/src/styles/globals.css
@@ -6,6 +6,17 @@
@theme {
--font-sans: var(--font-geist-sans);
--font-mono: var(--font-geist-mono);
+ --animate-progress-indeterminate: progress-indeterminate 1.4s ease-in-out
+ infinite;
+
+ @keyframes progress-indeterminate {
+ 0% {
+ transform: translateX(-100%);
+ }
+ 100% {
+ transform: translateX(300%);
+ }
+ }
}
@theme inline {