Release / Cria a release (push) Canceled after 0s
Release / CLI (linux-aarch64) (push) Canceled after 0s
Release / CLI (linux-x86_64) (push) Canceled after 0s
Release / CLI (macos-aarch64) (push) Canceled after 0s
Release / CLI (macos-x86_64) (push) Canceled after 0s
Release / CLI (windows-x86_64) (push) Canceled after 0s
Release / App desktop (linux-x86_64) (push) Canceled after 0s
Release / App desktop (macos-aarch64) (push) Canceled after 0s
Release / App desktop (macos-x86_64) (push) Canceled after 0s
Release / App desktop (windows-x86_64) (push) Canceled after 0s
Release / Publica a release (push) Canceled after 0s
- Explorador com grade/lista virtualizada, navegação por tipo, data EXIF, pasta e câmera, seleção estilo gerenciador de arquivos e visualização rápida. - Miniaturas no núcleo, metadados (EXIF, PDF, Office, MP4), datas MACB do NTFS/FAT/exFAT. - Hashes MD5/SHA-1/SHA-256, busca de texto no disco, exportação CSV, relatório HTML, detecção de duplicadas, tema, correções de interface e textos sem travessões. Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
351 lines
11 KiB
Rust
351 lines
11 KiB
Rust
//! Orquestra a varredura: layout → metadados dos sistemas de arquivos →
|
|
//! carving no espaço livre/desconhecido → fusão sem duplicatas.
|
|
|
|
use crate::carve::{carve_range, CarveCfg, CarveEvents, HitAction};
|
|
use crate::fs::{self, merge_extents, overlaps, Ctx};
|
|
use crate::partition;
|
|
use crate::signatures::category_for_ext;
|
|
use crate::source::Source;
|
|
use crate::types::*;
|
|
use serde::{Deserialize, Serialize};
|
|
use std::collections::HashMap;
|
|
use std::sync::atomic::AtomicU64;
|
|
|
|
#[derive(Debug, Clone, Serialize, Deserialize, Default)]
|
|
pub struct Progress {
|
|
pub phase: String,
|
|
pub done: u64,
|
|
pub total: u64,
|
|
}
|
|
|
|
/// Receptor dos resultados (a interface/CLI implementa).
|
|
pub trait Sink {
|
|
fn found(&mut self, f: Found);
|
|
/// Um achado de carving confirmou um arquivo já listado por metadados.
|
|
fn confirm(&mut self, id: u64, confidence: Confidence);
|
|
fn progress(&mut self, p: Progress);
|
|
fn warning(&mut self, w: String);
|
|
fn layout(&mut self, l: &DiskLayout);
|
|
}
|
|
|
|
#[derive(Debug, Clone, Serialize, Deserialize)]
|
|
pub struct RangeState {
|
|
pub start: u64,
|
|
pub end: u64,
|
|
pub next: u64,
|
|
}
|
|
|
|
/// Estado serializável para pausar/retomar.
|
|
#[derive(Debug, Clone, Serialize, Deserialize, Default)]
|
|
pub struct Checkpoint {
|
|
pub fs_done: bool,
|
|
pub ranges: Vec<RangeState>,
|
|
pub next_id: u64,
|
|
pub layout: Option<DiskLayout>,
|
|
}
|
|
|
|
pub enum Outcome {
|
|
Completed,
|
|
Interrupted(Checkpoint),
|
|
}
|
|
|
|
pub fn run(
|
|
src: &dyn Source,
|
|
opts: &ScanOptions,
|
|
resume: Option<Checkpoint>,
|
|
existing: &[Found],
|
|
cancelled: &(dyn Fn() -> bool + Sync),
|
|
sink: &mut dyn Sink,
|
|
) -> Outcome {
|
|
let mut cp = resume.unwrap_or_default();
|
|
let next_id = AtomicU64::new(
|
|
cp.next_id
|
|
.max(existing.iter().map(|f| f.id + 1).max().unwrap_or(1))
|
|
.max(1),
|
|
);
|
|
|
|
sink.progress(Progress {
|
|
phase: "Lendo tabela de partições".into(),
|
|
done: 0,
|
|
total: 0,
|
|
});
|
|
let layout = match cp.layout.clone() {
|
|
Some(l) => l,
|
|
None => {
|
|
let l = partition::analyze(src, opts.mode == ScanMode::Deep);
|
|
sink.layout(&l);
|
|
for w in &l.warnings {
|
|
sink.warning(w.clone());
|
|
}
|
|
l
|
|
}
|
|
};
|
|
cp.layout = Some(layout.clone());
|
|
|
|
// Índices para fusão: primeiro offset -> (id, confiança), e extents de achados por metadados.
|
|
let mut starts: HashMap<u64, (u64, Confidence)> = HashMap::new();
|
|
let mut fs_extents: Vec<Extent> = vec![];
|
|
for f in existing {
|
|
index_found(f, &mut starts, &mut fs_extents);
|
|
}
|
|
|
|
if !cp.fs_done {
|
|
let mut live_all: Vec<Extent> = vec![];
|
|
let mut free_all: Vec<Extent> = vec![];
|
|
let mut have_free_for: Vec<(u64, u64)> = vec![];
|
|
for p in &layout.partitions {
|
|
if p.encrypted.is_some() {
|
|
continue;
|
|
}
|
|
let phase = format!("Lendo metadados {} (partição {})", p.fs.name(), p.index + 1);
|
|
sink.progress(Progress {
|
|
phase: phase.clone(),
|
|
done: 0,
|
|
total: 0,
|
|
});
|
|
let mut emit_buf: Vec<Found> = vec![];
|
|
let res = {
|
|
let progress_phase = phase.clone();
|
|
let prog_cell = std::sync::Mutex::new((0u64, 0u64));
|
|
let progress = |d: u64, t: u64| {
|
|
*prog_cell.lock().unwrap() = (d, t);
|
|
};
|
|
let mut emit = |f: Found| emit_buf.push(f);
|
|
let mut ctx = Ctx {
|
|
src,
|
|
next_id: &next_id,
|
|
cancelled,
|
|
progress: &progress,
|
|
emit: &mut emit,
|
|
};
|
|
let r = match p.fs {
|
|
FsKind::Ntfs => Some(fs::ntfs::scan(&mut ctx, p.start)),
|
|
FsKind::Fat12 | FsKind::Fat16 | FsKind::Fat32 => {
|
|
Some(fs::fat::scan(&mut ctx, p.start, p.fs))
|
|
}
|
|
FsKind::Exfat => Some(fs::exfat::scan(&mut ctx, p.start)),
|
|
FsKind::Ext => Some(fs::ext::scan(&mut ctx, p.start, p.len)),
|
|
_ => None,
|
|
};
|
|
let _ = progress_phase;
|
|
r
|
|
};
|
|
for f in emit_buf {
|
|
index_found(&f, &mut starts, &mut fs_extents);
|
|
sink.found(f);
|
|
}
|
|
if let Some(r) = res {
|
|
for w in r.warnings {
|
|
sink.warning(w);
|
|
}
|
|
live_all.extend(r.live);
|
|
if let Some(fr) = r.free {
|
|
free_all.extend(fr);
|
|
have_free_for.push((p.start, p.start + p.len));
|
|
}
|
|
}
|
|
if cancelled() {
|
|
cp.next_id = next_id.load(std::sync::atomic::Ordering::Relaxed);
|
|
return Outcome::Interrupted(cp);
|
|
}
|
|
}
|
|
cp.fs_done = true;
|
|
cp.ranges = if opts.mode == ScanMode::Deep {
|
|
build_ranges(
|
|
&layout,
|
|
&live_all,
|
|
&free_all,
|
|
&have_free_for,
|
|
opts,
|
|
src.size(),
|
|
)
|
|
} else {
|
|
vec![]
|
|
};
|
|
}
|
|
fs_extents = merge_extents(fs_extents);
|
|
|
|
let total: u64 = cp.ranges.iter().map(|r| r.end - r.start).sum();
|
|
let done: u64 = cp
|
|
.ranges
|
|
.iter()
|
|
.map(|r| r.next.saturating_sub(r.start).min(r.end - r.start))
|
|
.sum();
|
|
let cfg = CarveCfg {
|
|
max_size: opts.max_carve_size,
|
|
categories: &opts.categories,
|
|
};
|
|
let mut handler = Handler {
|
|
sink,
|
|
starts,
|
|
fs_extents,
|
|
next_id: &next_id,
|
|
done,
|
|
total,
|
|
};
|
|
for ri in 0..cp.ranges.len() {
|
|
let (start, end) = (cp.ranges[ri].start, cp.ranges[ri].end);
|
|
let mut next = cp.ranges[ri].next;
|
|
if next >= end {
|
|
continue;
|
|
}
|
|
let finished = carve_range(src, start, end, &mut next, &cfg, cancelled, &mut handler);
|
|
cp.ranges[ri].next = next;
|
|
if !finished {
|
|
cp.next_id = next_id.load(std::sync::atomic::Ordering::Relaxed);
|
|
return Outcome::Interrupted(cp);
|
|
}
|
|
}
|
|
Outcome::Completed
|
|
}
|
|
|
|
struct Handler<'a> {
|
|
sink: &'a mut dyn Sink,
|
|
starts: HashMap<u64, (u64, Confidence)>,
|
|
fs_extents: Vec<Extent>,
|
|
next_id: &'a AtomicU64,
|
|
done: u64,
|
|
total: u64,
|
|
}
|
|
|
|
impl CarveEvents for Handler<'_> {
|
|
fn advance(&mut self, n: u64) {
|
|
self.done += n;
|
|
self.sink.progress(Progress {
|
|
phase: "Varredura profunda por assinatura".into(),
|
|
done: self.done.min(self.total),
|
|
total: self.total,
|
|
});
|
|
}
|
|
|
|
fn hit(&mut self, off: u64, hit: &crate::signatures::Hit) -> HitAction {
|
|
if let Some(&(id, conf)) = self.starts.get(&off) {
|
|
if hit.confidence > conf {
|
|
self.sink.confirm(id, hit.confidence);
|
|
self.starts.insert(off, (id, hit.confidence));
|
|
}
|
|
return HitAction::Skip;
|
|
}
|
|
if overlaps(&self.fs_extents, off, 1) {
|
|
return HitAction::Continue; // dentro de arquivo já listado (ex.: miniatura embutida)
|
|
}
|
|
let id = self
|
|
.next_id
|
|
.fetch_add(1, std::sync::atomic::Ordering::Relaxed);
|
|
let f = Found {
|
|
id,
|
|
name: format!("arquivo_{id:06}.{}", hit.ext),
|
|
dir: None,
|
|
ext: hit.ext.to_string(),
|
|
category: category_for_ext(hit.ext),
|
|
size: hit.len,
|
|
origin: Origin::Carved,
|
|
confidence: hit.confidence,
|
|
data: DataLoc::Extents {
|
|
extents: vec![Extent::new(off, hit.len)],
|
|
},
|
|
modified: None,
|
|
created: None,
|
|
accessed: None,
|
|
taken: None,
|
|
camera: None,
|
|
px: None,
|
|
gps: false,
|
|
sig: None,
|
|
note: if hit.confidence == Confidence::Low {
|
|
Some("sem rodapé/estrutura completa: pode estar truncado".into())
|
|
} else {
|
|
None
|
|
},
|
|
};
|
|
self.starts.insert(off, (id, hit.confidence));
|
|
self.sink.found(f);
|
|
HitAction::Skip
|
|
}
|
|
}
|
|
|
|
fn index_found(f: &Found, starts: &mut HashMap<u64, (u64, Confidence)>, ex: &mut Vec<Extent>) {
|
|
if let DataLoc::Extents { extents } = &f.data {
|
|
if let Some(o) = f.first_offset() {
|
|
starts.insert(o, (f.id, f.confidence));
|
|
}
|
|
ex.extend(extents.iter().filter(|e| !e.sparse).copied());
|
|
}
|
|
}
|
|
|
|
fn build_ranges(
|
|
layout: &DiskLayout,
|
|
live: &[Extent],
|
|
free: &[Extent],
|
|
have_free_for: &[(u64, u64)],
|
|
opts: &ScanOptions,
|
|
disk_size: u64,
|
|
) -> Vec<RangeState> {
|
|
let live = merge_extents(live.to_vec());
|
|
let free = merge_extents(free.to_vec());
|
|
let mut out: Vec<Extent> = vec![];
|
|
for p in &layout.partitions {
|
|
if p.encrypted.is_some() {
|
|
continue;
|
|
}
|
|
let whole = Extent::new(p.start, p.len.min(disk_size.saturating_sub(p.start)));
|
|
if opts.include_allocated {
|
|
out.push(whole);
|
|
} else if have_free_for
|
|
.iter()
|
|
.any(|&(s, e)| s == p.start && e == p.start + p.len)
|
|
{
|
|
out.extend(clip(&free, whole));
|
|
} else if p.fs.has_metadata_scanner() {
|
|
out.extend(subtract(whole, &live));
|
|
} else {
|
|
out.push(whole);
|
|
}
|
|
}
|
|
for g in &layout.gaps {
|
|
out.push(Extent::new(g.start, g.len));
|
|
}
|
|
if layout.partitions.is_empty() && layout.gaps.is_empty() {
|
|
out.push(Extent::new(0, disk_size));
|
|
}
|
|
merge_extents(out)
|
|
.into_iter()
|
|
.filter(|e| e.len >= 4096)
|
|
.map(|e| RangeState {
|
|
start: e.offset,
|
|
end: e.end(),
|
|
next: e.offset,
|
|
})
|
|
.collect()
|
|
}
|
|
|
|
fn clip(sorted: &[Extent], window: Extent) -> Vec<Extent> {
|
|
sorted
|
|
.iter()
|
|
.filter_map(|e| {
|
|
let s = e.offset.max(window.offset);
|
|
let t = e.end().min(window.end());
|
|
(t > s).then(|| Extent::new(s, t - s))
|
|
})
|
|
.collect()
|
|
}
|
|
|
|
fn subtract(window: Extent, sorted_live: &[Extent]) -> Vec<Extent> {
|
|
let mut out = vec![];
|
|
let mut cur = window.offset;
|
|
let i = sorted_live.partition_point(|e| e.end() <= window.offset);
|
|
for e in &sorted_live[i..] {
|
|
if e.offset >= window.end() {
|
|
break;
|
|
}
|
|
if e.offset > cur {
|
|
out.push(Extent::new(cur, e.offset - cur));
|
|
}
|
|
cur = cur.max(e.end());
|
|
}
|
|
if cur < window.end() {
|
|
out.push(Extent::new(cur, window.end() - cur));
|
|
}
|
|
out
|
|
}
|