Thread: FreeArc 0.40 pre-release version

i just have uploaded first pre-release version of FreeArc 0.40:

http://www.haskell.org/bz/arc040.htm
http://www.haskell.org/bz/FreeArc-0....ease1-win32.7z
http://www.haskell.org/bz/FreeArc-0.40-prerelease1 -linux-i386.tar.bz2
http://www.haskell.org/bz/FreeArc-0....se1-sources.7z

i expect to make release in a 3 weeks, which will include full English docs and fix many minor problems still remaining. please test this version using included pretest.cmd script to help me spot any remaining bugs

Hello everyone,

is it ok for you to spread the news?
Or do you want us to keep silence?

Best regards!

Thanks Bulat!

*Fantastic*
I'll try it as soon as possible to give some feedback...

The sourceforge project page is empty, on idea could be to link the web page to the sourceforge project web page and adding news in the provided section.
Probably Bulat wants to wait for the stable release, the pre-release is only for friends (i.e. anyone want to try it...)...

Originally Posted by Vacon

is it ok for you to spread the news?

well, this version is only for testing and doesnt contain any english docs, so if you want to make everyone know about single-legged horse

you may spread info among testers (in both meanings), others should be more interested to see final release

I've tested it on ~38000 files, didn't find any error.
Is it possible to add decompressors/lister for 7z, rar, zip files?

Originally Posted by nimdamsk

Ive tested it on ~38000 files, didnt find any error.

you mean "tested with supertest.cmd"?

Originally Posted by nimdamsk

Is it possible to add decompressors/lister for 7z, rar, zip files?

possible for 7z/zip, much harder for rar. it may be interesting project but i fear not very useful. users asks for the improved compression, GUI, using fa as backup utility so i hope to work in these directions

Yes, i tested with supertest.cmd.
I really need only one program for my archive needs. I don't want FreeArc to support all possible archive formats, but i think this "big three" support is a must. Most of my archives have these extensions and using many archivers is not so comfortable and powerful as using FreeArc's internal facilities Anyway, it's yours decision when or whether to do this support.

many years ago i maintained KAMAZ archive of pirated software so i worked with various archivers every day from sunset till dawn. and i was so ingenious that i've developed small batch files unifoming calls to all these archivers - there was x command to extract any archive, t to test it and so on

and now the answers to Gish questions:

> First, I haveВ* a very basic question: is FreeArc (or will it be) open source?
> I believe Open Source program are simply the best.В* Moreover, I
> wish that FreeArc could have a choice for success, but I think that
> it really need to be released as open source clearly (i.e. with a GPL license).

before all, we should attribute FreeArc success to the use of best
today's algorithms. noone can develop best lz77, ppm, bwt and audio
algorithms simultaneously, so other archivers that use their
proprietary compression libraries, just don't get any chances to
compete with FA (except for CCM/LPAQ that uses next-generation algorithms)

next, proposing me to make program free-licensed, you forget important
point - i also need some money for life. so, making FreeArc absolutely
free will mean that i will need to find some job and stop its
development. moreover, this will mean that my own resources i put into
this development will remain uncompensated. probably you don't suspect
that i work on it fulltime, but try to compare FreeArc amount of
changes in this year with that of other archivers

so, i try to balance here - FreeArc will remain free and open-sourced,
and i don't have anything against using its code in similar projects but i
need to find way to pay myself. to collect donations may be one possible
option, to sell compression algos i have developed - another one. if
anyone here have some ideas - please suggest

when i will find some reliable source of payment, i will put FreeArc
sources under appropriate license. *i* prefer the most liberal ways,
but life may be not so simple


> My second question is: do you plan to update and maintain FreeArc,
> to align it with the new releases of the compression algorithms
> (LZMA by Igor Pavlov, PPMD by Dmitry Shkarin and GRZipLib by Ilya
> Grebnov)? I fear that the choice of using the "Haskell" language
> would make more complex to write/change/update the code (I've not
> even heard of that language before...), than the development and
> maintaining of FreeArc will depend most of you. And FreeArc will
> become popular only if it will be safely maintained...

first, using Haskell makes it much easier for *me* to develop, debug
and improve it. i tried to compare with 7-zip and it seems that
FreeArc implemented the same functionality with 3 times less
amount of code (there is a rule of thumb: developers using any
language produce the same amount of code per month, the only thing
that changes over the time (and languages) is that this code does more
and more work), so you may estimate that using Haskell makes FreeArc
development 3 times faster than rar/7zip ones.

of course, using Haskell+Russian comments makes contribution for most
of developers harder. btw, i expect that Haskell popularity will grow.
at least, my previous beloved languages, perl and then ruby, become
pretty popular now, and amount of Haskell job offers significantly
increased this year


> Then the third question is: Do you plan to release in a future a
> more portable C/C++ code for it?

there is an Extractor sub-project which is intended to make portable C++
.arc decompressor and SFX modules. it's a bit obsolete now, but i plan
to fix it in next releases

the most important thing you should know about Haskell-C++ dilemma:
using C++ slows down the development process so i should avoid it as
much as possible. just one example: the whole project now consists of
1.2 mb of compression/encryption algorithms, 500 kb of Haskell code,
and 180 kb of boilerplate code for Compression library (i.e.
providing common interface to all the compression/encryption algos)!
if the last part was written in Haskell, it will become 10 times
smaller! but this is impossible because Compression library is also
used in C++ code. and now i need to port even more code from Haskell
to C++ in order to make development of Extractor possible

FreeArc includes many smart algorithms (flexible, user-controlled
compression settings; smart solid archives update, smart file sorting
and many other) which are "profitable" only because they was
implemented using higher-level programming language. i don't expect
these algos to be implemented in other archivers - it's too much work
for their authors

so, i don't get a penalty by doing Haskell-C++ interface - it's terse.
but i waste time developing glue code in C++ instead of Haskell for
the needs of Extractor/SFX project

all the compression/encryption/crc algos are written in C++. the glue
code is required to provide common interface to them all, and to
implement some additional features (such as changing algorithm
settings depending of amount of memory available) and these
are not expenses of interfacing to Haskell, these are expenses of
making common library and will be required even if program was
entirely written in C++


> My fourth question is: WillВ* a binary/source release of FreeArc
> (0.40+) for windows be available? I guess the window release is
> somewhat different from Linux, though I do not know why. I'm one of
> those 90% of PC users who have windows and not Linux...

as i said, i spent a week to port program to Linux first time. the
sources are common and contains small chunks of OS-specific code
selected at compile time

> My fifth (and luckily last) question is: will you use new Open
> Source algorithms in FreeArc. I've found two of them very
> interesting: LPAQ and SR2. LPAQ is the fast-light implementation of
> PAQ (20-30 time faster) and it is in my opinion the best "usable"
> compression algorithm.

as i said, ppmonstrI is even better, working 1.5 times faster with
close compression ratio. i will use either it or lpaq, if Dmitry can't
make his algorithm public

> SR2, instead, is one of the fastest (really
> *fast*, 3-4 times faster than Gzip) with a good compression rate,
> better than Gzip and Bzip2. I don't have figures to compare it with
> GRZipLib , but you can have a look at it.

you may find SR2 in http://www.maximumcompression.com/data/summary_mf2 .php
at 114'th place. let's look which places will have FreeArc in -m1
and -m2 modes

> I wish you success for you and FreeArc, it can really be the best
> Open Source compression algorithm, I mean it could become one of the
> most important piece of software for the Net and its users.
> To become popular, FreeArc needs, in my opinion two things:
> - a clear Open Source license (GPL),
> В*- a development roadmap for windows or something to make clear
> that windows development has not been abandoned...).

> Also, some english comments in the codea nd english web pageВ* would help a lot...

now i'm close to the point where it will become useful. until the last
weeks there was no even English docs

> Then I'm sure FreeArc will found many other OpenSource project
> interest to make it more popular. For instance, PEAZIP would provide
> a nice GUI frontend to it (the developer of PeaZip is interested to
> provide GUI for many interesting GPL compression algorithms, FreeArc would probably be the best one).

Giorgio said that it will support FA when linux version will arrive

Hi Bulat or anyone else that can assist me. I'm busy testing FreeARC 0.40, very good and fast archiver. I just want to ask a question. I hope I'm aloud to do this on this forum. I have difficulty extracting an archive to a different path than the original. Say for instance I want to extract the arcive test.arc to path c: emp, still maintaining the directory structure of the original archive, what command line must I use. Does the path c: emp have to exist first, or will FreeARC automatically create the path. Thank you

you should read docs (FreeArc036-eng.htm or http://www.haskell.org/bz/FreeArc-eng.htm). or this translation still too hard to understand?

the option you need is -dp:

arc x test -dpc:/temp

directories should be created automatically, but i see that you have spotted the bug - highest-level directories are not created. thank you

Originally Posted by Bulat Ziganshin

when i will find some reliable source of payment, i will put FreeArc
sources under appropriate license. *i* prefer the most liberal ways,
but life may be not so simple

Bulat, Ive replied to you in a long form by e-mail.
Just to summarize here is what I think:

FreeArc is good because it rely on many optimized OpenSource algorithm: they have a free usage license...
Open Source IMHO remains the simplest way (though not the only one) to get some money. However, you will hardly get money directly, neither by donations nor by selling portion of it, but indirectly: if FreeArc will spread, youll get enough visibility to obtain job offers and the success you deserve.

Hi Bulat. Thanks for your response. It was difficult to find the section on how to extract, but after a search with google, it lead me to the section in the help file which gave an exampe on how to extract. I had read the docs and tried using this same option you gave me in your post, but it did not work. I tried it again ans this is the message I receive in the command line. I tried FreeARC 0.36 and get a similar message. Any suggestions?

c: emp>arc x test.arc -dpc:/temp
ARC 0.40 (24.10) extracting archive: test.arc
Extracting 28 files, 7.839.491 bytes. Processed 0%
ERROR: c:/temp/d:: mkdir: invalid argument (Invalid argument)
c: emp>

give me results of "arc l". btw, most commands/options are compatible with RAR so you can read its manuals

(although -dp isn't there)

Here's the list Bulat. Basically, I want to extract archives to different paths. This extraction error message I get happens with other arc files too.

C: emp>arc l test
ARC 0.40 (24.10) listing archive: test.arc
Date/time Size Filename
----------------------------------------
2003-02-07 00:35:28 1.663 d:/audiograbber/audiograbber.cnt
2003-02-08 17:00:00 1.866 d:/audiograbber/German.cnt
2000-01-16 01:01:02 36.352 d:/audiograbber/ag12full.dll
2007-09-20 17:32:44 118.784 d:/audiograbber/in_flac.dll
2007-10-21 20:02:48 503.808 d:/audiograbber/lame_enc.dll
2002-07-19 12:06:00 119.296 d:/audiograbber/libVorbis.dll
2002-01-03 22:50:10 155.648 d:/audiograbber/WMA8Connect.dll
2004-02-09 05:48:52 899.072 d:/audiograbber/audiograbber.exe
2007-09-20 17:33:04 421.888 d:/audiograbber/flac.exe
2007-09-20 17:32:54 57.344 d:/audiograbber/flac_mac.exe
2007-09-20 17:32:54 57.344 d:/audiograbber/flac_ren.exe
2007-10-21 20:02:52 552.960 d:/audiograbber/lame.exe
2007-09-20 17:32:24 266.240 d:/audiograbber/metaflac.exe
2006-11-13 02:00:02 109.568 d:/audiograbber/mppenc.exe
2007-08-06 10:44:30 425.984 d:/audiograbber/neroAacDec.exe
2007-08-06 10:44:26 1.081.344 d:/audiograbber/neroAacEnc.exe
2007-08-06 11:06:16 1.744.896 d:/audiograbber/neroAacEnc_SSE.exe
2007-08-06 10:44:52 373.248 d:/audiograbber/neroAacTag.exe
1999-03-17 01:41:02 59.824 d:/audiograbber/uninstall.exe
2007-09-08 22:30:14 208.896 d:/audiograbber/wavpack.exe
2004-02-09 03:11:48 217.513 d:/audiograbber/audiograbber.hlp
2004-02-09 03:07:56 242.577 d:/audiograbber/German.hlp
2007-10-24 09:15:55 4.746 d:/audiograbber/audiograbber.ini
2003-02-09 12:04:00 46.092 d:/audiograbber/French.lng
2003-02-08 18:56:00 44.863 d:/audiograbber/German.lng
2001-12-21 00:15:16 43.771 d:/audiograbber/Italian.lng
2001-12-20 15:11:30 42.533 d:/audiograbber/Spanish.lng
2007-10-24 08:29:18 1.371 d:/audiograbber/install.log
----------------------------------------
28 files, 7.839.491 bytes
All OK

Originally Posted by Gish

if FreeArc will spread, youll get enough visibility to obtain job offers and the success you deserve.

im 33 years old professional programmer, so i have job opportunities without writing fa

btw, its the way many great compressors are stopped their development - they make their authors famous, authors get jobs, we get nothing. compressia, sbc, winrk

about winrk - it was not debugged enough to become really useful and i sure that making its algorithm open-source will not lead to that anyone will fix it. if Taylor was fixed his algorithm, i bet that he had no less chances than RAR and Squeez

Louis

arc x test.arc -dpc:/temp -apd:/audiograbber

in general, when you made archive, you should use the following command:

arc a test -dpd:/audiograbber -r

or just run in d:/audiograbber

arc a c:/dir/test -r

otherwise, directory also saved to the archive. i will add rar-compatible -ep option in the future, but just now it is that you have


btw, that is your preffered way to get support for freearc: maillist, SoundForge forum, some other forum, LJ group, your variant?

the same question about announces

Thanks Bulat. I now can extract my arc files Thanks for your help and for writing such a good archiver.

I did not quite understood what you meant by this. Could you perhaps rephrase?

"btw, that is your preffered way to get support for freearc: maillist, SoundForge forum, some other forum, LJ group, your variant?

the same question about announces"

you just get FreeArc support by asking me here. but it is not the best place in long terms, this forum is a community of compression hackers and as number of freearc users will grow, i will need to find better place where they can ask questions. so i collect opinions about where is the best place where you can ask and i will answer

the same question about releasing new versions - if you want to automatically get such announces what is your preferred way?

Originally Posted by Bulat Ziganshin

the same question about releasing new versions - if you want to automatically get such announces what is your preferred way?

I supposeyou know my answer: if you have one place just for FreeArc (SourceForge or any other you prefer) it would be better. In this forum you can still release beta, pre-release for beta testing, etc.

Bulat, I found a little bug: the speed computation is not very correct. For instance, Ive tried using "-m5" and "-mx" options: the compression time is nearly double (also according the measure of time given by FreeArc) but the speed is only 30% smaller...
In general, the speed measure given by freearc is never what I expected, i.e.
Speed =Uncompressed FileSize / Total Compression Time

look here:

Compressed 18 files, 10.145.063 => 3.692.648 bytes. Ratio 36.3%
Compression time 0.71 secs, speed 14.268 kb/s. Total 1.65 secs

speed is computed using uncompressed size and fresh compression time. it was mainly for my own puposes, i may change it to use total time (or print both)

Bulat

I didn't know I was joining a community of compression hackers, but thanks for telling me.

I would prefer a forum similar to this one where people can ask questions regarding FreeARC and releasing of new versions. I don't think a mailing list would work, but I could be wrong. People could be a bit hesitant in giving out there e-mail address.

Originally Posted by Bulat Ziganshin

ook here:
Compressed 18 files, 10.145.063 => 3.692.648 bytes. Ratio 36.3%
Compression time 0.71 secs, speed 14.268 kb/s. Total 1.65 secs
speed is computed using uncompressed size and fresh compression time. it was mainly for my own puposes, i may change it to use total time (or print both)

OK, its clear. In my test Ive used the overall time to have comparable data.

Originally Posted by Louis

I didnt know I was joining a community of compression hackers, but thanks for telling me.

Me too...

Hello everyone,

Originally Posted by Louis

I would prefer a forum similar to this one where people can ask questions regarding FreeARC and releasing of new versions.

So beloved Bulat made one for us (dont know if he knew himself )
http://sourceforge.net/forum/?group_id=207033
Still Im missing downloadable files from sf.net or link to FA#s homepage...
Bulat?

Best regards!

Originally Posted by Bulat Ziganshin

if Taylor was fixed his algorithm

I thought it was only GUI that was buggy?

Originally Posted by Black_Fox

I thought it was only GUI that was buggy?

i dont played with GUI too much. i just made scripts to archive/extract my testsets and in 10-20% of runs winrk failed! i can upload some of them so everyone could try yourself

FreeArc 0.40 prerelease-2 version implements RAR-compatible -ep option and fixes several bugs with handling absolute pathnames

http://sourceforge.net/project/showf...roup_id=207033

ps: i also add info about each release to News tracker but can't find a way to subscribe to the News

Thanks Bulat!

Thanx Bulat. Have you any plans to add in future version BCJ2 filter (instead of current BCJ)?