Compare commits

..

1 Commits

Author SHA1 Message Date
Xavier Roche
3a20de8ec4 Raise the test server's listen backlog so -c8 bigcrawl stops flaking on macOS
local-server.py used ThreadingHTTPServer's default request_queue_size of 5.
macOS/BSD drop SYNs once the accept queue overflows; Linux is lenient. Under a
parallel make check the single-threaded accept loop is starved for CPU, drains
the 5-deep queue slowly, and an -c8 bigcrawl burst overflows it. The
36_local-bigcrawl file count then flaked (e.g. 361 -> 283) because --retries=0
turns each dropped connection into a lost file. Raise the backlog to 128 for
ample headroom; -c8 is unchanged.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Signed-off-by: Xavier Roche <roche@httrack.com>
2026-07-12 10:36:15 +02:00
11 changed files with 27 additions and 267 deletions

View File

@@ -135,52 +135,21 @@ jobs:
- name: Build
run: make -j"$(sysctl -n hw.ncpu)"
- name: Add loopback aliases (macOS lacks 127.0.0.2/.3)
# 19_local-connect-fallback needs the dead 127.0.0.2/.3 to refuse
# instantly like Linux; alias them onto lo0 so they don't stall to timeout.
run: |
set -euo pipefail
sudo ifconfig lo0 alias 127.0.0.2 up
sudo ifconfig lo0 alias 127.0.0.3 up
- name: Test
# bigcrawl's sustained -c8 crawl drops fetches on macOS's loopback when
# it competes with other crawls, flaking its exact file count (the #527
# macOS drop). Run everything else in parallel, then bigcrawl alone (its
# serial-safe condition). Linux tolerates the full parallel run.
run: |
jobs=$(( $(sysctl -n hw.ncpu) * 2 )); [ "$jobs" -le 16 ] || jobs=16
make check -j"$jobs"
rest=$(cd tests && ls *.test | grep -v '^36_local-bigcrawl\.test$' | tr '\n' ' ')
make check -j"$jobs" TESTS="$rest"
make check TESTS=36_local-bigcrawl.test
- name: Print the test log on failure
if: failure()
run: cat tests/test-suite.log 2>/dev/null || true
# Runtime smoke of the WebHTTrack launcher on macOS: it carries a Darwin-only
# browser path (open -W) and nothing else exercises htsserver. Install into a
# temp prefix, then check webhttrack brings up htsserver and serves the UI.
webhttrack-macos:
name: webhttrack smoke (macOS arm64)
runs-on: macos-14
steps:
- uses: actions/checkout@v6
with:
submodules: recursive
- name: Install build dependencies
run: |
set -euo pipefail
brew install autoconf automake libtool autoconf-archive
- name: Build and install into a temp prefix
run: |
set -euo pipefail
ssl="$(brew --prefix openssl@3)"
autoreconf -fi
./configure CPPFLAGS="-I${ssl}/include" LDFLAGS="-L${ssl}/lib" \
--prefix="$RUNNER_TEMP/inst"
make -j"$(sysctl -n hw.ncpu)"
make install
- name: Smoke-test webhttrack
run: bash tests/webhttrack-smoke.sh "$RUNNER_TEMP/inst"
# Portability/hardening: 32-bit (i386) build on the x86-64 runner via multilib
# -- no extra hardware. Exercises the 32-bit size_t/pointer ABI, where size
# and bounds math can truncate or wrap in ways 64-bit never reveals (the axis

View File

@@ -14,10 +14,10 @@ the operational checklist: toolchain, invariants, and how to ship a change.
automatically; only a test slower than the current longest raises the floor.
On a few-core Linux box, `-j` at 2x the core count is faster still: the tests
spend much of their wall time asleep (server trickles, httrack self-pacing),
so an idle core covers a sleeping one. CI uses `min(2*cores, 16)` on every
platform, macOS included: the test server raises its listen backlog
(`request_queue_size`) so macOS/BSD don't drop connections under a parallel
`-c16` bigcrawl the way Python's default backlog of 5 did.
so an idle core covers a sleeping one. CI uses `min(2*cores, 16)`. macOS runs
36_local-bigcrawl alone in a second pass: its sustained `-c8` crawl overloads
the macOS loopback when it competes with other crawls and flakes its exact
file count (Linux tolerates the full parallel run).
Or run `sh build.sh` to do bootstrap + configure + make in one shot.
## Hard invariants

View File

@@ -121,18 +121,12 @@ int fa_strjoker_dual(int type, char **filters, int nfil, const char *nom1,
return use1 ? jok1 : jok2;
}
/* Real filters/URLs fit HTS_URLMAXSIZE*2; past it a hostile pattern recurses
O(len) deep or runs O(n^2*stars). Cap length (depth) and steps (work). */
#define STRJOKER_MAXLEN (HTS_URLMAXSIZE * 2)
#define STRJOKER_MAXSTEPS 2000000u
/* Failure memo for the recursive matcher: one bit per (chaine, joker) offset
pair keeps star-heavy patterns polynomial instead of exponential (#501). */
typedef struct strjoker_memo {
const char *chaine0, *joker0; /* offsets are relative to these bases */
size_t stride; /* strlen(joker0) + 1 */
unsigned char *failed; /* failed-pair bitmap; NULL: no memo */
size_t *nsteps; /* shared work counter; NULL: unbounded (oracle) */
} strjoker_memo;
static const char *strjoker_impl(const strjoker_memo *memo, const char *chaine,
@@ -147,8 +141,6 @@ static const char *strjoker_rec(const strjoker_memo *memo, const char *chaine,
size_t bit = 0;
const char *adr;
if (memo->nsteps != NULL && ++*memo->nsteps > STRJOKER_MAXSTEPS)
return NULL; /* work budget spent: fail the match safely */
if (memo->failed) {
bit = (size_t) (chaine - memo->chaine0) * memo->stride +
(size_t) (joker - memo->joker0);
@@ -161,24 +153,22 @@ static const char *strjoker_rec(const strjoker_memo *memo, const char *chaine,
return adr;
}
/* Match chaine against joker with a shared work budget *nsteps (a strjokerfind
sweep passes one counter, bounding the whole scan). */
static const char *strjoker_bounded(const char *chaine, const char *joker,
LLint *size, int *size_flag,
size_t *nsteps) {
strjoker_memo memo = {chaine, joker, 0, NULL, nsteps};
// wildcard comparator: match chaine against joker (pattern), case-insensitive
// returns the address of the first matched letter past any leading joker
// (ie. *[..]toto.. returns the address of toto within chaine), NULL on mismatch
HTS_INLINE const char *strjoker(const char *chaine, const char *joker,
LLint *size, int *size_flag) {
strjoker_memo memo = {chaine, joker, 0, NULL};
unsigned char stackbits[2048];
hts_boolean onheap = HTS_FALSE;
const char *adr;
if (chaine != NULL && joker != NULL) {
const size_t l1 = strlen(chaine), l2 = strlen(joker);
const size_t n1 = strlen(chaine) + 1;
if (l1 > STRJOKER_MAXLEN || l2 > STRJOKER_MAXLEN)
return NULL; /* beyond any real filter/URL: bound depth+work */
memo.stride = l2 + 1;
if (l1 + 1 <= (SIZE_MAX - 7) / memo.stride) { // overflow-safe bitmap size
const size_t bytes = ((l1 + 1) * memo.stride + 7) / 8;
memo.stride = strlen(joker) + 1;
if (n1 <= (SIZE_MAX - 7) / memo.stride) { // overflow-safe bitmap size
const size_t bytes = (n1 * memo.stride + 7) / 8;
if (bytes <= sizeof(stackbits)) {
memset(stackbits, 0, bytes);
@@ -195,16 +185,6 @@ static const char *strjoker_bounded(const char *chaine, const char *joker,
return adr;
}
// wildcard comparator: match chaine against joker (pattern), case-insensitive
// returns the address of the first matched letter past any leading joker
// (ie. *[..]toto.. returns the address of toto within chaine), NULL on mismatch
HTS_INLINE const char *strjoker(const char *chaine, const char *joker,
LLint *size, int *size_flag) {
size_t nsteps = 0;
return strjoker_bounded(chaine, joker, size, size_flag, &nsteps);
}
/* Test-only oracle: the same matcher without the failure memo. */
const char *strjoker_nomemo(const char *chaine, const char *joker, LLint *size,
int *size_flag) {
@@ -451,18 +431,12 @@ static const char *strjoker_impl(const strjoker_memo *memo, const char *chaine,
// d'un strcpy sur une variable ayant un nom en lettres et copiant une chaine de chiffres
// ATTENTION!! Eviter les jokers en début, où gare au temps machine!
const char *strjokerfind(const char *chaine, const char *joker) {
size_t nsteps = 0; /* one budget for the whole scan */
const char *adr;
if (chaine != NULL && strlen(chaine) > STRJOKER_MAXLEN)
return NULL;
while(*chaine) {
if ((adr = strjoker_bounded(chaine, joker, NULL, NULL,
&nsteps))) { // ok trouvé
if ((adr = strjoker(chaine, joker, NULL, NULL))) { // ok trouvé
return adr;
}
if (nsteps > STRJOKER_MAXSTEPS) // scan budget spent: no match
return NULL;
chaine++;
}
return NULL;

View File

@@ -2575,11 +2575,6 @@ int ident_url_absolute(const char *url, lien_adrfil *adrfil) {
// recopier adrfil->adresse www..
strncatbuff(adrfil->adr, p, ((int) (q - p)));
// *( adrfil->adr+( ((int) q) - ((int) p) ) )=0; // faut arrêter la fumette!
// fil[] holds the path plus a possible leading '/' the top strlen() misses.
if (strlen(q) >= sizeof(adrfil->fil) - (q[0] != '/' ? 1 : 0))
return -1;
// recopier chemin /pub/..
if (q[0] != '/') // page par défaut (/)
strcatbuff(adrfil->fil, "/");

View File

@@ -740,40 +740,6 @@ static int st_filterdual(httrackp *opt, int argc, char **argv) {
return 0;
}
/* Length/work caps stop a hostile pattern stack-overflowing or hanging the
process (OSS-Fuzz 5060751291908096 / 5745936014573568). */
static int st_filterbounds(httrackp *opt, int argc, char **argv) {
const size_t big = 100000; /* well past the length cap */
const size_t stars = 900; /* star-heavy pattern, under the cap */
char *subj = malloct(big + 1);
char *pat = malloct(2 * stars + 2);
size_t i;
(void) opt;
(void) argc;
(void) argv;
memset(subj, 'a', big);
subj[big] = '\0';
/* '*' matches anything, but an over-length subject is refused by the cap */
assertf(strjoker(subj, "*", NULL, NULL) == NULL);
assertf(strjokerfind(subj, "*") == NULL);
/* within-cap star-heavy dead-end: the step budget keeps it bounded (a hang
would time the test out) */
for (i = 0; i < stars; i++) {
pat[2 * i] = '*';
pat[2 * i + 1] = 'a';
}
pat[2 * stars] = 'b'; /* never matches an all-'a' subject */
pat[2 * stars + 1] = '\0';
subj[stars] = '\0';
assertf(strjoker(subj, pat, NULL, NULL) == NULL);
assertf(strjokerfind(subj, pat) == NULL);
freet(pat);
freet(subj);
printf("filterbounds: OK\n");
return 0;
}
static int st_simplify(httrackp *opt, int argc, char **argv) {
(void) opt;
if (argc < 1) {
@@ -1300,30 +1266,6 @@ static int st_identurl(httrackp *opt, int argc, char **argv) {
return 0;
}
/* Regression for the one-byte fil[] overflow: a 2047-byte hostless "?"-URL used
to abort in strncat_safe_ when the missing leading '/' pushed fil to 2048. */
static int st_identabs(httrackp *opt, int argc, char **argv) {
lien_adrfil af;
const size_t len =
sizeof(af.fil) - 1; /* 2047: max URL the top guard admits */
char *url = malloct(len + 1);
(void) opt;
(void) argc;
(void) argv;
url[0] = '?';
memset(url + 1, 'a', len - 1);
url[len] = '\0';
assertf(ident_url_absolute(url, &af) == -1);
freet(url);
/* valid URLs still parse, so the guard is not over-rejecting */
assertf(ident_url_absolute("http://www.example.com/a/b/c.html?x=1", &af) ==
0);
assertf(ident_url_absolute("www.foo.com?bar=1", &af) == 0);
printf("identabs self-test OK\n");
return 0;
}
/* Extra args are key=value: adr= cdispo= statuscode= status= strip= urlhack=
no-www= no-slash= no-query= n83= type=, plus repeatable prior=adr|fil|sav
registering an already-crawled link (dedup/collision paths). */
@@ -2595,8 +2537,6 @@ static const struct selftest_entry {
"memoized vs unmemoized matcher differential", st_filtermemo},
{"filterdual", "<string1> <string2> <filter>...",
"merged two-form filter verdict (fa_strjoker_dual)", st_filterdual},
{"filterbounds", "", "matcher length/work caps reject hostile patterns",
st_filterbounds},
{"simplify", "<path>", "collapse ./ and ../ in a path", st_simplify},
{"stripquery", "", "--strip-query pattern/key stripping self-test",
st_stripquery},
@@ -2627,8 +2567,6 @@ static const struct selftest_entry {
{"resolve", "<link> <adr> <fil>", "resolve a link against an origin",
st_resolve},
{"identurl", "<url>", "split an absolute URL into (adr, fil)", st_identurl},
{"identabs", "", "ident_url_absolute one-byte fil[] overflow self-test",
st_identabs},
{"header", "<raw-header-line> ...", "response header-line parsing",
st_header},
{"headerlong", "[header-name:]",

View File

@@ -12,9 +12,10 @@ if test -n "${BROWSER}"; then
fi
# Patch for Darwin/Mac by Ross Williams
if test "$(uname -s)" == "Darwin"; then
# Darwin's 'open -W' launches the default browser and waits for it to quit.
# Keep it only as a fallback so a configured $BROWSER still wins the search.
BROWSEREXEFALLBACK="/usr/bin/open -W"
# Darwin/Mac OS X uses a system 'open' command to find
# the default browser. The -W flag causes it to wait for
# the browser to exit
BROWSEREXE="/usr/bin/open -W"
fi
BINWD=$(dirname "$0")
SRCHPATH=("$BINWD" /usr/local/bin /usr/share/bin /usr/bin /usr/lib/httrack /usr/local/lib/httrack /usr/local/share/httrack /opt/local/bin /sw/bin "${HOME}/usr/bin" "${HOME}/bin")
@@ -80,7 +81,6 @@ for i in "${SRCHBROWSEREXE[@]}"; do
done
test -n "$BROWSEREXE" && break
done
test -n "$BROWSEREXE" || BROWSEREXE="${BROWSEREXEFALLBACK}"
test -n "$BROWSEREXE" || ! log "Could not find any suitable browser" || exit 1
# "browse" command

View File

@@ -192,6 +192,3 @@ test "$(with_timeout httrack -O /dev/null -#test=filter "$pat" "$subj")" == "$su
# (guards against a single-backtrack-point matcher rewrite)
match '*[aX]X*[a]Y' 'aXaXaY'
nomatch '*[aX]X*[a]Y' 'aXbXaY'
# hostile patterns must not stack-overflow or hang: length + work caps
test "$(with_timeout httrack -O /dev/null -#test=filterbounds)" == "filterbounds: OK" || exit 1

View File

@@ -1,8 +0,0 @@
#!/bin/bash
#
set -euo pipefail
# A 2047-byte hostless "?"-URL used to overflow fil[] by one byte and abort in
# strncat_safe_; the guard now returns -1 while valid URLs still parse.
test "$(httrack -O /dev/null -#test=identabs)" == "identabs self-test OK" || exit 1

View File

@@ -53,4 +53,4 @@ bash "$top_srcdir/tests/local-crawl.sh" --rerun \
--log-found ', no files updated' \
--max-mirror-bytes 700000 \
--min-mirror-bytes 500000 \
httrack 'BASEURL/big/index.html' --retries=0 -c16 -%c100 -A100000000
httrack 'BASEURL/big/index.html' --retries=0 -c8 -%c100 -A100000000

View File

@@ -44,7 +44,6 @@ TESTS = \
01_engine-header.test \
01_engine-idna.test \
01_engine-identurl.test \
01_engine-identabs.test \
01_engine-escape-room.test \
01_engine-inplace-escape.test \
01_engine-java.test \

View File

@@ -1,104 +0,0 @@
#!/bin/bash
# Smoke-test an installed webhttrack: launch it with a stub browser and assert
# htsserver comes up and serves the web UI. Arg: the install prefix.
set -euo pipefail
prefix="${1:?usage: webhttrack-smoke.sh <install-prefix>}"
wht="$prefix/bin/webhttrack"
test -x "$wht" || {
echo "no webhttrack at $wht" >&2
exit 1
}
work="$(mktemp -d)"
# webhttrack backgrounds htsserver, which outlives it; reap any stray one (scoped
# to this prefix) so a lingering server can never hold the CI step open.
trap 'pkill -f "$prefix/bin/htsserver" 2>/dev/null || true; rm -rf "$work"' EXIT
export HOME="$work/home"
mkdir -p "$HOME/websites"
marker="$work/marker"
stubdir="$work/bin"
mkdir -p "$stubdir"
# On Darwin webhttrack hardcodes "open -W", which launches a real GUI browser and
# blocks headless. Shadow uname so it takes the generic path and picks the stub
# browser below; htsserver and webhttrack's path resolution still run for real.
cat >"$stubdir/uname" <<'EOF'
#!/bin/bash
[ "${1:-}" = "-s" ] && {
echo Linux
exit 0
}
exec /usr/bin/uname "$@"
EOF
chmod +x "$stubdir/uname"
# Stub browser: webhttrack tries its browser-name list in order and runs the
# first it finds, so shadow the first entry, "x-www-browser". It gets the server
# URL, fetches it, and records the result (htsserver only lives until webhttrack
# exits, so the check has to happen here).
# -a: the UI is served ISO-8859-1, so grep must not treat it as binary.
cat >"$stubdir/x-www-browser" <<EOF
#!/bin/bash
echo "stub browser invoked with: \$1" >&2
if body="\$(curl -fsSL --max-time 20 "\$1")" && printf '%s' "\$body" | grep -qai httrack; then
echo PASS >"$marker"
else
echo "FAIL: unexpected response from \$1" >"$marker"
fi
EOF
chmod +x "$stubdir/x-www-browser"
export PATH="$stubdir:$prefix/bin:$PATH"
echo "launching webhttrack"
"$wht" </dev/null >"$work/webhttrack.log" 2>&1 &
whpid=$!
# Hard watchdog: macOS has no timeout(1), and webhttrack can block if htsserver
# never publishes a URL, so guarantee the tree dies regardless of the poll below.
(
sleep 40
kill "$whpid" 2>/dev/null || true
pkill -f "$prefix/bin/htsserver" 2>/dev/null || true
) &
wdpid=$!
for i in $(seq 1 45); do
test -f "$marker" && {
echo "marker written after ${i}s"
break
}
kill -0 "$whpid" 2>/dev/null || {
echo "webhttrack exited on its own after ${i}s"
break
}
sleep 1
done
# Reap webhttrack and the htsserver it spawned; stop the watchdog. Confirm death
# with a bounded poll instead of a blocking wait (which could hang on macOS).
echo "tearing down"
kill "$whpid" 2>/dev/null || true
if pkill -f "$prefix/bin/htsserver" 2>/dev/null; then
echo "reaped a lingering htsserver"
else
echo "no lingering htsserver"
fi
kill "$wdpid" 2>/dev/null || true
for _ in $(seq 1 10); do
kill -0 "$whpid" 2>/dev/null || break
sleep 1
done
echo "--- webhttrack.log ---"
cat "$work/webhttrack.log" 2>/dev/null || true
echo "--- end ---"
echo "marker=[$(cat "$marker" 2>/dev/null || echo NONE)]"
if test "$(cat "$marker" 2>/dev/null || true)" = PASS; then
echo "webhttrack smoke: PASS"
else
echo "webhttrack smoke: FAIL" >&2
exit 1
fi