Files
httrack/tests/01_zlib-cache-writefail.test
Xavier Roche 484fc47eab One bad cache entry ends the whole mirror (#494)
* Reword the "bogus state" cache-skip warnings

Users read "file not stored in cache due to bogus state" as a crash or
cache corruption; it is a benign skip when the transfer is shorter than
the Content-Length. Say what happened, the consequence, and the -%B
override instead, and give the delayed-type variant a plain wording.
Test pins updated; 22_local-broken-size gains set -e (its first crawl's
audit failure was masked by the second crawl's exit status).

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
Signed-off-by: Xavier Roche <roche@httrack.com>

* Make one bad cache-entry write skippable instead of ending the mirror

Since #426 any new.zip write failure cleanly aborts the whole mirror.
Keep that for storage-level trouble (fatal errno such as ENOSPC, or
every entry failing), but let an isolated failure drop only the current
entry: abandon it, warn with the URL, and keep the mirror and the cache
stream going. A streak of CACHE_MAX_WRITE_FAILURES consecutive failures
still aborts. Also degrade the >2GB assertf crash in cache_add: an
oversized on-disk body is stored headers-only (X-In-Cache: 0), an
in-memory one drops the entry.

The cache-writefail self-test now pins all four regimes (fatal errno,
persistent streak, isolated skip with sibling round-trip, oversize); on
the previous code it fails the new assertions and hits the oversize
assertf (SIGABRT).

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
Signed-off-by: Xavier Roche <roche@httrack.com>

* review: pin streak semantics, assert the skip warning, add EDQUOT

Adversarial test audit found the policy's core semantics unpinned: no
test distinguished a consecutive-failure cap from a total count, or
proved a stored entry resets the streak. Phase 1 now asserts the abort
lands exactly on the 8th consecutive failure, and a new phase drives 10
failures interleaved with successes and asserts no abort. The .test now
greps a URL-bearing skip warning. check_fatal_io_errno gains EDQUOT
(quota exhaustion is disk-full for our purposes).

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
Signed-off-by: Xavier Roche <roche@httrack.com>

* review: trim comments to the one-line default

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
Signed-off-by: Xavier Roche <roche@httrack.com>

---------

Signed-off-by: Xavier Roche <roche@httrack.com>
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
2026-07-05 15:15:07 +02:00

30 lines
976 B
Bash

#!/bin/bash
#
# Keep this POSIX-portable: the harness runs it via $(BASH), which is a plain
# POSIX /bin/sh on some platforms (e.g. macOS), so avoid bashisms and GNU-only
# tool flags despite the #!/bin/bash above.
# Cache write-failure policy (-#test=cache-writefail <dir>). #174/#219: disk
# full or a failure streak aborts cleanly; an isolated failure or an oversized
# entry is only dropped.
set -eu
dir=$(mktemp -d)
trap 'rm -rf "$dir"' EXIT
out=$(httrack -#test=cache-writefail "$dir")
# Match the exact success line (error logs also go to stdout); a renamed/removed
# test prints the registry to stderr, which exits non-zero but never prints this.
printf '%s\n' "$out" | grep -qx "cache-writefail: OK" || {
echo "expected 'cache-writefail: OK', got: $out" >&2
exit 1
}
# A skipped entry must be warned about with its URL.
printf '%s\n' "$out" | grep -q "entry not cached: example.com/" || {
echo "expected a URL-bearing skip warning" >&2
exit 1
}