Maybe in your case not @shermanp, but there are people who have systems reduced by space, low power processor, arm processors (Raspberry Pi case) with flash memories or for some reason in particular that can not afford to subtract space (and the list goes on ...), those people are directed to this ...

shouldn't this be moved to patches so that it can get included in the source? No, because it shouldn't be required. I just tested a freshly cloned version of PCem on Kubuntu 18.04 $ hg clone https://bitbucket.org/pcem_emulator/pcem $ cd pcem $ autoreconf -i $ ./configure $ make Make compiled PCem w...

Patch has been updated see this post for details => viewtopic.php?f=4&t=3056&p=11503#p11503

Main user facing change is better GUI integration with PCem.

Ok, I think I'm finally done with this. New patch added to this post. As stated in my previous post, the VHD changes are mostly around serialization/deserialization of VHD metadata. Tweaked VHD creation a bit as well. The core data read/write code has not changed. Quite a few changes on the PCem sid...

While I'm far from an expert on the autotools system, I believe Makefile.in is automatically generated from Makefile.am and should not be edited manually. In other words, it should be enough to run autoreconf before configure. On the other hand, the bespoke platform specific makefiles do require alt...

SarahWalker wrote: ↑Tue 13 Nov, 2018 6:54 pm Merged at rev 1197.

Just a heads up, I noticed when preparing my VHD work against the latest revision that the platform specific Makefiles haven't been updated with the changes made in this revision. Might cause some grief for those attempting to compile!

Just a heads up that I've temporarily removed the patch while I do some more work on it. I've since switched from manually serializing/deserializing the VHD header/footer, to using using #pragma pack. This should be much less error prone compared to the previous method, and I think modern ARM archit...

The config files would need to be UTF-8... however, valid ASCII is also valid UTF-8, so the config files won't have to be rewritten. Especially if we don't set the (entirely optional and unnecessary) UTF-8 BOM.

Hi folks, I've been thinking about unicode support lately (Differencing support may be coming to MiniVHD...), and wanted to have a discussion what would be the best way to make PCem unicode aware. My thoughts on the matter are as follows: Use null terminated UTF-8 internally. UTF-8 has the advantage...

Tested as much as I could, and encountered no issues. Fiddled with mounting the pcem-made VHD in Hyper-V machines, or even directly under Windows, and viceversa (VHDs made in Hyper-V mounted under PCem), and all works flawlessly. Thanks! Thanks for testing it. Glad to hear that you haven't found an...

Have you tried creating a new config? There may be some issue with the current config file you're using that's causing a problem.

I am unable to reproduce the issue here. CD images load just fine for me.

Go to Video, Render Driver and put it from Auto to OpenGL 3.0. It makes my machines work faster. It was on D3D already. OGL just displays ~999 FPS instead, but the emualtion percentage remains the same (below 100%). At desktop, just standing there, can0t hit 100%. My own CPU is barely used, ~30%. I...

In case anyone in this thread missed it, I have updated my patch in the patches subforum to add support for dynamic VHD's

MAJOR REFACTOR I decided that projects other than PCem may find my VHD code useful. To that end, I've disentangled the VHD code from the PCem code as much as possible. I am now calling this implementation MiniVHD , or Minimalist VHD. The code is available at https://github.com/shermp/MiniVHD , relea...

EDIT 2018-07-09: Updated patch available with post #2 Hi all, I wanted to see if I could extend myself, and implement sparse, or dynamic VHD support to PCem. I am happy to announce that I appear to have succeeded. This patch replaces my earlier VHD patch, which I consider depreciated (I'll edit the...

So I decided to run ATTO under Windows 98. I don't know how representative it is etc etc, but it hopefully gives some indication of any performance differences between fixed and dynamic disks. The setup was as follows: All disk images are stored on my Crucial MX100 512GB SSD. The guest OS (Windows 9...

But OS in VirtualBox works faster if it has Fixed size. Does that mean PCem will lose some of it's performance when having Dynamically allocated hard disk? Yes, potentially. But I don't think there will be too much in it. The format stores data in (typically) 2MB blocks. When a read or write occurs...

Coming soon to a PCem near you! (Hopefully... If I don't stuff anything ELSE up...) DynamicVHD.PNG Oh, and BTW, the VHD file was created by PCem, and Windows 10 mounts it as well. Up next will be to release a patch in the next day or two, where I can hopefully get some victims—er—"kind voluntee...

Just thought I'd post a quick update. I have got dynamic VHD support semi-working. Current status is that I can attach a pre-formatted image to PCem as a secondary drive, and copy some files to/from Windows 98. I have just tried installing Win 98 to a fresh (dynamic) VHD (including formatting).... I...

I have tested the newly implemented VHD support in how many scenarios as possible without a hitch. Old VHDs created under Virtualbox and Hyper-V were mounted and operated normally, and vice-versa PCem-made VHDs mount just fine everywhere else. Great thanks to Sherman, Sarah, and everyone who was in...

(Hopefully) set the HDD type when the MFM controller is selected, and the user loads a VHD with one of the supported geometries. (Does this actually do anything elsewhere in PCem?) It's mainly a hint to the user when setting up machines. Some of the early MFM controllers only support a few hard dri...

Patch has been updated again. Changes are: I did some more refactoring. No more malloc()... Added some checks to test for possible VHD footer corruption/issues. (Hopefully) set the HDD type when the MFM controller is selected, and the user loads a VHD with one of the supported geometries. (Does this...

Large VHD image files should now work properly (cross fingers...). OP has updated patch.

Bit of a face-palm that one. I forgot to cast the file size calculations to large enough (64 bit) integers.

Great job on this patch. Issue I have is any vhd I create larger than 2048mb is not mountable under windows. I tried different sizes because the first one I tried was 4096mb and pcem worked fine. I installed Windows ME. When I tried mounting the vhd in windows it did not work. So I created smaller ...

A new patch has been posted, now with more VHD creation capabilities.

Ok, I've updated my OP with a new patch, which now allows for creation of VHD's as well as opening them. With regards to my previous posts, I have implemented the VHD algorithm for calculating CHS in the size dialog when a VHD file is selected for creation. More specifically, the algorithm is used w...

For those who are interested, I ported the algorithm for both approaches to a python script and compared them. The script takes the size in MB, and calculates the geometry using the VHD method. It then uses the CHS output from the VHD method as the size parameter of the PCem method. Differing geomet...

I've noticed some differences in the way PCem and the VHD spec calculates disk geometry. PCem almost seems too specific, or not lenient enough. For example, a 100MB image (104,857,600 bytes) PCEM: Cyl: 203 Head: 16 SPT: 63 VHD : Cyl: 1003 Head: 12 SPT: 17 However, if one passes 104,761,344 bytes (10...

Looking at https://github.com/libLAS/libLAS-1.2/blob/master/include/liblas/guid.hpp I think the variant and version should be at offsets 6 and 8 rather than 7 and 9: guid[6] &= 0x0F; guid[6] |= 0x40; /* Type 4 */ guid[8] &= 0x3F; guid[8] |= 0x80; /* Variant 1 */ #endif } After reading the R...

Thanks JohnElliott. I'll adapt that. I guess I should have mentioned that a string representation is not required. The raw bytes are saved directly to the footer. EDIT: Turns out Mingw-w64 doesn't include a definition for CoCreateGuid in its objbase.h for some reason (legacy MinGW does--go figure). ...