This topic was archived here:

https://web.archive.org/web/20230000000000fw_/https://www.zenhax.com/viewtopic.php?t=1953

shekofte, posted Fri Dec 18, 2015 12:29 pm (10488)

---

La mia cara amica

for example , i mean i had an archive that is just made of compressed blocks (zlib) that stacked continuously without gaps between them or having a clear index table , that fragmentates these blocks by pointing to their offset or their zipsize !

so i just wrote a simple bms with enough great imaginary(random) values for both zsize(compressed size of block) and size(uncompressed size of the data block) :

Code:

comtype zlib

set zsize long 160000
set size long 400000
clog head01 260 zsize size

quickbms traditional OUTPUT:

Code:

  offset   filesize   filename
--------------------------------------
  00000104 400000     head01

- 1 files found in 0 seconds
  coverage file 0     3%   160000     4970432

Press RETURN to quit

as you saw this return the real size of uncompressed data block true but the zsize is false !and it is exactly the imaginary value i determined ...

by this way , i have two question :

1) is it possible just simply print this value ?

2) it would be so magical if quickbms while extraction process can dynamically invoke this value and assign it to another variable ... in this case i can make a cycle that automatically extract the full archive (in a BLIND EXTRACTION MODE)

your idea please ?

aluigi, posted Sat Dec 19, 2015 11:38 am (10500)

---

The amount of bytes read from the input is something that only a very limited amount of compression algorithms return back, basically just zlib and few others do it.
So this is something that doesn't depend by quickbms.

Now, on the technical side that's something that can be done using an alternative and slower solution (memory monitoring) but it may be just a waste of time.
The reason is that, other than the slowness, most of the compression algorithms aren't able to handle unexpected data so if your plan is a sort of offzip for N algorithms, then it will work probably only with a few. It's enough to use comtype_scan2 on a random file to see how many crashes you get.

P.S.: "il mio caro amico"

shekofte, posted Sat Dec 19, 2015 1:25 pm (10505)

---

thanks for your technical advise !
anyway i waited for unexpected prodigious abilities in quickbms