[xnguard]

WindBlade1, on Jun 27 2007, 03:54 PM, said:

hmmm... Any Torrents or mirrors for 1.51?

Thanks..

Sorry, not unless someone's willing to provide me hosting space for it. I can put it up on more of the free file hosting services, but that's about it.

[urie]

xnguard, on Jun 27 2007, 08:55 PM, said:

Hanschke, on Jun 27 2007, 03:28 PM, said:

where did u get it cause it is nof official?

I compiled it myself, using the InnoSetup sources that Jared makes freely available on his site.

Yeah i done that a while back but i just updated innounp.exe and instead of extract context menus because i use winrar in context menus also i added send to Universal Extractor, save confusion quite handy was allways gonna ask nitro322 to add send to option plus if possible, explain switches to disable context e.c.t.

but i know nitro322 is busy so i just did my own.

[jim0615]

Where are the Dev builds Nitro322 mentioned before?

[nitro322]

Surfer, on Jun 25 2007, 11:08 AM, said:

Uniextractor 1.5 fails to extract this file (InstallShield)

Will check it out.

MagicAndre1981, on Jun 27 2007, 10:00 AM, said:

This is the output for WindowsXP-KB935989-x86-DEU.exe:

That means that TrID doesn't even recognize it as a Microsoft Patch. I added special signatures specifically for them since they're an odd case. There's not a whole lot I can do without scanning it, unfortunately.

Hanschke, on Jun 27 2007, 02:28 PM, said:

where did u get it cause it is nof official?

No, it's not official, but I'm aware of it. I obviously can't endorse it or anything like that, but it looks like he fixed a few of the outstanding bugs from v1.5, as well as updated some of the tools. If any of the addressed issues are causing problems for you and you can't wait until the next official version (which, honestly, is likely still a little ways off), it may be worthwhile checking this out.

You can see this thread for a bit more information about the release if interested.

[MagicAndre1981]

Hanschke, on Jun 27 2007, 09:28 PM, said:

where did u get it cause it is nof official?

Do you mean me?

If yes, I've requested the hotfix from the MS-Support

[xnguard]

nitro322, on Jun 27 2007, 11:17 PM, said:

That means that TrID doesn't even recognize it as a Microsoft Patch. I added special signatures specifically for them since they're an odd case. There's not a whole lot I can do without scanning it, unfortunately.

I'm finding that none of the recent Microsoft hotfixes -- official or unofficial -- are being extracted correctly by UniExtract. I have a perfect, public example: one of the June security hotfixes.

The bulletin is here.
The patch download (for WinXP SP2) is here.

(This applies to both Universal Extractor 1.5 and to Universal Extractor x 1.5.0.1 with updated TrID defs.)

This post has been edited by xnguard: 28 June 2007 - 11:01 AM

[nitro322]

xnguard, on Jun 28 2007, 12:00 PM, said:

I'm finding that none of the recent Microsoft hotfixes -- official or unofficial -- are being extracted correctly by UniExtract. I have a perfect, public example: one of the June security hotfixes.

I had some code in 1.4.2, before I started using TrID, that would try to detect if crap files were extracted instead of files with the proper names since I couldn't distinguish between the two from analyzing the file. The code was ugly, hackish, and slow, so I removed it in favor of properly identifying the files with TrID, but unfortunately the old method seemed to be more reliable. I'll look into this more deeply once I start working on the code again, and may add that code back if I can't get the TrID detection to work more reliably.

[xnguard]

nitro322, on Jun 29 2007, 12:21 AM, said:

I had some code in 1.4.2, before I started using TrID, that would try to detect if crap files were extracted instead of files with the proper names since I couldn't distinguish between the two from analyzing the file. The code was ugly, hackish, and slow, so I removed it in favor of properly identifying the files with TrID, but unfortunately the old method seemed to be more reliable. I'll look into this more deeply once I start working on the code again, and may add that code back if I can't get the TrID detection to work more reliably.

Here's another hackish idea: USSF.

Not many executables support the peculiar selection of switches that modern MS hotfixen do...

EDIT: Or, here's something a little less hackish. At some point after an attempt to identify the hotfix with TrID, try to list the contents with cabextract:

Quote

# cabextract.exe -l 'C:\Download\WindowsServer2003-KB923845-x86-ENU.exe'
Viewing cabinet: C:\Download\WindowsServer2003-KB923845-x86-ENU.exe
File size | Date Time | Name
-----------+---------------------+-------------
2630 | 26.05.2007 07:48:12 | _sfx_manifest_
30720 | 17.03.2007 06:14:22 | _sfx_.dll
193 | 26.05.2007 07:44:48 | _sfx_0026._p
176 | 26.05.2007 07:44:48 | _sfx_0019._p
1015 | 26.05.2007 07:44:48 | _sfx_0030._p
141 | 26.05.2007 07:44:48 | _sfx_0022._p
360 | 26.05.2007 07:44:48 | _sfx_0005._p
144 | 26.05.2007 07:44:48 | _sfx_0014._p
217 | 26.05.2007 07:44:48 | _sfx_0027._p
139 | 26.05.2007 07:44:48 | _sfx_0020._p
135 | 26.05.2007 07:44:48 | _sfx_0031._p
821 | 26.05.2007 07:44:48 | _sfx_0021._p
335 | 26.05.2007 07:44:48 | _sfx_0004._p
5000 | 26.05.2007 07:44:48 | _sfx_0013._p
9524 | 26.05.2007 07:44:48 | _sfx_0025._p
1315 | 26.05.2007 07:44:48 | _sfx_0017._p
174 | 26.05.2007 07:44:48 | _sfx_0032._p
176 | 26.05.2007 07:44:48 | _sfx_0023._p
237 | 26.05.2007 07:44:48 | _sfx_0003._p
183 | 26.05.2007 07:44:48 | _sfx_0015._p
191 | 26.05.2007 07:44:48 | _sfx_0028._p
141 | 26.05.2007 07:44:48 | _sfx_0018._p
139 | 26.05.2007 07:44:48 | _sfx_0033._p
139 | 26.05.2007 07:44:48 | _sfx_0024._p
148589 | 26.05.2007 07:44:48 | _sfx_0002._p
144 | 26.05.2007 07:44:48 | _sfx_0016._p
1668 | 22.03.2007 00:31:50 | _sfx_0012._p
41248 | 22.03.2007 00:31:50 | _sfx_0001._p
325 | 26.05.2007 07:44:48 | _sfx_0037._p
449 | 22.03.2007 00:31:50 | _sfx_0035._p
6308 | 26.05.2007 07:44:48 | _sfx_0034._p
7182 | 22.03.2007 00:31:50 | _sfx_0011._p
249639 | 22.03.2007 00:31:50 | _sfx_0000._p
1015 | 26.05.2007 07:44:48 | _sfx_0029._p
267 | 30.04.2007 15:21:50 | _sfx_0036._p
199 | 26.05.2007 07:44:48 | _sfx_0009._p
5468 | 26.05.2007 07:44:48 | _sfx_0006._p
179 | 26.05.2007 07:44:48 | _sfx_0008._p
232 | 26.05.2007 07:44:48 | _sfx_0007._p
118273 | 22.03.2007 00:31:50 | _sfx_0010._p

That's pretty definitive. But, if it's not definitive enough, here're the contents at the top of "_sfx_manifest_":

Quote

[Options]

Run = "update\update.exe"
PatchDLL = "_sfx_.dll"

You'd think that cabarc would be able to extract a cabinet from an SFX, but noooo... Anyhow, cabextract appears to be [L]GPLed. It's over here, and it's available as a DOS or Cygwin binary. Or perhaps a way to use the author's libmspack directly from AutoIt?

This post has been edited by xnguard: 29 June 2007 - 12:55 AM

[nitro322]

sherpya, on Apr 26 2007, 10:02 PM, said:

you can add also http://oss.netfarm.i...ad/unshield.exe
it's able to decompress most installshield cabs

sherpya, do you know the origin of that file? I know unshield comes from the SynCE project (I've looked at it many times), but I've never found a native version of it before that doesn't require cygwin. This is a really good find, and I'd like to make sure I credit both the original author(s) at the SynCE product as well as the porter.

And yeah, I know it's been a while since you posted that. Hopefully you're still following this thread. :-)

[nitro322]

xnguard, on Jun 29 2007, 01:31 AM, said:

EDIT: Or, here's something a little less hackish. At some point after an attempt to identify the hotfix with TrID, try to list the contents with cabextract:

That's actually very similar to what I did in 1.4.2, though I use 7-Zip instead of cabextract.

[nitro322]

erthy976, on Apr 30 2007, 06:35 AM, said:

Also, here is a packer not yet supported:

LiberalInstaller: http://www.liberalsoft.com/

I just took a look, and unfortunately I'm too lingually challenged to attempt this one. If anyone can provide a utility or directions to extract the files then I'll be happy to add it, but given that I can't read anything about it I can't do a whole lot myself.

[nitro322]

anoymous_person, on Apr 26 2007, 11:00 PM, said:

The description for the SystemBoosterXP installer says its a self-extracting installation program though Universal Extractor has trouble with it.

Sorry, I can't download this. It wants me to register for a trial version, and I'm really not interested in providing personal information to that site.

[anonymous_user]

Well heres the direct link for trial. Also have you ever tried providing fake information?

This post has been edited by anoymous_person: 30 June 2007 - 10:04 PM

[nitro322]

nitro322, on Jun 28 2007, 11:21 PM, said:

xnguard, on Jun 28 2007, 12:00 PM, said:

I'm finding that none of the recent Microsoft hotfixes -- official or unofficial -- are being extracted correctly by UniExtract. I have a perfect, public example: one of the June security hotfixes.

I had some code in 1.4.2, before I started using TrID, that would try to detect if crap files were extracted instead of files with the proper names since I couldn't distinguish between the two from analyzing the file. The code was ugly, hackish, and slow, so I removed it in favor of properly identifying the files with TrID, but unfortunately the old method seemed to be more reliable. I'll look into this more deeply once I start working on the code again, and may add that code back if I can't get the TrID detection to work more reliably.

Ok, I think we should be good on this for the next version. I made some updates to the TrID defs file file a while back (forgot the specifics), and it seems to be much more reliable than the old version. I just tried extracting all hotfixes that I had archives, including several newish ones such as linked to in post 606, and TrID detection is working fine.

Just in case this becomes an issue again in the future, though, I also added back the manual checking/validation that I performed in 1.4.2. It's ugly code and I was really hoping to deprecate it, but I guess it's best left there as a failsafe than removing it altogether. Should TrID detection fail on these hotfixes, the manual check should now catch it and instruct UniExtract to do the right thing.

[nitro322]

roytam1, on May 30 2007, 09:10 PM, said:

will UniExtract support .lzma files in future?
there's some softwares packed directly with lzma program in lzma sdk.

LZMA support will be included in the next version.

Thanks for the suggestion and links.

[Camarade_Tux]

nitro322, on Jul 1 2007, 04:13 AM, said:

sherpya, on Apr 26 2007, 10:02 PM, said:

you can add also http://oss.netfarm.i...ad/unshield.exe
it's able to decompress most installshield cabs

sherpya, do you know the origin of that file? I know unshield comes from the SynCE project (I've looked at it many times), but I've never found a native version of it before that doesn't require cygwin. This is a really good find, and I'd like to make sure I credit both the original author(s) at the SynCE product as well as the porter.

And yeah, I know it's been a while since you posted that. Hopefully you're still following this thread. :-)

Maybe he's not but I think I can answer: there are two main ways to run linux/unix apps under windows. The first one is cygwin which aims at providing full support for unix apps, no matter the speed. The other one is mingw32 which iirc is a fork from cygwin which focuses on performance first.
Both produce native apps but mingw-compiled apps do not need any dll file.
I have a working mingw environnement here; just give me a sourcecode and I can compile it for you. =)
edit: I downloaded the source; there's a compilation problem but it seems minimal. I'll see if I can optimize it for size. And before I forget: which file could I test it on?

This post has been edited by Camarade_Tux: 01 July 2007 - 02:20 AM

[nitro322]

anoymous_person, on Jun 30 2007, 11:03 PM, said:

Well heres the direct link for trial.

I can't figure out what kind of installer it is. It's some kind of SFX-based installer (run it with ' /?' appended to see supported arguments), but I can't find any way to either rip out the files or instruct the installer to only extract the files without installing.

anoymous_person, on Jun 30 2007, 11:03 PM, said:

Also have you ever tried providing fake information?

Certainly. I just generally prefer to not register for stuff unless I absolutely need to, whether it's using fake or real data. Feel free to call me paranoid.

[nitro322]

Camarade_Tux, on Jul 1 2007, 03:15 AM, said:

I have a working mingw environnement here; just give me a sourcecode and I can compile it for you. =)
edit: I downloaded the source; there's a compilation problem but it seems minimal. I'll see if I can optimize it for size. And before I forget: which file could I test it on?

The copy of unshield.exe that was linked to seems to already be a native win32 app (in this case, I'm using native to mean non-dependent on Cygwin). I wouldn't spend much time tryng to compile this yourself - unless I'm mistaken, it'd just be a duplication of work. What I was more interested in is tracking down where this came from, because I spent at least a couple hours, probably longer, looking for this at several points since I began working on UniExtract. And now, seemingly out of the blue, but turns up in a locked down directory with no other information about it. That's just a bit frustrating.

As for testing, I haven't gotten that far yet. Mostly been trying to get myself back on track tonight, remember where I left off on a bunch off stuff. I'll certainly let you know if I come across anything that could use some additional testing.

[nitro322]

Molecule, on Jun 9 2007, 08:57 PM, said:

I tried this Kingston 98se USB-memory stick driver, to no avail (it produced 0-byte inf files) and so I tried to extract it using UE1.5. It has an older (2000?) InstallShield installer.

I'm not sure what was causing your problem, but it both files extract perfectly for me with the current dev version. I have not tested v1.5, though, because I've already made many significant changes and it'd be a waste of time to troubleshoot the older version at this point. Hold out a little bit for the next version and it should work fine for you.

[Camarade_Tux]

I see.

For the compilation thing, I just thought it would be a bit cleaner and whether gcc size optimizations were efficient.
Moreover, I wanted to see how much work had to be done to port a basic app to mingw. Quick but it seems gcc compiled ASCII C while the code contains non-ASCII things so it could have been quicker.
Btw, my binary is 3354o, no matter the level of optimization I use (-O2 [level2], -Os [size], or even -O3 [level 3] which is said to be space hungry).

Now, let's go back to my package manager for windows.