Compare commits

..

5 Commits

Author SHA1 Message Date
Xavier Roche
f49365fee3 ci: harden the webhttrack smoke and drop the leaky watchdog
Review found two issues. The content check matched only the HTTrack brand
string, which is a literal in every page header, so a truncated/degraded
template served 200 would still pass; also require the step-2 form action.
The hard watchdog subshell was un-redirected, so its orphaned sleep kept the
CI step stdout open ~35s after every run; the poll loop plus teardown already
bound the run, so drop the watchdog (SIGKILL webhttrack if it ignores TERM).

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Signed-off-by: Xavier Roche <roche@httrack.com>
2026-07-12 13:32:41 +02:00
Xavier Roche
d0ba2c67fa ci: route the webhttrack smoke past macOS open -W
On Darwin webhttrack hardcodes the browser to "open -W", which launches a real
GUI browser and blocks headless (the macOS smoke stalled here). Shadow uname so
webhttrack takes the generic path and uses the stub browser; htsserver and
webhttrack path resolution still run for real on macOS.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Signed-off-by: Xavier Roche <roche@httrack.com>
2026-07-12 13:04:41 +02:00
Xavier Roche
8b4faa5e2f ci: bound and instrument the webhttrack smoke test
The smoke step hung on the macOS runner. Add a hard watchdog (macOS has no
timeout(1)), drop the blocking wait for a bounded liveness poll, and log each
step plus webhttrack.log so a macOS-specific stall is visible instead of
running to the job timeout.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Signed-off-by: Xavier Roche <roche@httrack.com>
2026-07-12 12:54:30 +02:00
Xavier Roche
61935f28eb ci: reap htsserver in the webhttrack smoke test
webhttrack backgrounds htsserver and leaves it running; killing only webhttrack
orphaned the server, which held the macOS CI step open until timeout. Kill
htsserver explicitly (scoped to the test prefix) in teardown and on exit, and
dump webhttrack.log so a real failure is diagnosable.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Signed-off-by: Xavier Roche <roche@httrack.com>
2026-07-12 12:41:28 +02:00
Xavier Roche
8ae30c1f2c ci: smoke-test webhttrack on macOS
Nothing exercised the WebHTTrack GUI launcher or its htsserver backend, and
webhttrack has a Darwin-only browser path (open -W). Install into a temp prefix
on a macOS runner and assert webhttrack brings up htsserver and serves the UI
(fetched via a stub browser that shadows the first name in webhttrack's list).

The UI is served ISO-8859-1, so the content check uses grep -a.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Signed-off-by: Xavier Roche <roche@httrack.com>
2026-07-12 12:14:26 +02:00
9 changed files with 23 additions and 130 deletions

View File

@@ -135,18 +135,16 @@ 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()

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

@@ -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 \