Hi all,
I said this in a reply in one of the threads (the one about Windows registry), and got zero reply. I'm bringing the subject up again here.
Back in 1996 (and until around 2000) I was project manager for a project called "GTFormat". This was a project used by the late Packard Bell, as well as NEC in Japan, to put the programs bundled with the machines on the hard disks shipped out to customers. The tools consist of a tool that understands what the original installation did, a database to do offline conflict resolution and other stuff, and a front end to perform a (silent) installation of the result. I have written most of the code in the last part of this project, and as I said before, managed the entire thing. We also had a tool the generated files for the last part directly from the first part, without going through the database.
Now, the tool was written when I was an employee of the company that produced them (G.Tek Technologies, today called gteko), but I believe that given enough persuasion I can get their approval to either freely distribute or actually open source the detection and the installation tool. I believe most of the distinguishing value of the tool as far as Gteko is concerned is in the database, which is not relevant for Wine in any way.
The tool is written mostly in C++ for Visual Studio 6. It may require pulling out a single proprietary library (compression), but should not pose a problem (zlib does a wonderful job, after all).
The question, therefor, is this. Should I try? The tool has proven itself over a long period of time, and is fairly reliable (at least was back at the time). It CAN solve some of our installer related problems. Your opinions are welcome.
So, what say you?
Shachar
On Mon, 14 Mar 2005 22:09:54 +0200, Shachar Shemesh wine-devel@shemesh.biz wrote:
the hard disks shipped out to customers. The tools consist of a tool that understands what the original installation did, a database to do offline conflict resolution and other stuff, and a front end to perform a (silent) installation of the result. I have written most of the code
Putting on my sys admin hat that hasn't been worn in a long time...
Yeah, I think this would be useful. Here's an example of why:
I tried running the current Word Viewer 2003 install program it failed with MSI errors. I simply tried copying over the installation directory from a Windows partition but it didn't work. Well, at that point I could either corrupt my clean Wine install with a native MSI, or I could try to sort out what the install program does on Windows. I used a program called InstallSpy 2.0 from 2BrightSparks to figure out what the install program did. I noticed it dumped some files in c:\Program Files\Common Files\blah\blah\blah. 'Lo and behold, as soon as I copied that folder over to my fake Windows drive, Word Viewer worked just fine. All in all, I probably spent about 15 minutes on the solution and avoided native MSI.
Now, a program that monitored a Windows install, copied all of the files created, generated a .reg file with registry changes, noted INI file changes, and then built an RPM that would install on Linux.. that would be cool.
-Brian
Brian Vincent wrote:
Now, a program that monitored a Windows install, copied all of the files created, generated a .reg file with registry changes, noted INI file changes, and then built an RPM that would install on Linux.. that would be cool.
-Brian
Thinking about it, maybe it would be easier to write one than to get my ex-boss to change copyright on the existing tool + renovate it. Also, as the current tool is C++, it is bound to be an external tool anyways.
Hmm....
Shachar