HTTrack / WinHTTrack 3.50 Beta 1 (2026)
=======================================

Build:    WinHTTrack 3.50-beta-1, commit 98a513d, engine 3.49-15 (nothing is tagged yet)
Baseline: everything below is new since 3.49-2 (20 May 2017), the release
          offered on https://www.httrack.com/page/2/en/index.html

These builds are BETA. They come out of the development branch; they are not a
release and have had no release testing. Everything executable is signed: what we
build with an Authenticode certificate issued to Open Source Developer, Xavier
Roche, and the bundled Microsoft runtime with Microsoft's own. They need
Windows 7 SP1 or later, so Windows XP and Vista users must stay on 3.49-2. The
Visual C++ runtime now travels with the program, so nothing has to be installed
first, for either the installers or the *_noinst.zip archives. On Windows 7 SP1
that runtime still needs KB2533623, which comes from Windows Update.

The complete source these builds came from is named in SOURCE.txt, next to them.

Report anything you find at https://github.com/xroche/httrack/issues (engine)
or https://github.com/xroche/httrack-windows/issues (Windows interface).


WinHTTrack (Windows interface), since 3.49-2
--------------------------------------------
+ New: a "Write WARC archive" option, with the log, index and cache options regrouped on their own tab
+ New: the interface exposes the engine options added since 3.49-2: a cookies.txt to preload, a random pause between files, query-key stripping, the individual URL-hack toggles and the CONNECT-tunnel proxy mode
+ New: the About box names the selected language, and the wizard's question dialog shows the whole link
+ Fixed: a crash when a mirror ended with the Options or View-transfers panel open
+ Fixed: use-after-free on the option-page tab titles after a language change, and in the Transfers dialog reading the engine's live state
+ Fixed: non-ASCII arguments reached the engine in the ANSI codepage instead of UTF-8, so accented paths mirrored into mojibake directories
+ Fixed: DBCS trail bytes containing 0x5C were corrupted on their way to the display
+ Fixed: a stale language index bricked startup instead of being rebuilt
+ Fixed: the directory tree missed entries in large directories and re-read them needlessly
+ Fixed: deleting a mirror descended into reparse points instead of unlinking them
+ Fixed: dial-up (RAS) gave up after a failed dial
+ Fixed: bounded the input-driven buffer writes in the URL, filter and log paths
+ Fixed: the crash reporter now unwinds properly and names the faulting call
+ Changed: rebuilt with Visual Studio 2022 and OpenSSL 3.x; the Windows floor is now Windows 7 SP1
+ Changed: the installer only ever associated .whtt, and now says so
+ Changed: the Visual C++ runtime ships beside the program rather than as Microsoft's redistributable, so the no-install ZIP needs nothing installed first
+ Changed: the x64 installer is 8 MB rather than 33 MB; debug symbols and the bundled copy of the source are no longer part of the download, and SOURCE.txt beside it says where the source is
+ Changed: retired the Internet Explorer toolbar and the dead Java/Flash parser targets


Engine (httrack), since 3.49-2
------------------------------
3.49-15
+ New: --single-file rewrites each saved page with its assets inlined as data: URIs (#713)
+ New: --changes reports what a crawl added, updated or removed against the previous mirror (#714)
+ New: --sitemap and --sitemap-url ingest sitemaps, so pages nothing links to are still found (#712)
+ New: webhttrack exposes --warc-cdx, --wacz and --warc-max-size (#862)
+ Fixed: --update and --purge-old destroyed a good local copy when the re-fetch got no response, was aborted mid-read, or when the backup meant to protect it failed (#746, #748, #758, #775)
+ Fixed: an FTP re-fetch truncated the mirrored file, resumed a complete mirror with REST and spliced the old body into the new one, and a successful transfer was blanked when its backlog slot was swapped out (#771, #797, #798, #823)
+ Fixed: a chunked response cut at a chunk boundary was stored and cached as complete (#840)
+ Fixed: cache repair deleted the old cache before a rename it never checked, and the unlink-then-rename fallback could lose the destination (#779, #786, #790, #824)
+ Fixed: several backward scans from strlen(s) - 1 read before their buffer on an empty string (#730, #768, #770, #814, #821)
+ Fixed: a URL could be saved onto the engine's own temporary files, and the final path segment was never clamped (#774, #842, #852)
+ Fixed: a query-string character reference the page charset cannot represent was left unescaped, changing how the query parses (#854)
+ Fixed: a second --update pass overwrote the previous WARC and regenerated a page-less WACZ (#759)
+ Fixed: WARC output dropped URLs of 1005 bytes or more, archived nothing for a 304 revisit, and marked an engine-forced not-modified as one the server sent (#778, #785, #826, #838, #839)
+ Fixed: ProxyTrack overflowed its .arc header block, walked past the end of an .ndx buffer, trusted an unparsed offset, and crashed on a PROPFIND or on an entry with no usable Last-Modified (#793, #820, #825, #828)
+ Fixed: webhttrack leaked the session id into crawled pages, overflowed its command line so a quoted value could inject flags, let a posted project path repoint the served root, and built its redirect Location in a 256-byte stack buffer (#700, #706, #707, #710)
+ Fixed: in the web GUI, options ticked on by default could not be un-ticked, "max site size" set a per-file cap instead of the HTML one, and the mirror link on the finished page could not be followed (#708, #709, #725)
+ Fixed: htsserver labelled every PNG as image/gif and offered JPEG as a download, and an unauthenticated GET of a directory spun the server forever (#724, #875)
+ Fixed: webhttrack hung at exit when no mirror had been launched (#753)
+ Fixed: oversized cache and header fields aborted the engine or overflowed a neighbouring field instead of being clipped (#701, #715, #717, #722, #732)
+ Fixed: a -%S rules file of 4 GB or more overran the heap (#702)
+ Fixed: the CLI display did not repaint when the terminal was resized (#97)
+ Fixed: a document with no declared charset double-encoded the title lifted for the local index (#848)
+ Fixed: a fragment on an inlined reference was dropped, losing an SVG sprite selector (#766)
+ Fixed: several time helpers handed out libc's shared gmtime/localtime static rather than a reentrant breakdown (#794, #806)
+ Changed: fatal-signal backtraces name engine frames instead of a bare module and offset (#705)
+ Changed: --without-zlib is rejected at configure time rather than failing at link (#735)
+ Changed: the offline documentation gains one GUI guide with screenshots, an Android option reference, and a restructured index
+ Changed: multiple internal hardening, test and CI improvements

3.49-14
+ New: WARC/1.1 archive output (--warc), with a sorted CDXJ index (--warc-cdx) and WACZ packaging (--wacz), also available from webhttrack (#668)
+ New: -%F takes named footer fields such as {url}, {lastmodified}, {mime}, {charset} and {status} instead of a fixed layout (#667)
+ New: a command-line guide organized by task ships with the offline documentation (#649)
+ Fixed: on Windows, paths beyond MAX_PATH truncated files, and mirroring into a long or non-ASCII directory silently failed (#133)
+ Fixed: the top index showed mojibake for non-ASCII project names and categories on Windows (#216)
+ Fixed: webhttrack handed the engine the web form's charset rather than UTF-8, so a non-ASCII path mirrored into a mojibake directory (#629)
+ Fixed: a non-ASCII single -O left the logs and the cache in a mangled twin directory on Windows (#630)
+ Fixed: a path-ceiling truncation dropped the .delayed marker, losing the file (#623)
+ Fixed: an oversized -%F footer aborted the crawl instead of being skipped (#669)
+ Fixed: a rejected 206 resume could loop and lose the file rather than refetch it whole (#581)
+ Fixed: default-port stripping was scheme-blind and dropped explicit ports from https and ftp URLs, and a :80 written with leading zeros mangled the host (#627, #638)
+ Fixed: -K silently reset the -c socket count (#650)
+ Fixed: signed-shift undefined behaviour in the zip-repair local-header read (#639)
+ Changed: the offline documentation drops stale facts, gains an Android help page, and documents the filter wildcards and the real long option forms
+ Changed: multiple internal hardening, test and CI improvements

3.49-13
+ New: SOCKS5 proxy support, with scheme-aware -P URLs (socks5://, socks5h://, connect://) and plain HTTP tunneled through a CONNECT-only proxy (#563, #564)
+ New: decode brotli and zstd content codings, advertised over TLS only as browsers do (#556)
+ New: webhttrack exposes the engine options added since 3.49-2, among them --cookies-file, --pause and --strip-query (#587)
+ Fixed: files of 2 GB or more were mishandled on Windows and on every 32-bit build (#569)
+ Fixed: --update destroyed a good local copy when the re-fetch returned an HTTP error, was aborted by -M/-E, failed to decode, or came in short (#176, #521, #557, #562)
+ Fixed: a self-redirect cookie wall was dropped instead of being re-fetched with the cookie (#15)
+ Fixed: a stalled TLS handshake ignored --timeout, and synchronous DNS resolution could wedge a crawl past --max-time (#607, #613)
+ Fixed: -M metered saved bytes rather than received volume, and overshot its cap under a slow server (#77, #520)
+ Fixed: several network-facing overflows and denial-of-service paths in the Content-Range, chunked-transfer, cookie, filter and ProxyTrack cache parsers
+ Fixed: a failed connect did not fall back to the next address on Windows (#579)
+ Fixed: -P took an out-of-range port as a garbage port, and scanned past an IPv6 literal's closing bracket (#598, #602)
+ Fixed: reject a port outside 1..65535 wherever one is parsed (a crawled URL, the htsserver and proxytrack listen arguments, an ftp:// URL), instead of letting a bare sscanf wrap a huge value into a plausible port and silently use it (#614, #626)
+ Fixed: a configured proxy still resolved and dialed the origin itself (#592)
+ Fixed: ~/ in the -O base path was never expanded (#270)
+ Fixed: a non-ASCII -O output path was double-encoded on Windows once argv became UTF-8 (#621)
+ Fixed: files under a non-ASCII project path were saved to a mangled directory on Windows (#217)
+ Fixed: --build-top-index (-%i) and --protocol (-@i) were taken for the -i continue flag, wiping the URL list and exiting on the usage screen (#615)
+ Fixed: webhttrack ignored LC_ALL/LC_MESSAGES and picked the wrong Chinese and Portuguese (#95)
+ Fixed: webhttrack wrote its base path and httrack.ini to the filesystem root when $HOME was empty (#625)
+ Fixed: crawls on a non-default port were slowed by a per-request pre-resolve (#181)
+ Changed: Windows builds moved to Visual Studio 2022 and OpenSSL 3.x, the VS2008 project files are retired, and the binaries carry a version resource
+ Changed: removed the obsolete Java-applet (.class) parser and the dead SWF module remnants
+ Changed: multiple internal hardening, test and CI improvements (Windows and macOS crawl suites, HTML-parser fuzzing, parallel make check)

3.49-12
+ New: --why explains which filter rule accepts or rejects a given URL, then exits (#505)
+ Fixed: links carrying raw UTF-8 bytes were fetched double-encoded and 404'd (#516)
+ Fixed: an uncompressed body mislabeled as gzip no longer loses the page (#515)
+ Fixed: remote stack overflow and uninitialized read in Content-Type/-Encoding parsing (#506)
+ Fixed: several over-reads and a leak in the filter, URL and IDNA parsers (#499)
+ Fixed: bound the cache-index (.ndx) parser to its buffer (#507)
+ Fixed: catastrophic backtracking on '*'-heavy filter patterns (#513)
+ Fixed: cookies.txt was created world-readable (#511)
+ Fixed: a single corrupt cache entry no longer aborts the whole mirror (#494)
+ Fixed: cache-reconcile policy was broken for zip caches (#491, #493, #495)
+ Fixed: cancelling a crawl mid type-check no longer orphans .delayed placeholders (#496)
+ Fixed: detect URLs after the first inline script and in mid-tag attributes (#497)
+ Changed: build with _FORTIFY_SOURCE and -fstack-protector-strong (#504)
+ Changed: removed the pre-3.31 (.dat/.ndx) cache import (#512)
+ Changed: multiple internal hardening and build improvements (libFuzzer harnesses, CodeQL, dead-code removal)

3.49-11
+ New: parse robots.txt Allow rules and path wildcards per RFC 9309 (#452)
+ New: advertise deflate in Accept-Encoding and decode deflate responses (#450)
+ New: follow <source> and <track> media elements as embedded links (#451)
+ New: added modern web MIME types to the type/extension table (#448)
+ Fixed: enforce the -E time limit during a slow transfer instead of only between files (#481)
+ Fixed: sniff the leading bytes of a download so a misdeclared Content-Type no longer renames a correct URL extension
+ Fixed: fast transfers could be saved under their temporary .delayed placeholder name (#5, #107)
+ Fixed: follow a redirect that maps to the same saved file instead of writing a self-pointing stub (#159)
+ Fixed: several network-facing buffer overflows in the FTP, Java and HTML parsers
+ Fixed: the htsjava plugin could not be loaded (hidden entry points, stale library name)
+ Fixed: HTML-escape truncation and a cache-buffer leak in the parser
+ Changed: modernized the default User-Agent to an honest HTTrack identifier (#449)
+ Changed: decode the full WHATWG set of HTML named character references (#443)
+ Changed: refreshed stale HTTP status, proxy-port and TLS-floor constants (#453)
+ Changed: multiple internal hardening, build, test and CI improvements

3.49-10
+ New: --cookies-file to preload a Netscape cookies.txt before crawling (#215)
+ New: --pause to space out file downloads by a random delay (#185)
+ New: --strip-query to drop selected query keys from the dedup naming (#112)
+ Changed: split the -%u URL hacks into independent --keep-www-prefix, --keep-double-slashes and --keep-query-order toggles (#271)
+ Fixed: follow a redirect Location after dropping its #fragment, instead of requesting the fragment and polluting the saved name (#204)
+ Fixed: escaped brackets inside a *[...] filter character class (#148)
+ Fixed: honor the server's Content-Range when resuming a partial download, instead of appending overlapping bytes (#198)
+ Fixed: abort the download as soon as the response type is excluded by -mime:, instead of fetching then discarding the body (#58)
+ Fixed: keep size-based filter rules neutral until the file size is known (#143)
+ Fixed: stop the mirror with a clean fatal error on a cache write failure, instead of crashing (#174, #219)
+ Fixed: stop the 412/416 partial re-get loop on --continue and --update (#206)
+ Fixed: keep an unrecognized URL tail instead of mangling it to .html (#115)
+ Fixed: honor --tolerant (-%B) on a broken Content-Length, and fix an out-of-bounds read it exposed (#32, #41)
+ Fixed: fall back to the next resolved address when a connection fails or stalls, instead of hanging on a dead IPv6 address
+ Fixed: report why a -%L URL list could not be loaded (#49)
+ Changed: multiple internal hardening, build and CI improvements

3.49-9
+ Fixed: file-type detection from the Content-Type header: trust a declared type over a binary URL extension, honor --assume under the delayed type check, and keep a known extension against a bogus or empty Content-Type (#267, #29, #56)
+ Fixed: an uninitialized-buffer read when the Content-Type is empty (#411)
+ Fixed: restored C++ source-compatibility of the installed headers so reverse dependencies (httraqt) build again (#413)
+ Changed: multiple internal build, packaging and test-harness improvements

3.49-8
+ New: tunnel HTTPS downloads through the configured HTTP proxy via CONNECT (#85)
+ New: parse every candidate URL in <img> and <source> srcset lists (#326)
+ Changed: dropped the obsolete OpenSSL linking exception (OpenSSL 3.0+ is Apache-2.0 and GPL-compatible); httrack is now plain GPLv3-or-later
+ Fixed: several out-of-bounds reads in the HTML/CSS parser on hostile input (#94, #396)
+ Fixed: stored XSS via an unescaped URL in the generated page footer (#165)
+ Fixed: hardened buffer copies throughout the engine against overflow
+ Fixed: capture conditional CSS @import URLs (#94)
+ Fixed: don't crawl xmlns namespace declarations as links (#191)
+ Fixed: don't mistake the method argument of XMLHttpRequest.open for a URL (#218)
+ Fixed: percent-encode parentheses when rewriting CSS url() targets (#163)
+ Fixed: collapse ../ in file:// URLs and widen relative-link handling (#137, #162)
+ Fixed: drop the obsolete $Version/$Path attributes from the request Cookie header, per RFC 6265 (#151)
+ Fixed: keep empty quoted arguments when reloading doit.log for --update/--continue (#106)
+ Fixed: raise the User-Agent and custom-header length limits (#152)
+ Fixed: abort on a long log path (lock-file buffer too small) (#183)
+ Fixed: race in lazy mutex initialization (#297)
+ Fixed: sub-second mtime precision when comparing local files on POSIX (#383)
+ Fixed: modernize OpenSSL TLS initialization for the 3.x to 4.x transition (#308)
+ Fixed: in-place changes made by the postprocess callback were not applied (Roman Sęk)
+ Fixed: "preffered" typo in the help text and man page (yosinn1-blip)
+ Fixed: corrections and updates of the Russian translation (German Aizek)
+ Fixed: corrections and updates of the Danish translation (scootergrisen)
+ Fixed: link libhtsjava and the libtest examples directly against libc
+ New: documented the public library API headers and typed the option fields as named enums
+ Fixed: numerous build, packaging, CI and test-coverage improvements (out-of-tree builds, sanitizer/distcheck CI, shell and Python linting, AppStream metainfo)

3.49-7
+ Fixed: keep generated config.h architecture-independent (Debian #1133728)
+ Fixed: man page rendered the -%! warning as bogus options (Debian #1061053)

3.49-6
+ Fixed: puny_decode CVE-2017-14062

3.49-5
+ Fixed: MiniZip CVE-2023-45853
+ Lintian fixes, multiple build fixes

3.49-4
+ New: Push default bandwidth to 100kiB/s, max to 10MiB/s
+ Fixed: Multiple packaging and build fixes, added security flags

3.49-3
+ Fixed: Corrections and updates of the Brazilian Portuguese translations (Paulo Neto)
+ Fixed: Fixed for Android
+ Fixed: Add mimetypes for documents (Pablo Castellano)
+ Fixed: Add firefox-developer-edition (sickcodes)
+ New: Support data-src and data-srcset for img (Brian 'Redbeard' Harrington)
+ Fixed: When using the '%N' structure naming key, separate the filename and extension with a dot (Steve Mokris)
+ Fixed: Multiple build updates