Re: [Qemu-discuss] Native 32bit Windows Compile of QEMU 2.1.1 and/or QEMU 2.0.2

[Paul Gydos]

Yes - I've explored Vagrant and that is not thee solution for me - I believe that QEMU is what I want - the solution is to be available offline - they will be able to ssh into the virtual machine offline and the virtual machine will act as a server alr- In my case use I'm assuming no consistent internet access - Again - I'm looking for help to compile natively an up-to-date QEMU on Windows. Please respond if you have any current (2014 or late 2013) experience in doing this. Also it is important that the tools will be made a minimal size - that is why I'm not aiming to do this with Virtualbox or VMWare - and that is why the core I'm building it on is about 5MB, Unfortunately when I add the dependencies to the core for just git and python 2 it probably will be closer to 50MB and as far as I'm concerned that is too big.The philosophy is important which is the elimination over time of unnecessary code is more important even than the writing of the new code. So in the future this project will get into hacking the up-to-date linux kernel down by eliminating portions not needed for case use (as opposed to using an old kernel) and also discovering the ways to only include the elements of the QEMU code base that is necessary - Up-to-date (well except in the case of vi vs vim) and smaller (except in the case of using glibc instead of uclibc) and assuming off-line is the direction of this project. The exceptions have been thought through also. I have already decided on a design of QEMU as a virtual layer, 4MLinux Core (linux,busybox,and glibc) as the Core, and Python2 and Git as the Development Layer, there will also be 2 Applications added for both example and designed utility) - Right now I'm only looking for help in getting started with a Natively compiled QEMU build on Win7 32 bit which though perhaps most is difficult there are reasons why I've decided to start my exploration of QEMU from this angle.

On Mon, Sep 15, 2014 at 12:12 AM, Paul Gydos <address@hidden> wrote:

Yes - I've explored Vagrant and that is not thee solution for me - I believe that QEMU is what I want - the solution is to be available offline - they will be able to ssh into the virtual machine offline and the virtual machine will act as a server alr- In my case use I'm assuming no consistent internet access - Again - I'm looking for help to compile natively an up-to-date QEMU on Windows. Please respond if you have any current (2014 or late 2013) experience in doing this. Also it is important that the tools will be made a minimal size - that is why I'm not aiming to do this with Virtualbox or VMWare - and that is why the core I'm building it on is about 5MB, Unfortunately when I add the dependencies to the core for just git and python 2 it probably will be closer to 50MB and as far as I'm concerned that is too big.The philosophy is important which is the elimination over time of unnecessary code is more important even than the writing of the new code. So in the future this project will get into hacking the up-to-date linux kernel down by eliminating portions not needed for case use (as opposed to using an old kernel) and also discovering the ways to only include the elements of the QEMU code base that is necessary - Up-to-date (well except in the case of vi vs vim) and smaller (except in the case of using glibc instead of uclibc) and assuming off-line is the direction of this project. The exceptions have been thought through also. I have already decided on a design of QEMU as a virtual layer, 4MLinux Core (linux,busybox,and glibc) as the Core, and Python2 and Git as the Development Layer, there will also be 2 Applications added for both example and designed utility) - Right now I'm only looking for help in getting started with a Natively compiled QEMU build on Win7 32 bit which though perhaps most is difficult there are reasons why I've decided to start my exploration of QEMU from this angle.

On Sun, Sep 14, 2014 at 11:45 PM, Blair Bethwaite <address@hidden> wrote:

Paul,

It strikes me that what you are trying to do is overkill to achieve your goals of introducing *nix and OSS tool-chain. Is there any reason you wouldn't just run a server (cloud VM or similar) and give your students access via SSH (then you teach them that critical step too). And if you want them to have local access on their own machine instead then just pick a standard *nix and give them VMs - Vagrant might be a good option for that. Then you could rather spend your time focusing on producing good content.

Cheers,

On 15 September 2014 13:37, Paul Gydos <address@hidden> wrote:

I am currently trying to compile on a 32 bit Windows 7 Pro SP1 system a recent and up-to-date QEMU (2.1.1 or 2.0.2 right from the primary download page as opposed to from git)right in my CMD command prompt window preferably, but willing to do it any way a reasonable beginner can do it.

The virtual machine I need will simply be a an x86.

It has been suggested to use GTK+ instead of the currently listed dependencies for creating a MinGW environment. It has also been suggested to use a 32 bit version of MinGW-W64 instead of mingw32. After the person making suggestions ceased making suggestions and then said he didn't have actual experience doing this.

So far all documentation I have found is years and years old and leads to failure. There seems to be an endless number of possible combinations of the needed tools, which is the combination that will work?

I realize that windows is not popular among the open and free community but for my case use it is essential. My case use is to introduce new students coming from multiple popular platforms Win7/Win8.1, Ubuntu 14.04 LTS, Etc. to an exact alike simple command line linux plus busybox system with a c library. The system will also have python2 and git and the goal is to learn how to navigate a simple linux system, but mostly to learn vi, git, and python2 so that they learn how to use a simple and classic text editor, I'm going to build my OS on top of the 4MLinux Core and keep it as light as possible even though it has current linux kernel, current busybox, and a current glibc. I'm hoping to put that all inside a current qemu.

I have used QEMU Manager to run the 4MLinux Core. However I would like to become proficient in QEMU and I would like an up to date build which I have full command line access to. It seems wise to have all that I would need to compile my build if I actually want to be accessing QEMU in that manner - but hey I'm a noob to this so what do I know.

I am willing to follow reasonable directions to make my environment, including specific files, changing environmental variables like PATH, from someone who actually has experience doing it.

I will document every step I follow, and make public my findings for documentation purposes breaking it down ad nauseum for the newbie. I believe that qemu is a wonderful tool and would like to see the newbie community have more access to it. 

It does seem that up-to-date QEMU documentation that would be helpful to the beginner and doesn't exclude the Windows community couldn't hurt the project.

Any help please? I have notes on what I have done thus far but since I have no evidence that I'm going the right direction I'm putting this out as a more general topic,

--
Cheers,
~Blairo