This blog is devoted to nonconstructive criticism, bashing, and other things that, being neither answers nor questions, are evidently unwelcome anywhere else in Linux Questions.
Linux is not Windows. Watch your steps.
Posted 03-22-2009 at 04:57 PM by AGer
Subj! That is, if you want something done be ready to use the trial and error method. Or, if you think something is working, please double check. This weekend example follows.
Due to the nature of Windows, a development project often presents a choice between screwing up your PC or testing in the "user environment" prepared in a virtual machine. Normally I choose the second option. In this case I have Delphi, MySQL, and the necessary crap installed on a virtual disk and source code on a host NTFS partition. Thus I can edit and compile in both the VM and the host. The VM gets access to the source using a virtual folder.
On Windows the things are as they should be - the VM feels as fast as the host (less junk installed?) and building a project takes a bit longer, possibly due to less memory and the network access to the source code.
Now I try Linux. The setup is identical. The VM feels a bit faster (better disk caching or what?), but the project rebuild takes minutes instead of seconds. So, Linux does not work. I do not care if ntfs-3g is slow or Virtual Box is slow or VB Addons are slow. Linux does not work, period.
However, if I replace the virtual folder with direct access to the partition with the source code, everything (OS startup, Delphi startup, project rebuild) is faster in the VM than in the Windows host. Linux rocks? If not the previous experience, I could falsely conclude it does.
Naturally, I cannot test or use this setup in Windows since the Windows way to prevent concurrent access to the partition with the source code is much less convenient than umount.
Due to the nature of Windows, a development project often presents a choice between screwing up your PC or testing in the "user environment" prepared in a virtual machine. Normally I choose the second option. In this case I have Delphi, MySQL, and the necessary crap installed on a virtual disk and source code on a host NTFS partition. Thus I can edit and compile in both the VM and the host. The VM gets access to the source using a virtual folder.
On Windows the things are as they should be - the VM feels as fast as the host (less junk installed?) and building a project takes a bit longer, possibly due to less memory and the network access to the source code.
Now I try Linux. The setup is identical. The VM feels a bit faster (better disk caching or what?), but the project rebuild takes minutes instead of seconds. So, Linux does not work. I do not care if ntfs-3g is slow or Virtual Box is slow or VB Addons are slow. Linux does not work, period.
However, if I replace the virtual folder with direct access to the partition with the source code, everything (OS startup, Delphi startup, project rebuild) is faster in the VM than in the Windows host. Linux rocks? If not the previous experience, I could falsely conclude it does.
Naturally, I cannot test or use this setup in Windows since the Windows way to prevent concurrent access to the partition with the source code is much less convenient than umount.
Total Comments 0
