Beruflich Dokumente
Kultur Dokumente
-------------------------------------This file lists all changes and fixes that have been made for HTTrack.
3.47-13
+ Fixed: random closing of files/sockets, leading to "zip_zipWriteInFileInZip_fa
iled" assertion, "bogus state" messages, or random garbage in downloaded files
+ Fixed: libssl.dylib is now in the search list for libssl on OSX (Nils Breunese
)
+ Fixed: bogus charset because the meta http-equiv tag is placed too far in the
html page
+ Fixed: incorrect \\machine\dir structure build on Windows (TomZ)
+ Fixed: do not force a file to have an extension unless it has a known type (su
ch as html), or a possibly known type (if delayed checks are disabled)
+ Fixed: HTML 5 addition regarding "poster" attribute for the "video" tag (Jason
Ronallo)
+ Fixed: memory leaks in proxytrack.c (Eric Searcy)
+ Fixed: correctly set the Z flag in hts-cache/new.txt file (Peter)
+ Fixed: parallel patch, typo regarding ICONV_LIBS (Sebastian Pipping)
+ Fixed: memory leak in hashtable, that may lead to excessive memory consumption
+ Fixed: on Windows, fixed possible DLL local injection (CVE-2010-5252)
+ Fixed: UTF-8 conversion bug on Linux that may lead to buggy filenames
+ Fixed: zero-length files not being properly handled (not saved on disk, not up
dated) (lugusto)
+ Fixed: serious bug that may lead to download several times the same file, and
"Unexpected 412/416 error" errors
+ Fixed: images in CSS were sometimes not correctly detected (Martin)
+ Fixed: links within javascript events were sometimes not correctly detected (w
quatan)
+ Fixed: webhttrack caused bus error on certain systems, such as Mac OSX, due to
the stack size (Patrick Gundlach)
+ Fixed: bogus charset for requests when filenames have non-ascii characters
+ Fixed: bogus charset on disk when filenames have non-ascii characters
+ New: support for IDNA / RFC 3492 (punycode) handling
3.46-1
* New: Unicode filenames handling
* Fixed: fixed bug in handling of update/continue with erased files or renamed f
iles, leading to "Unexpected 412/416 error (Requested Range Not Satisfiable)" an
d/or "Previous cache file not found" (-1)" errors
* Fixed: escape characters >= 128 when sending GET/HEAD requests to avoid server
errors
* Fixed: do not use "delayed" extensions when the mirror is aborting
* Fixed: generate error pages when needed (Brent Palmer)
* Fixed: parsing issue with js files due to "script" tags (Vasiliy)
* Fixed: anonymous FTP without password (Luiz)
* Fixed: Makefile issues regarding parrallel build and examples (Sebastian Pippi
ng)
* Fixed: removed deprecated and annoying "Accept-Charset" header in requests (Pi
otr Engelking) (closes:#674053)
3.45-4
* New: source license is now GPLv3
* New: added a "K5" feature to handle transparent proxies (Brent Palmer)
* New: option -y to control ^Z behavior (Julian H. Stacey)
* Fixed: replace // by / when saving rather than _/ (Brent Palmer)
* Fixed: do not interpret ^C before mirror is finished, or after
* Fixed: webhttrack: do not use md5sum to produce a temporary filename, but mkte
mp (Ryan Schmidt)
* Fixed: document %k for custom structure (full query string)
3.45-3
* Fixed: spurious "Previous file not found (erased by user ?)" messages leading
to retransfer existing files in cache (Alain Desilets)
* Fixed: --max-time now stops the mirror smoothly (Alain Desilets)
3.45-2
* Fixed: number of simultaneous connections was often only one
* Fixed: "Unexpected 412/416 error" leading to have broken files on disk
3.45-1
* Fixed: interrupting an update/continue mirror session should not delete anymor
e previously downloaded content (William Roeder, Alain Desilets and many others)
* Fixed: --continue/--update bug leading to download again already cached data i
n some cases (especially redirect/error pages)
3.44-5
* Fixed: crash when using -d with non-fully-qualified hostname (Alain Desilets)
* Fixed: typo in logs (Pascal Boulerie)
3.44-4
* Fixed: random crash when interrupting the mirror (spotted by -fstack-protector
) in htscoremain.c (closes:#657878)
3.44-3
+ Fixed: Linux build (closes:#657334)
3.44-2
+ Fixed: malformed format htslib.c (Moritz Muehlenhoff)
+ Fixed: default footer print format
+ New: clever "^C" handling
+ New: added --do-not-generate-errors option
+ New: increased maximum cookie name
3.44-1
+ Fixed: Randomly corrupted files during updates due to "engine: warning: entry
cleaned up, but no trace on heap"/"Unexpected 412/416 error" errors (Petr Gajdus
ek ; closes:#614966)
3.43-12
+ Fixed: buffer overflow while repairing httrack cache if a damaged cache is fou
nd from a previous mirror (closes:#607704)
3.43-11
+ Fixed: webhttrack fixes for icecat (closes:#605140)
3.43-10
+ Fixed: capture URL not working properly when IPv6 is installed (John Bostelman
)
3.43-9
+ Fixed: application/xhtml+xml not seen as "html" (Peter Fritzsche)
+ Fixed: various linux fixes for desktop files (closes:#563691)
3.43-8
+ Fixed: URL encoding bugs with filenames containing '%' characters (sandboxie32
)
+ Fixed: MacPorts Darwin/Mac fixes to webhttrack (Ross Williams)
+ Fixed: no more character escaping when not needed (such as UTF-8 codes)
+ Fixed: possibly temporary files left on disk with bogus servers giving compres
sed content on HEAD reuests
+ Fixed: URL hack caused unexpected filename collisions (index.html vs INDEX.HTM
L)
+ Fixed: "do not erase already downloaded file" option now correctly works (it l
eaves files linked in the mirror)
+ Fixed: UCS2 encoded pages are now converted properly into UTF-8
+ New: "near" option now also catch embedded (images, css, ..) files
+ Fixed: bogus chunked multimedia link text files (such as x-ms-asf files)
+ Fixed: compilation problems on Un*x version
3.33
+ Fixed: Bogus redirects with same location in https
+ Fixed: Bogus file naming with URL hack
+ Fixed: Extremly slow redirections and empty files
+ Fixed: Bogus names with directories ending with a "."
+ New: Number of connection per second can now be.. decimal, to delay even more
+ New: Enforce stronger ISO9660 compliance
+ Shell: "URL Hack" in interface
+ Shell: "Save settings" now rebuild categories
+ Shell: "Shutdown PC after mirror" option
+ Shell: Sound at the beginning/end or the mirror (configurable through system s
ound properties)
+ Shell: Fixed drag & drop, .url import
+ Shell: Fixed "wizard" mode (crash)
+ Fixed: Crash at the end due to unterminated pending threads
+ Fixed: \ is not anymore transformed into / after the query (?) delimiter
+ New: Two new callbacks for pre/post-processing html data
+ New: link-detected2 callback (additional tag name parameter)
+ Fixed: Broken ISO9660
+ Fixed: Crash on file:// links
+ Fixed: Unescaped ampersands (&) in URLs
+ Fixed: Transfer hangs introduced in 3.33-beta-2
+ Fixed: Display bug "Waiting for scheduled time.."
+ Fixed: Bug "Waiting for scheduled time.." (NOT a display bug, actually)
+ Fixed: CaSe SenSiTiViTy bugs with mutliple links reffering to the same URL but
using different case styles
+ Fixed: Failed to build from sources (FTBFS) on amd64 archs because of cast pro
blems (Andreas Jochens)
+ Fixed: & were converted into (Leto Kauler)
+ Shell: Fixed crash with long URL lists (Libor Striz)
+ Fixed: connection/seconds limiter replugged
+ Fixed: "no files updated" display bug
+ Fixed: bogus links encoded with UTF (Lukasz Wozniak)
+ New: --assume can be used to force a specific script type (Brian Schrder)
3.32
+ Fixed: css and js files were not parsed!
+ Fixed: again broken file:// (infinite loops with local crawls)
+ Fixed: Bandwidth limiter more gentle with low transfer rate
+ Fixed: external wrappers were not called during updates/continue
+ New: additional callback examples
+ Fixed: overflow in unzip.c fixed
+ New: tests are now cached for better performances!
+ New: %r (protocol) option for user-defined structure
+ Fixed: Broken engine on 64-bit archs
3.31
+ New: Experimental categories implemented
+
+
+
+
+
+
+
+
3.20-2
+ New: HTTPS support (SSL)
+ New: ipv6 support
+ New: 'longdesc' added
+ New: new file 'new.txt' generated for transfer status reports
+ New: ISO9660 compatibility option
+ New: empty mirror/update detection improved
+ New: Update hack now recognizes "imported" files
+ New: Option to disable ipv4/ipv6
+ New: Filters now recognize patterns like -https://*
+ Fixed: The engine should be now fully reentrant
+ Fixed: Fixes for alpha and other 64-bit systems
+ Fixed: Files downloaded twice if not found in cache
+ Fixed: ftp problems with 2xx responses
+ Fixed: ftp problems with multiple lines responses
+ Fixed: ftp %20 not escaped anymore
+ Fixed: ftp RETR with quotes problems
+ Fixed: now tolerent to empty header responses
+ Fixed: hts-log closed
+ Fixed: Compressed pages during updates
+ Fixed: Crash when receiving empty compressed pages
+ Fixed: Random crashes in 'spider' mode
+ Fixed: bcopy/bzero not used anymore..
+ Fixed: various code cleanups
+ Fixed: Better UTF8 detection
+ Fixed: External links now work with https and ftp
+ Fixed: Top index.html corrupted or missing
+ Fixed: URL list crashes
+ Fixed: Random crashes with large sites due to bogus naming handler
+ Fixed: Freezes on some robots.txt files
+ Fixed: Compressed files not stored
+ Fixed: SVG fixes
+ Fixed: Raw HTML responses
+ Fixed: 406 error workaround
+ Fixed: Crashes due to binary files with bogus HTML type (not parsed anymore)
+ Fixed: External https and ftp links broken, relative https links broken
+ Fixed: Automatic resizing of filter stack
+ Fixed: Various ampersand (&) elements added
+ Fixed: https with proxy temporary workaround (direct connection)
+ Fixed: "base href" with absolute uris
+ Fixed: stack frame too large on some systems
+ Fixed: random bad requests due to bogus authentication
+ Shell: Several fixes, including registration type problems
+ Shell: "template files not found" fixed
3.16-2
+ Fixed:
+ Fixed:
+ Fixed:
+ Fixed:
+ Shell:
Zlib v1.4
Gzipped files now downloaded without problems (HTTP compression bug)
Ending spaces in URLs now handled correctly
META-HTTP bug
Type registration done only once
3.15
+ Fixed:
+ Fixed:
+ Fixed:
+ Fixed:
+ Fixed:
+ Shell:
+ Shell:
3.10
+ Fixed: Broken pipes on Linux version
+ Fixed: Commandline version bug with gzipped files
+ Fixed: Crash when reaching compressed error pages
+ Fixed: Bogus html-escaped characters in query strings
+ Fixed: Files skipped (bogus anticipating system)
+ Fixed: Crash when showing stats (div by zero)
+ Fixed: Problems with URLs/redirects containing spaces or quotes
+ Fixed: Slash added when ~ detected
+ Fixed: Ugly VT terminal
+ New: Faster and cleaner mirror interrupt
3.09
+ Fixed: Several problems with javascript parsing
+ Fixed: Elements after onXXX not parsed
+ New: Source update wrapper
+ New: Style url() and @import parsed
+ Shell: Word database and maximum number of links
+ Shell: Option changes taken in account immediately
+ Shell: Cleaner installer (registry keys)
3.08
+ New: HTTP compression is now supported
+ New: Faster response analysis
+ Fixed: External page in html if cgi
+ Fixed: Mix between CR and CR/LF for comments
+ Fixed: Top index corrupted
+ Shell: Better refresh during parsing
+ Shell: DLL error
3.07
+ Fixed: Random crashes with HTTP redirects
+ New: New rate limiter (should be sharper)
+ New: Code cleaned up, new htscore.c/httrack.c files
3.06
+ Fixed: Redirect to https/mailto now supported
+ New: Top index/top dir for Un*x version
+ New: Sources more modular (.so)
+ New: Quicktime targetX= tags
+ New: HTTP 100 partially supported
3.05
+ Fixed: Non-scannable tag parameters ("id","name",..)
+ Fixed: Java classes not found when using "." as separator
+ Fixed: Java classes not found when missing .class
3.04
+ Fixed: URLs with starting spaces
+ Fixed: bogus URLs when using "base href"
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
Fixed:
Fixed:
Fixed:
Fixed:
Fixed:
Fixed:
Fixed:
Fixed:
Fixed:
Fixed:
Fixed:
Fixed:
Fixed:
Fixed:
Fixed:
Fixed:
Fixed:
Fixed:
Fixed:
Fixed:
Fixed:
Fixed:
Shell:
Shell:
Shell:
Shell:
Fixed:
2.2
Note: 3.00 alpha major bug fixes are included in the 2.2
2.02
+ New: Cache system improved, compatible with all platforms
+ New: Update process improved (accurate date)
+ New: Remote timestamp for files
+ New: ETag (HTTP/1.1) supported
+ Shell: Portugese interface available
+ Fixed: Bug with links containing commas
+ Fixed: 'file://' bug with proxy
+ New: Engine a little bit faster
+ Shell: Some bugs fixed in the interface
2.01
+ New: ftp through proxy finally supported!
+ New: Sources cleaned up
+ New: Again some new marvelous options
+ New: Speed improved (links caught during parsing, faster "fast update")
+ New: Tool to catch "submit" URL (forms or complex javascript links)
+ Shell: German interface available
+ Shell: Dutch interface available
+ Shell: Polish interface available
+ Fixed: Level 1 bug fixed
+ Fixed: Still some parsing/structure problems
+ Fixed: Referer now sent to server
+ Fixed: Cookies did not work properly
+ Fixed: Problems with redirect pages
+ New: Better javascript parsing
+ Fixed: Problems with URL-parameters (foo.cgi?param=2&choice=1)
+ Fixed: Problems with ftp
+ New: ftp transfers are now in passive mode (firewall compliant)
Interface improved
Multiple setups
Redefine options
Continue interrupted mirror improved
1.22
+ Fixed: Parsing up/down did not work well
+ Fixed: Several files not catched, bugs fixes
+ Fixed: Problems with classes (1.21)
+ New: Transfer rate can be limited (-A option)
+ Shell: Smooth refresh
+ New: ftp basic protocol a little bit improved
1.21