Chronic deficiencies:
- Automated testing system. Should accommodate multiple operating
systems (at least Unix, VMS and Windows?), so Perl-based?
- Testing.
- Performance analysis and improvement.
- Localized messages.
New stuff to be added to the main list (if not already there):
- Check if -X is implemented on platforms other than Unix and VMS
(8 Jul 2010 SMS)
- Check if UnZip ?? are appropriately displayed in file names. (12 Jul 2010 EG,
from forum thread "Info-ZIP Software Discussions and Feature Requests >
Info-ZIP UnZip > How to deal with a list of filenames with '??'")
(Also thread "Info-ZIP Bugs > UnZip Bugs > file listing in UTF-8 shows
question marks")
- Explain path matching in extended help, such as *a.txt matches a/b/a.txt.
(12 Jul 2010 EG, from forum thread "Info-ZIP Software Discussions and
Feature Requests > Info-ZIP UnZip > How to extract a single file among
using unzip?")
- Add -O option to set the code page. (12 Jul 2010 EG, from forum thread
"Info-ZIP Bugs > UnZip Bugs > Unzip 6.0 is missing option -O")
Seems to be in 6.10b.
- Look at extracting files with names larger than local OS limit (maybe
by adding a truncation option. (12 Jul 2010 EG, from forum thread
"Info-ZIP Bugs > UnZip Bugs > unzip archives with very long filenames")
- Add compression method 98 (PPMd). See the thread for references. (12 Jul
2010 EG, from forum thread "Info-ZIP Bugs > UnZip Bugs > unsupported
compression method 98")
Done in 6.10c09.
- Fix for funzip. (12 Jul 2010 EG, from forum thread "Info-ZIP Bugs > UnZip
Bugs > compilation error: Uz_Globs has no disk_full")
Fixed in 6.10c09.
- Add more specifics about editing flags. Is check needed that all data
from extra field was read? (12 Jul 2010 EG, from forum thread "Info-ZIP
Bugs > UnZip Bugs > Valgrind reports uninitialized reads in unzip")
- Handle archives that truncate large files at 32 bits. (12 Jul 2010 EG,
from forum thread "Info-ZIP Bugs > UnZip Bugs > Unable to extract
large ZIP files")
- Probably should update the non-generic targets of UNIX Makefile. (12
Jul 2010 EG, from forum thread "Info-ZIP Software Discussions and
Feature Requests > Info-ZIP UnZip > How to include large file support
in unzip 6.00")
- Recover ftp damaged file. (12 Jul 2010 EG, from forum thread "Info-ZIP
Software Discussions and Feature Requests > Info-ZIP UnZip > Recover
ascii ftp damaged zip file - I've done it!")
- Support for more Mac-specific file attributes in AppleDouble scheme.
(3 Aug 2010 SMS)
- Testing Mac-specific attribute code on (different-endian) Intel Mac
(including crossing between Power PC and Intel). (3 Aug 2010 SMS)
- Look at controlling attribute extraction more finely, maybe using a
new --restore-attributes=(protection,owner,acl) option. (3 Aug
2010 SMS, EG)
- Shared object (Unix) or shareable image (VMS) for the object library.
================================
For UnZip 6.1/who knows:
================================
o add extraction support for other compression algorithms used by new
PKZIP, WinZIP, 7-Zip versions
- LZMA, compression type 14 (most important, because of its efficiency)
- PPMd, compression type 98 (maybe, less important)
- WavPack, compression type 97 (maybe, less important)
LZMA is first-level priority for 6.1, other formats may be taken
into consideration
LZMA and PPMd done in 6.10c09.
o add support for reading AES encrypted archives
- WinZIP format (priority 1)
- PKZip format (priority 2)
top level item for 6.1
o add multi-part zipfile handling
major feature for 6.x!
could happen for 6.1
o better support for multilingual uses and different codepages;
support unicode (UTF-8 coded) filenames and comment texts
a requested feature getting more and more important,
- partially done for the Windows port in 6.0
(support restricted for chars of the current system codepage)
- partially done (beta state) for Unix
(requires native codepage to be UTF-8)
o complete support for UTF-8 coded entry names (and comments)
- add new "win32_wide" port to extend unicode support on Windows
beyond the restrictions of the current (ANSI) system codepage
- revise/extend the WinDLL interface to allow passing of "wide"
string argument data
- add simple built-in character translation between UTF-8 and the
old (ISO-8851-1 / IBM850) code pages to allow old systems without
standard UTF-8 support to read UTF-8 encoded archives.
- extend the built-in translation tables to support other language
regions besides "Western_Latin1" (e.g. Russian-kyrillic, Japanese,
Chinese)
There has been talk on the forum to link to the libiconv library
to do translations.
- streamline the multilingual codepage and UTF-8 support for the UNIX
port (standard codepage translation facility?, like WideChar<->AnsiCP
translation functions under MS Windows)
should happen for 6.1
(there is internal alpha-state code for better "wide" support on
Windows available at the time of the 6.0 release)
Planned for UnZip 6.10b.
o revise the "extended charcodes" handling in decryption password to
support UTF-8 encoding on Unicode-aware systems where the "native"
character coding is NOT UTF-8 (e.g. Windows).
o revise the command line interface for more compatibility with Zip'
command parser
- implement the versatile command parser from Zip 3.0.
- add "long option" definitions for all existing options; revise
the UnZip user manual to document the long-option alternatives.
- add support for reading the "process these entries" and the "skip
these entries" pattern lists from a file (or from separate files ?).
- add a (long) option to switch off UnZip's internal pattern matching
on filename arguments.
Initial implementation done in 6.1
o add command line options for miscellaneous features requested by users
and/or development team members:
- display the Info-ZIP software license (done as --license in 6.1)
- more fine-tuning for file attributes set/restored at extraction, like:
set/clear archive attribute on DOS/OS2/WIN32;
apply/skip standard or user-defined umask filter on UNIX (& Unix-alike)
- additional time-stamp related processing filters
- more listing display modifications
- overriding the default date-time display style
- ...
All these options are of minor importance and/or would collide with
existing "one-character" options. The current UnZip maintainer does not
want to reserve any of the few not-yet-occupied short option characters.
for one of these features. So, any implementation effort for items
of this feature wish-list has to be delayed until the "long option"
support of the revised command line parser becomes available.
------------------------------
New stuff:
- option to extract a file using a new name
(2009-6-9, Kamers on forum
(http://www.info-zip.org/board/board.pl?m-1244582716/))
- support Windows extra long paths
(2009-7-8, fritz.mueller on forum
(http://www.info-zip.org/board/board.pl?m-1246994079/))
------------------------------
some option may get implemented in 6.1
o support for
and/or development team members:
o add new low-level, binary API; rewrite "normal" (command-line) UnZip
to use it
maybe soon (maybe 6.1)
o MSDOS/WIN32/others: detection of "reserved" names (= names of character
devices, or system extensions that look like a characters device driver)
at runtime; with the goal of emitting "meaningful" error messages and/or
rename queries.
(Currently, these reserved names are catched as "non-deletable files".
On MSDOS and WIN32, when the RTL stat() function allows to identify
character devices, the "reserved" names are automatically prefixed with
an underscore.)
o redesign "file exists -- is newer/older -- overwrite/skip/rename"
logic in extract.c and the corresponding system specific mapname()
services; to prevent superfluous decryption key prompts for entry
that will be skipped, later.
o rewrite to use fread/fseek/etc. [eventually: test
write(bytes) vs. fwrite(words), especially on Crays/Alphas]
soon (probably in conjunction with multi-part handling)
o incorporate new backfill version of inflate()
wait for zlib version
o check NEXTBYTE for EOF in crypt.c, funzip.c and explode.c, too
whenever
Done in crypt.c and funzip.c in 6.10c09.
o add option to force completely non-interactive operation (no queries
for overwrite/rename, password, etc.); also allow some sort of non-
interactive password provision? (file? command-line? env. variable?)
someday?
o add testing of extra fields (if have CRC)
later
o rewrite to allow use as a filter
way, way later...
o add Unix hard-link support?
way, way later...
o add ".ini" file support as a (more elaborate) alternative to the presently
supported preconfiguring abilities via special environment variables
(UNZIP on many systems...)?
way, way later (if ever)...
o add option to search zipfile contents for a string and print the
results? ("zipgrep" option--e.g., unzip -g or unzip -S) (easy for
fixed strings, hard for wildcards/true regex's)
way, way later, if at all...probably use libregex
o add -y "display symlinks" option to zipinfo? various sorting options?
(-St date/time, -Sn name)?
who knows
o add "in-depth" option to zipinfo? (check local headers against
central, etc.)--make it a better debugging tool (or just create
zipfix)
who knows (zip -F, -FF already exist)
Some maintenance or OS specific topics for 6.x release:
* add "unix-style-path -> partitioned-dataset filename" conversion
to MVS port
* should we add support for (null) entry names (empty entry name field), to
conform to the PKWARE specification?
Done in 6.10c09.
=======================================
Requested features:
- extract or exclude on basis of UID [Armin Bub, Armin.Bub@bk.bosch.de, 970904]
=======================================
o miscellaneous little stuff: whenever
--------------------------
- change DOS -f/-u stuff to use DOS API for getting filetimes, not stat()
- add (-N?) option to lose all user input and/or switch to "(*input)()"
function, replaceable by UzpAltMain() param
- add -@ option to read from stdin (zip) or from file (PKZIP)? (go32 built-in)
- add -oo option to overwrite OS/2 and DOS system and hidden files, too
- add option to compute MD5 checksum on files and/or on entire zipfile?
- decide whether to use WinGUI "skipping" diagnostics in extract.c
- combine "y/n/A/N" query/response stuff into unified section with query
function(s) (InputFn?)
- disable ^V code in remaining mapname() routines
- change filename-matching logic so case-insensitive if case-sensitive fails?
- allow multiple dir creation with -d option? [Bob Maynard]
- use gcc -pg, gprof to do profiling on unzip
- Doug Patriarche (doug.patriarche.bvdhp01@nt.com) Northern Telecom Canada Ltd.
"I need to do a port of zip/unzip for Wind River Systems' VxWorks OS"
[GRR: 15 March 95 -> "early June"]
Features from old BUGS file (mostly duplicates of other entries above):
- ignore case for internal filename match on non-Unix systems, unless file-
specs enclosed in single quotes
- modify to decompress input stream if part of a pipe, but continue using
central directory if not (BIG job!)--extended local header capability
- add zipinfo option(s) to sort alphabetically, by date/time, in reverse, etc.
- when listing filenames, use '?' for non-printables? [Thomas Wolff, 92.6.1]
- add zipinfo "in-depth" option? (check local vs. central filenames, etc.)
- create zipcat program to concatenate zipfiles
- add -oo option (overwrite and override)? no user queries (if bad password,
skip file; if disk full, take default action; if VMS special on non-VMS,
unpack anyway; etc.)
- add -Q[Q[Q]] option (quiet mode on comments, cautions, warnings and errors)?
forget -oo, or make synonym? Default level -Q?
