Hey Bulat, I have a feature suggestion.
It's something for distant future, but I'd like to know what do you think about it.
Incremental backups. FreeArc takes as parameters an archive and a set of files and creates a new archive with new / changed files and a reference to the previous archive for the rest.
Something like DAR.
Hey! It was planned for BIT!![]()
BIT Archiver homepage: www.osmanturan.com
Strange. +RTS option works (if "--" doesn't occur before) with 23 June version, but doesn't with current(16 Jan '09). So probably Bulat compiled that new version in another way.
Last edited by IsName; 18th January 2009 at 17:39.
yes, i fixed handling unicode filenames in cmdline, but it leads to broken +RTS handling
vi6hd8, try to add -mcd- switch. i have found that delta filter sometimes crashes in arc
btw, anyone who have reported broken compression in arc - please yell again. i even have some data you have snet, but don't remember what yoy have reported. i was busy last 6 monts, now it's time to fix these
>Incremental backups...
saved to my huge to-do list![]()
updated http://www.haskell.org/bz/arc1.arc
- fixed bug in Delta filter (leading to crashes)
- -m3 - improved compression by 1% on binary files
"-mcd-" didnt help
I checked 23 June version, increased stack size using "+RTS -K16M" and file was successfully compressed and tested without any errors
edit:
new version (with fixed delta) also gives this error
Last edited by vi6hd8; 19th January 2009 at 01:56.
updated http://www.haskell.org/bz/arc1.arcIf you support some FreeArc translation, please look for "??" in your language file and translate it
- added Compressed/Total compressed to progress dialog
>I checked 23 June version, increased stack size using "+RTS -K16M" and file was successfully compressed and tested without any errors
this means that without +RTS switch, Jun23 version was also crashed??
Hello everyone,
thank you Bulat!
DoneIf you support some FreeArc translation, please look for "??" in your language file and translate it
But what does:
mean...?0253 Total compressed=Gesamt-Kompression <- gives you a clue where it is
=Всего сжатых <= THIS IS MEANT
Best regards!
just run any compression command with new freearc
on the left it shows current archive size, on the right - estimated final archive size
vi6hd8, finally i just tried myself to compress huge file and got the same problem. thank you for report and sorry that i don't tried it myself
Hello everyone,
thank you Bulat for your kind answer. Next time I will try to be even more clear (though I thought my arrows were a good attempt...)
Line 0253 has english text and two question-marks:
which were translated by me into german.0253 Total compressed=Gesamt-Kompression
So far, so good.
Right underneath there is a "crippeld" line, starting with "=" and some russian (?) text:
I believe / think / suppose this is by accident. But I don't know. So I'm asking here, because it's possibly:=Всего сжатых
-> left over from original language file (russian) and needless
or
-> english string is missing, meaning some feature of FreeArc will either be shown in russian or not shown at all.
Please enlight me / us.
Best regards!
yes, it is "total compressed" in russian. small mistake
thank you for translation, i will update it in a moment
Hello everyone,
thank you again, Bulat!
I can think of things that are worsesmall mistake
Since I had no russian in school, I had to ask people who are more competent than I.
So from now on let's concentrate on more important things!
Best regards!
updated http://www.haskell.org/bz/arc1.arc
- fixed "+RTS -Ksize" bug on some huge files
Last edited by Bulat Ziganshin; 20th January 2009 at 01:43.
http://anonym.to/?http://nanoflooder...taller-cmd.sfx
A stub that launches "setup.cmd" instead of "setup.exe" after the extraction. Might be useful for those who make unattended distributions![]()
Fear not, to have a basic overview what the thing may mean there's Google translator![]()
I am... Black_Fox... my discontinued benchmark
Hello everyone,
That's why I like this forum -> at any time someone helps one outFear not, to have a basic overview what the thing may mean there's Google translator
When will Google have a site where one can find out if something is written really intensionally?
Best regards!
i the same problem once...
see : http://encode.ru/forum/showpost.php?...&postcount=368
and solved it : http://encode.ru/forum/showpost.php?...&postcount=372
The clue was to set an extra environment variable like :
set GHCRTS=-K20000k
Thanks for the update !
Yes it also crashed...>I checked 23 June version, increased stack size using "+RTS -K16M" and file was successfully compressed and tested without any errors
this means that without +RTS switch, Jun23 version was also crashed??
I guess it has something to do with the length of the parameters given to ARC.
See my other post about it :
set GHCRTS=-K20000k
I reported this error before with the 28-10-09 version :
http://encode.ru/forum/showpost.php?...&postcount=368
http://encode.ru/forum/showpost.php?...&postcount=372
I'll test this new version and let you know.
--edit--
Results from testing this version with same parameters on the same file :
Code:G:\test\gentoo\Gentoo 64-bit.vmwarevm>arc a mexe+rep:1300mb+tempfile+rep:1300mb+tempfile+rep:1300mb+tempfile+7zx:0 -lc1600mb -ld1600mb gentoo_tar gentoo.tar FreeArc 0.50 alpha (Jan 20 2009) started: 0.00 secs Compressing 1 file of 4.519.589.888 bytes: 0.01 secs Using exe+rep:1300mb+tempfile+rep:1300mb+tempfile+rep:1300mb+tempfile+7zx:0 Memory for compression 1556mb, decompression Compressing Gentoo.tar 99.9% Compressing 1924863692 bytes with 7za a -mx0 -mmt=3 $$arcpackedfile$$.7z $$arcdatafile$$.tmp 99.9% 7-Zip (A) 4.63 Copyright (c) 1999-2008 Igor Pavlov 2008-12-31 Scanning Creating archive $$arcpackedfile$$.7z Compressing $$arcdatafile$$.tmp Everything is Ok 99.9% Solid block compression results (223.332 seconds) exe: 4.519.589.888 bytes in 10.858 seconds 99.9% rep:1300mb: 2.239.734.974 bytes in 74.818 seconds tempfile: 2.239.734.974 bytes in 3.416 seconds rep:1300mb: 1.946.846.587 bytes in 40.716 seconds tempfile: 1.946.846.587 bytes in 1.794 seconds rep:1300mb: 1.924.863.692 bytes in 38.610 seconds tempfile: 1.924.863.692 bytes in 1.357 seconds 7zx:0: 1.924.863.827 bytes in 51.762 seconds Writing directory: 391.99 secs Found 1 directory names: 392.02 secs Directory written: 392.0 Compressed 1 file, 4.519.589.888 => 1.924.863.827 bytes. Ratio 42.5% Compression time: cpu 249.26 secs, real 392.09 secs. Speed 11.527 kB/s G:\test\gentoo\Gentoo 64-bit.vmwarevm>
Seems OK to me !!
Last edited by pat357; 20th January 2009 at 20:33.
>I guess it has something to do with the length of the parameters given to ARC.
no, the problem was progress indicator internals which showed itself on huge files compressed using rep
>I reported this error before
yes, i recalled that there were some errreports, but don't recalled where exactly i seen these (unfortunately, search on forum doesn't work)
btw, if you reported any (de)compression problems with freearc - please say it again. now i'm looking for all these bugs
updated http://www.haskell.org/bz/arc1.arc>Arc.exe t b.arc -p2 -p-
- reporting when data cannot be decrypted:
FreeArc 0.50 alpha (Jan 21 2009)
Testing archive: b.arc
Testing 2 files, 121 bytes. Processed 0%
ERROR: Bad password for compile.cmd in archive b.arc
updated http://www.haskell.org/bz/arc1.arc
fixed two severe bugs:
- program was terminated when trying to detect filetype of locked file (such as pagefile.swp)
- "divide by zero" error when calculating "Total compressed" for GUI progress indicator
- also now "Compressed/Total compressed" line displayed only for compression commands
i don't remember who asked it, but now i checked this problem: it was a bug in docs, this switch is just -ep actually. btw, one can find this info in rar.txt - i maintain a high level of compatibility with rarDocumentation mentions -ep0 switch, while FreeArc doesn't accept it
FreeArc (and it's companion libraries) sources are now available via SVN:
https://freearc.svn.sourceforge.net/svnroot/freearc
you can even ask me about write access![]()
Wow! Fantastic!
Only now I discovered this forum, with the mighty author of the fantastic arc!!! :-)
First of all, thanks very much for arc, a wonderful, fast and well made program! I use it really very much.
Second of all, recently I really was astonished by a compression argument I came through.
I had to compress a lot of isos containing a huge number of audio samples. Some of them are couples of 16bit or 24bit integers, others may be floating point (but I don't think many of them).
I tried arc -m9 -mt2 as usual, then I tested the compression level with 7zip and nanozip. They ALL finished with a 15%-20% worse ratio then the old and fast WinRAR 3.8, that generally stays behind them by light years.
I even tried with StuffIt, that gave me large compression on already compressed media files (actually jpg) but didn't give me any better news.
This incredible result made me look for forums and here I came! Is it possible that compressor full of specific high-technology algorhythms are so badly defeated by a quick general-purpose software!?
A recent iso, made by a setup file containing samples, probably with a already slight compression, gave me these results (KBytes):
Original file 880027
nanozip opt1 (1.3GB) 653425
7zip LZMA Ultra 651059
arc -m9 645500 (v 0.50)
WinRar Max 582162
I cancelled WinRK and didn't run paq because I just could not wait one week... :-) Anyway, with other files, for which I didn't keep statistics, the gap was much larger then this 6%-7%.
These results are really astonishing. Is there a way to implement (if it wasn't done already by using some external compressor) a algorhythm to compress those wavs?
This file I was compressing was from a personal owned commercial program, and it was a setup file, without any extension, and containing several audio files; but there are really many free samples, available online, for example from http://www.freesound.org/ and http://www.sampleswap.org/ .
Thanks for your great work!!
Beppi.
Last edited by Beppi; 27th January 2009 at 20:40.