Back to home page

DOS ain't dead

Forum index page

Log in | Register

Search:  
Back to index page
Thread view  Board view
boeckmann

Aachen, Germany,
11.11.2022, 15:11
(edited by boeckmann, 11.11.2022, 17:50)
 

Ranish Partition Manager revived (Announce)

Post reply

Hi,

does anyone remember Ranish Partition Manager, a software from the end of the 90s?

The source up to version 2.37.11 is under public domain. I started maintaining the code and ported the C code from Borland Turbo C to Open Watcom C. I also ported the assebmler code, with the help of user ecm, to NASM.

At the moment two binary releases can be obtained from

https://codeberg.org/boeckmann/ranish/releases

Version 2.90.00 is the first development release of the port. Version 2.37.99 is basically a recompiled (Borland C++ 3.1) version of the last publically available public domain version from the original author.

Be aware that at the moment the software is lacking support of hard disks >8GB. But you may use the software with larger disks to create partitions within that limit. My goal is to eventually remove that restriction by implementing LBA support.

What this software sets apart from other software is its advanced boot manager capability. Especially in a multi OS environment you can effectively shield the different OS partitions from each other while letting the OSes share some common data partitions. So if you want to setup DOS / Windows / OS2 etc. on the same host its worth giving it a try. Be aware that it is not the simplest software to use and there is some learning curve.

I strongly recommend to not use the software on any production system but would be happy to get some feedback. When using this software use its save function to make a backup of your current partition layout before doing anything else with it!!! ;-)
glennmcc

Homepage E-mail

North Jackson, Ohio (USA),
11.11.2022, 18:40

@ boeckmann
 

Ranish Partition Manager revived

Post reply

> Hi,
>
> does anyone remember Ranish Partition Manager, a software from the end of
> the 90s?
>
> The source up to version 2.37.11 is under public domain. I started
> maintaining the code and ported the C code from Borland Turbo C to Open
> Watcom C. I also ported the assebmler code, with the help of user ecm, to
> NASM.
>

Small point.

README.TXT in the ZIP is the original from 1998 and still states.

"IMPORTANT: First of all run INSTALL.BAT"

However, INSTALL.BAT is not included in the ZIP.

>
> Be aware that at the moment the software is lacking support of hard disks
> >8GB. But you may use the software with larger disks to create partitions
> within that limit. My goal is to eventually remove that restriction by
> implementing LBA support.
>

I look forward to testing that feature. :)

---
--
http://glennmcc.org/
boeckmann

Aachen, Germany,
11.11.2022, 21:53

@ glennmcc
 

Ranish Partition Manager revived

Post reply

> Small point.
>
> README.TXT in the ZIP is the original from 1998 and still states.
>
> "IMPORTANT: First of all run INSTALL.BAT"
>
> However, INSTALL.BAT is not included in the ZIP.
>
Thanks for the feedback :-) The original README.TXT is gone in version 2.90.01.

> I look forward to testing that feature. :)
 Will take me some time.
bretjohn

Homepage E-mail

Rio Rancho, NM,
11.11.2022, 19:42

@ boeckmann
 

Ranish Partition Manager revived

Post reply

> does anyone remember Ranish Partition Manager, a software from the end of
> the 90s?
I certainly do. I still have a copy of it and use it on occasion when I need to mess with partitions. I still think it's one of the best DOS partition editors around -- much easier to use and understand than something like FDISK. I'm glad to see somebody move the ball forward with it -- good luck.

> What this software sets apart from other software is its advanced boot
> manager capability.
I experimented with the Ranish boot manager capability way back when, but ultimately ended up going a different route (I ending up using BootIt, which I still use). I remember the Ranish web site also had a download of XOSL (Extended Operating System Loader), which was a more "complete" version of what Ranish itself provided as a boot manager.

The other thing you might want to think about adding is support for GPT.
boeckmann

Aachen, Germany,
11.11.2022, 22:04

@ bretjohn
 

Ranish Partition Manager revived

Post reply

> > does anyone remember Ranish Partition Manager, a software from the end
> of
> > the 90s?
>
> I certainly do. I still have a copy of it and use it on occasion when I
> need to mess with partitions. I still think it's one of the best DOS
> partition editors around -- much easier to use and understand than
> something like FDISK. I'm glad to see somebody move the ball forward with
> it -- good luck.
 Thanks :-)

> I experimented with the Ranish boot manager capability way back when, but
> ultimately ended up going a different route (I ending up using BootIt,
> which I still use).
 I was not aware BootIt until now. Surely will have a look at it. But if I understand correctly it is a commercial software? https://www.terabyteunlimited.com/bootit-bare-metal/

> The other thing you might want to think about adding is support for GPT.
 I am afraid that would result in a nearly complete rewrite. For now I will concentrate on making it usable for disks of up to 2TB. I think that would make it a reasonably usable partition / boot manager for "legacy" operating systems.

I doubt if it would make much sense competing with the GPT partition managers already out when the main target operating systems even don't understand that type of partition table. But, time will tell ;-)
bretjohn

Homepage E-mail

Rio Rancho, NM,
14.11.2022, 15:12

@ boeckmann
 

Ranish Partition Manager revived

Post reply

> I was not aware BootIt until now. Surely will have a look at it. But if I
> understand correctly it is a commercial software?
> https://www.terabyteunlimited.com/bootit-bare-metal/
It is commercial software (not open-source) and works under a very different methodology than a partition manager. While it does work with multiple partitions, the thing that makes it useful for me is that it allows booting from multiple operating systems (including multiple versions of DOS) from the SAME partition. In my test environment, I have about 20 different versions of DOS (different manufacturers including MS-DOS, PC-DOS, DR-DOS, FreeDOS, and PTS-DOS) in a single partition and can select which one to boot. I can also boot into different versions of Windows and Linux with it. It moves around the files needed for booting. For my purposes that's a much better methodology than dozens of partitions.

Again, a very different methodology than a partition manager, but I think both are needed.

>> The other thing you might want to think about adding is support for GPT.
> I am afraid that would result in a nearly complete rewrite. For now I will
> concentrate on making it usable for disks of up to 2TB. I think that would
> make it a reasonably usable partition / boot manager for "legacy" operating
> systems.
Agreed, for now that's what you should concentrate on. Eventually DOS (at least FreeDOS) is going to need to learn how to "get along with" GPT, though. Newer OS's (like Windows 11) require GPT and won't even boot from MBR any more (they claim MBR isn't "secure"). Unless everyone is going to be 100% satisfied running DOS in Virtual Machines instead of bare metal (I won't), we'll need to do something different.
Zyzzle

17.11.2022, 03:20

@ bretjohn
 

Ranish Partition Manager revived

Post reply

> Agreed, for now that's what you should concentrate on. Eventually DOS (at
> least FreeDOS) is going to need to learn how to "get along with" GPT,
> though. Newer OS's (like Windows 11) require GPT and won't even boot from
> MBR any more (they claim MBR isn't "secure"). Unless everyone is going to
> be 100% satisfied running DOS in Virtual Machines instead of bare metal (I
> won't), we'll need to do something different.
100% agree. However, even if DOS is patched to support GPT, the raw partition limit will still be 2048 GiB, correct? No way around that limit, even with 4k sector-support, is there? And, further, the filesize limit shall always remain at 4096 MiB (or even 2048 MiB)?
boeckmann

Aachen, Germany,
17.11.2022, 11:42

@ Zyzzle
 

Ranish Partition Manager revived

Post reply

> 100% agree. However, even if DOS is patched to support GPT, the raw
> partition limit will still be 2048 GiB, correct? No way around that limit,
> even with 4k sector-support, is there? And, further, the filesize limit
> shall always remain at 4096 MiB (or even 2048 MiB)?
Well I do not have insight into the FreeDOS Kernel, but if it could be patched on a source level to support GPT partitions natively chances are that the 2048 GiB limit would also fall, because that 2TiB limit is dictated by constraint of the MBR partition scheme. But I think there will be other limiting factors I am not aware of like 32-bit data types for storing LBA sector numbers etc in the kernel that had to be extended >32-bit, probably introducing compatibility problems and other stuff...

Older DOS will propably never support GPT natively but user Japheth is doing some interesting research on backward compatible workarounds according to this thread:

http://www.bttr-software.de/forum/board_entry.php?id=19502

This requires replacing the standard DOS boot loader with a loader capable of understanding GPT and doing some INT 13h trickery. Here the usual constraints (max. 2TB etc.) apply.

In my opinion the greater risk of DOS someday not beeing usable anymore on modern systems is not the large disk support but the removal of legacy BIOS API support from modern UEFI-enabled boards. But even in that field there is work going on to reimplement that. The guys from ArcaOS (OS/2) seem to have had some success with that:

https://www.arcanoae.com/uefi-support-in-arcaos-5-1-update-for-february-2022/
tom

Homepage

Germany (West),
17.11.2022, 16:22
(edited by tom, 17.11.2022, 18:03)

@ boeckmann
 

Ranish Partition Manager revived

Post reply

> > 100% agree. However, even if DOS is patched to support GPT, the raw
> > partition limit will still be 2048 GiB, correct? No way around that
> limit,
> > even with 4k sector-support, is there? And, further, the filesize limit
> > shall always remain at 4096 MiB (or even 2048 MiB)?
>
> Well I do not have insight into the FreeDOS Kernel, but if it could be
> patched on a source level to support GPT partitions natively
which has already been done ;)

now limits are:

a MBR disk can have at most 2^32 sectors; with 512 byte sectors this gives you 2TB. with 4k native disks (which FreeDOS does not support) you could have 8TB disks, which is probably the trick that some external disk USB enclosers play when they present 512e disks as 4k sectors to the outside world.

a FAT32 partition on a GPT disk can also not be larger than 2^32 sectors, but you can have several of them, even beyond sector 2^32. the amount of work required in the FreeDOS kernel is limited; basically all work is done on the partition level, and would only require the addition of the disk offset in the last few instructions before calling INT13.

of course, disk caching software that doesn't expect 48-bit offsets would be a serious problem ;)

a single file can't be larger than 2^32=4GB byte on FreeDOS/MSDOS file systems

edited to add: not to "can also not be larger than" and "on FreeDOS/MSDOS file systems"
Zyzzle

18.11.2022, 01:12

@ boeckmann
 

Ranish Partition Manager revived

Post reply

> In my opinion the greater risk of DOS someday not beeing usable anymore on
> modern systems is not the large disk support but the removal of legacy BIOS
> API support from modern UEFI-enabled boards. But even in that field there
> is work going on to reimplement that. The guys from ArcaOS (OS/2) seem to
> have had some success with that:
>
> https://www.arcanoae.com/uefi-support-in-arcaos-5-1-update-for-february-2022/
Thanks for the link. I had not heard about it, Sounds very promising. Quoting from the description:
"Essentially, this is a 64-bit environment which will provide a rather complete BIOS emulation for the ArcaOS kernel, including VGA services for video display."

Impressive, since it provides BIOS emulation *and* VGA compatibility, in a sense, creating a de-novo "16-bit legacy" system inside a UEFI bootshell.

SEABIOS has also attempted this. The main problem is getting 100% VGA-compatible code into the abstraction layer of the "new" 16-bit VBIOS, "replacing" the VBIOS and making it compatibile with 30+ years of 16-bit DOS code, etc.

This is a difficult project, but I have no doubt that those of us who want to run DOS baremetal in the future will work together to circumvent this artificial castration that Microsoft, Intel, and others have worked so hard to foist upon the world for "security reasons". Talk about a damn walled garden!
boeckmann

Aachen, Germany,
19.11.2022, 21:03

@ Zyzzle
 

Ranish Partition Manager revived

Post reply

One question of a more technical kind.

I managed to modify RANISH to use LBA disk access if provided by BIOS and modified it from top downto its disk access functions to only use LBA addressing (with CHS reads / writes as a fallback if no LBA BIOS is provided).

I am now confused that my BIOS, when querying drive parameters via INT 13 function 48h returns a totally different geometry in comparison to what function 8 returns. While function 8 is giving me CHS values 1024,255,63 as expected for a large disk, function 48 seems to passthrough the CHS values the IDE controller returns, that is 16383,16,63.

I am now wondering which one to trust when creating partitions and MBR entries. My guess would be that I use the geometry returned by function 8 for compatibility reasons and function 48 to get the disk size (total sectors). From both values one could calculate another "virtual" cylinder count.

I am missing something essential here or could that be the way to go forward?
ecm

Homepage E-mail

Düsseldorf, Germany,
19.11.2022, 21:16

@ boeckmann
 

Ranish Partition Manager revived

Post reply

> I am now confused that my BIOS, when querying drive parameters via INT 13
> function 48h returns a totally different geometry in comparison to what
> function 8 returns. While function 8 is giving me CHS values 1024,255,63 as
> expected for a large disk, function 48 seems to passthrough the CHS values
> the IDE controller returns, that is 16383,16,63.
>
> I am now wondering which one to trust when creating partitions and MBR
> entries. My guess would be that I use the geometry returned by function 8
> for compatibility reasons and function 48 to get the disk size (total
> sectors). From both values one could calculate another "virtual" cylinder
> count.
>
> I am missing something essential here or could that be the way to go
> forward?
I think both geometries are essentially fictional.

13.08 is the way to go for things that expect to use CHS access functions, because it should match what these functions (13.02 and 13.03) use. This includes old loaders and OSes. Therefore this also includes the CHS triplets in the MBR partition table entries, as well as the H and S values in a FAT FS's BPB. (But note that one controller + ROM-BIOS combo could use different geometry than another combo for the same disk.)

The 13.48 function is documented as giving a "physical" geometry with three dwords but I believe this is mostly fictional as well. It may fit what the controller uses or not. In any case, there's an amount of total sectors qword which you can use to get the exact size of the disk without any geometry calculations. And for the 13.4x functions you do only need the Logical Block numbers, and therefore only the total blocks amount, not the "physical geometry".

---
l
boeckmann

Aachen, Germany,
19.11.2022, 21:46

@ ecm
 

Ranish Partition Manager revived

Post reply

> I think both geometries are essentially fictional.
>
> 13.08 is the way to go for things that expect to use CHS access functions,
> because it should match what these functions (13.02 and 13.03) use. This
> includes old loaders and OSes. Therefore this also includes the CHS
> triplets in the MBR partition table entries, as well as the H and S values
> in a FAT FS's BPB. (But note that one controller + ROM-BIOS combo could use
> different geometry than another combo for the same disk.)
>
> The 13.48 function is
> documented as giving a "physical" geometry with three dwords but I
> believe this is mostly fictional as well. It may fit what the controller
> uses or not. In any case, there's an amount of total sectors qword which
> you can use to get the exact size of the disk without any geometry
> calculations. And for the 13.4x functions you do only need the Logical
> Block numbers, and therefore only the total blocks amount, not the
> "physical geometry".
Thanks for the feedback. So it is essentially like I thought it should be. While the geometry is not strictly needed when working with LBA I want to leave the CHS view Ranish provides intact. So I have to calculate the cylinder count from the H and S of 13.08 and the total sectors returned by 13.48 by myself. Of course all relevant source positions have to be verified that the fallen 1024/255/63 CHS barrier does not lead to corrupt CHS entries anywhere on disk. Certainly on many places warning / error messages have to be implemented to inform the user of non-compatible partition locations etc.

I think it is a good idea to make the LBA view the default when operating on disk >8GB but I want to leave the user the option to switch back to CHS view.
bretjohn

Homepage E-mail

Rio Rancho, NM,
21.11.2022, 01:59

@ boeckmann
 

Ranish Partition Manager revived

Post reply

> I think it is a good idea to make the LBA view the default when operating
> on disk >8GB but I want to leave the user the option to switch back to CHS
> view.
You can do that, but of course when they see a disk >8GB in CHS mode it will just look like it's 8GB. I personally think the default view should be LBA and CHS should just be for "informational" purposes (but things should appear red when the numbers don't agree).

I discovered as I have been writing by USBDRIVE program that you really can't trust much of anything a disk tells you, particularly when comparing LBA to the various CHS values (MBR, PBR/VBR, GPT, BPB, ...). You will find the same entries for various places even on the same disk don't agree with each other. I've even found some USB disks that don't report the correct LBA value. That is, one of the parameters that is returned by the low-level functions is the Maximum LBA which should always be an odd number since the total number of sectors should always be even and LBA numbering starts at zero. Some disks actually return the total number of sectors (which is the Maximum LBA + 1) which must be adjusted.

I have also been doing a little bit of testing with Ranish to see what the MBRs provided by various VMs look like and see a lot of red (inconsistent/invalid) values showing up. Also, sometimes VMs provide MBR disks and other times provide PBR/VBR ("superfloppy") disks which makes thing even more troublesome. I have some routines in USBDRIVE that try to detect whether something is an MBR or a PBR/VBR and operate accordingly.
boeckmann

Aachen, Germany,
23.11.2022, 15:38

@ bretjohn
 

Ranish Partition Manager revived

Post reply

> You can do that, but of course when they see a disk >8GB in CHS mode it
> will just look like it's 8GB. I personally think the default view should
> be LBA and CHS should just be for "informational" purposes (but things
> should appear red when the numbers don't agree).
>
My current implementation is that in CHS mode you see the adjusted CHS values (cylinder can be over 1023) calculated from LBA value. Otherwise the values would always "be red" because of CHS LBA disagree. When storing the MBR to disk the CHS value gets set to the largest possible value 1023/255/63 if the 8GB bounds gets excceded to indicate an error condition. I am not 100% sure yet it is the best possible solution. But initial testing shows it is working well. In the next time I will examine how MS-DOS 7.1 FDISK is handling this case. I think it is best to follow the Microsoft way here.

> I discovered as I have been writing by USBDRIVE program that you really
> can't trust much of anything a disk tells you, particularly when comparing
> LBA to the various CHS values (MBR, PBR/VBR, GPT, BPB, ...). You will find
> the same entries for various places even on the same disk don't agree with
> each other. I've even found some USB disks that don't report the correct
> LBA value. That is, one of the parameters that is returned by the
> low-level functions is the Maximum LBA which should always be an odd number
> since the total number of sectors should always be even and LBA numbering
> starts at zero. Some disks actually return the total number of sectors
> (which is the Maximum LBA + 1) which must be adjusted.
 Thanks that is a very good hint. I will incorporate that to the LBA disk size detection.
boeckmann

Aachen, Germany,
26.11.2022, 17:36

@ boeckmann
 

Ranish Partition Manager revived

Post reply

A first version capable of doing LBA and disk sizes >8GB can now be downloaded.

https://codeberg.org/boeckmann/ranish/releases
Zyzzle

29.11.2022, 06:06

@ boeckmann
 

Ranish Partition Manager revived

Post reply

> A first version capable of doing LBA and disk sizes >8GB can now be
> downloaded.
>
> https://codeberg.org/boeckmann/ranish/releases
 That's very nice progress. Has a compiled binary been provided for this new version, or have I missed it, and we need to compile it ourselves?
boeckmann

Aachen, Germany,
29.11.2022, 09:33

@ Zyzzle
 

Ranish Partition Manager revived

Post reply

> > A first version capable of doing LBA and disk sizes >8GB can now be
> > downloaded.
> >
> > https://codeberg.org/boeckmann/ranish/releases
> That's very nice progress. Has a compiled binary been provided for this new
> version, or have I missed it, and we need to compile it ourselves?
Version 2.90.04 can be downloaded in binary form. When running part.exe LBA support is indicated by "LBA" in the lower left corner of the screen. If it is present and the disk is still 8GB in size, then this is a limitation of the BIOS.

Direct download link: https://codeberg.org/attachments/8191bbce-605b-4671-9017-a34b36a71cc1
Zyzzle

29.11.2022, 09:35

@ boeckmann
 

Ranish Partition Manager revived

Post reply

> Dicect download link:
> https://codeberg.org/attachments/8191bbce-605b-4671-9017-a34b36a71cc1
 Thank you. For some reason, I couldn't see that binary download.
boeckmann

Aachen, Germany,
29.11.2022, 11:02
(edited by boeckmann, 29.11.2022, 12:14)

@ Zyzzle
 

Ranish Partition Manager revived

Post reply

> > Dicect download link:
> > https://codeberg.org/attachments/8191bbce-605b-4671-9017-a34b36a71cc1
> Thank you. For some reason, I couldn't see that binary download.
Released version 2.90.05, which changes the way the virtual CHS geometry is determined when working in LBA mode. Additionaly, a custom disk geometry may be given as a command line argument in case the one detected by Ranish is wrong.

The standard geometry table (given in README.md) Ranish uses is now only used if geometry information returned by BIOS is obviously wrong. It is obviously wrong if head count is <= 16 while the disk is greater than >504 MB.

In the previous release the head count was unconditionally set to 255 heads if disk is larger than 4GB. Having thought about it for some time it seems not such a great idea, because some BIOSes for example return 256 as a head count. If in such a case 255 heads and 63 sectors per track must be used it may be forced with part.exe -g 0,255,63. Total cylinder count then is calculated based on disk size and given heads, sectors per track.

A also think the releases are a little bit hard to find. In that area Gitea could use a little bit of rework.
fritz.mueller

Homepage

Munich, Germany,
09.01.2023, 20:09

@ boeckmann
 

Ranish Partition Manager revived

Post reply

(this is a copy as the first version of this text seems to have died when sending it)

Hi,
I just tested Ranish 2.90.06alpha, read that it should support 2TB HDs and so
tested it with 2,00 TB in virtualbox and on a vmware with 3 TB. At each of the tests
were problems, so I started again and tried to documentate it and to add proposals for
it.
I am aware that the GUI is more than twenty years old and not originally made by
Mr. Böckmann. I think it is not intuitive.

My main request would be a simple text file that explains with which steps you
can create one or two primary partitions. The html file starts with 31 partitions,
but I think it is more important that users understand how the basic GUI works
with simple cases. And I needed a long time to find out the following steps (still not sure if everything is ok).


a) I created a 2,00 TB virtual hd (vhd) with virtualbox.
b) I booted from a FD floppy (img) with part.exe on it (self created).
c) I started part
d) It shows me Disk 1, 2097151 MiB Base, 0, Total 4,294,965.248 sectors
e) I want to create a 2 TB FAT32 partition (max. size)
f) I went to Nr. 1, section "Active" to make it active (spacce bar, till " 1 -Active - YES" is shown.
g) Pressed F2 and get message "MBR was saved to hard disk.
h) goto "File System Type", press "insert" and choose 0x0C FAT-32 LBA
PROPOSALS:
- there is enough space to replace "ins" by "insert",
- "ins": could you add the max size at 0x0B (FAT32, 0x0C FAT-32 LBA) etc.
as it is already done at 0x04, FAT-16 (32 M). Not each user knows
the differences between all this FAT types (etc.) This would make
decisions which format has to be chosen easier.
- "ins": It is not self explaining why some file system types are in blue and
some are in black colour, although the comments say YES YES YES
i) Pressed F2 and get message: "MBR was saved to hard disk.
j) Partition 1 says: Start sector 63, number of sectors: 4.294.965.185,
ending sector: 4.294.965.247 (dots were added by me)
PROPOSAL: adding dot or commas as 1.000.000 separator would be nice.
k) No idea if I also have to press "ENTER" and what to change in the following window.
I decided to do nothing.
Proposals:
(At the end it showed the following:
Minimum partition size: 4173837 sectors = 2,086,918 kB
Current partition size: -2111 sectors = 2,147,482,592 kB
Maximum partition size: 1073739789 sectors = 536,869,894 kB
there was nothing in red, how can there be negative sectors and a current size bigger
than maximum? My failure or a bug? Ah, entering a volume label and F2 made the
current size red! But it is accepted.)
l) I pressed "f" format. Lasts a while.
PROPOSAL: The following window says: "Verifying" - NOT "Formatting",
x percent "verified" instead of "formatted" Is this correct?
m) Got message: "Format correct". Entered "Quit" and then rebooted.
n) After reboot entered C:\, worked. C: is free as expected.
o) Sys c: worked fine too. Created folder: "C:\freedos\bin. Copied all files from
my diskette there.
(I had different results at this point, from crazy filenames, not booting system
etc.
p) Rebooted and tried to install FreeDOS live CD FDT2301. It had problems with the
"move" packet and chkdsk (supports FAT 16 only) reports errors.
q) Rebooted and started live mode and dosfsck:
Dosfsck says: File system has -267 clusters but only space for 1.073.741.566 FAT entries.
(dots were added by me).

Some more proposals from my side:

In most cases you have to enter F2 - Save changes although you do not want to change.
ESC often also does not work so that it is confusing how to make mistakes undone. There
is no "X" (close program) or Alt-F4 or another key available (found none).
It is not really intuitive.

I) / A) / U): Could you replace "Advanced" by "Advanced boot manager"? Would be easier
to understand. Referring to this: I also had problems to understand that the standard
IPL is installed by "doing nothing". But

I checked the vhd with my Win Hexeditor - it shows a 2,0 TB partition but says
"unknown filesystem". So I am sure that anything went wrong (see k)).
boeckmann

Aachen, Germany,
13.01.2023, 16:41

@ fritz.mueller
 

Ranish Partition Manager revived

Post reply

Thank you Fritz for documenting your findings! It will take some time for me to work through it.

Some general notes on using the software
1) Pressing "insert" only creates a partition of the type selected from the shown list. It does NOT format a partition.

2) Press "F" over a partition for formatting a FAT16 or FAT32 partition. The disk surface gets verified, therefore "verfying" is displayed. That takes a long time. You may skip verification by entering "/quick" as format argument.

3) Pressing "return" over a partition shows FILE SYSTEM information. If the partitions is not formatted that information contains GARBAGE.

Large disk support
Until now I only tested large disk support for disk up to 137GB, because my BIOS does not support larger disks. If using a 2TB disk or even larger there may be some edge cases that get triggered. I will setup a VM and try to reproduce your errors.

The boot loaders are not patched for LBA access yet. That means, they are not able to boot from partitions starting beyond the 8.4GB barrier. The simple IPL gets installed by default when not using part.exe in advanced boot manager mode.

Interface
The inferface is "complicated" to say at least, and it would benefit from a complete overhaul. Sadly, I can not accomplish that in a reasonable timeframe. In the code, there is no clear distinction between interface and partition logic. Therefore a major overhaul of the interface would nearly result in a complete rewrite.

But I will try to incorporate the "cosmetic" changes you mentioned.

Questions
1) Have you tried to format the created 2TB partition with the operating system supplied format utility?
fritz.mueller

Homepage

Munich, Germany,
14.01.2023, 11:41
(edited by fritz.mueller, 14.01.2023, 12:11)

@ boeckmann
 

Ranish Partition Manager revived

Post reply

Hi,
first of all I wanted to let you know that fdisk behaves strange on virtual hds that are a little bigger than 2 tb, see:
https://gitlab.com/FreeDOS/base/fdisk/-/issues/11
I would have expected that it says hd to big or no hd but that it works with the size bigger than 2 tb and creates this oversize as partition astonished me. i will try to find out what happens with 4 tb of size.
In the meantime I am able to create up to 7 smaller working partitions that behave like fdisk by using an extended partition and 4 logical volumes.
the trick is to enter the extended partition like in fdisk, save all this and with 2x enter you come into a new partition menu.
ranish shows a maximum size of 2 tb on the left top.
you will get more feedback later.

> Thank you Fritz for documenting your findings! It will take some time for
> me to work through it.
>
> Some general notes on using the software
> 1) Pressing "insert" only creates a partition of the type selected from the
> shown list. It does NOT format a partition.
>
> 2) Press "F" over a partition for formatting a FAT16 or FAT32 partition.
> The disk surface gets verified, therefore "verfying" is displayed. That
> takes a long time. You may skip verification by entering "/quick" as format
> argument.
>
> 3) Pressing "return" over a partition shows FILE SYSTEM information. If the
> partitions is not formatted that information contains GARBAGE.
>
> Large disk support
> Until now I only tested large disk support for disk up to 137GB, because my
> BIOS does not support larger disks. If using a 2TB disk or even larger
> there may be some edge cases that get triggered. I will setup a VM and try
> to reproduce your errors.
>
> The boot loaders are not patched for LBA access yet. That means, they are
> not able to boot from partitions starting beyond the 8.4GB barrier. The
> simple IPL gets installed by default when not using part.exe in advanced
> boot manager mode.
>
> Interface
> The inferface is "complicated" to say at least, and it would benefit from a
> complete overhaul. Sadly, I can not accomplish that in a reasonable
> timeframe. In the code, there is no clear distinction between interface and
> partition logic. Therefore a major overhaul of the interface would nearly
> result in a complete rewrite.
>
> But I will try to incorporate the "cosmetic" changes you mentioned.
>
> Questions
> 1) Have you tried to format the created 2TB partition with the operating
> system supplied format utility?
boeckmann

Aachen, Germany,
14.01.2023, 12:48

@ fritz.mueller
 

Ranish Partition Manager revived

Post reply

> ranish shows a maximum size of 2 tb on the left top.
Thats right. That is a hard limit resulting from the used data type. 2^32 512 byte sectors leads to 2TB.

It MAY be possible to create partitions starting in the 2TB boundary and spanning across it, but the software is not able to access the sectors beyond 2TB. This could lead to all sort of errors, even completely destroying to the data by overwriting the first sectors of the disk when the sector number overflows.

HIGHLY recommending not doing it.
boeckmann

Aachen, Germany,
14.01.2023, 13:06

@ fritz.mueller
 

Ranish Partition Manager revived

Post reply

> Hi,
> first of all I wanted to let you know that fdisk behaves strange on virtual
> hds that are a little bigger than 2 tb, see:
> https://gitlab.com/FreeDOS/base/fdisk/-/issues/11
Sadly reaching or crossing the 2TiB barrier triggeres undefined behaviour in many DOS programs.

The FreeDOS format utility for example crashes with a division by zero error when formatting partitions exactly or slightly less in size than 2TiB. Not sure if there is a bug report opened yet.
boeckmann

Aachen, Germany,
14.01.2023, 11:47

@ boeckmann
 

Ranish Partition Manager revived

Post reply

As a primarly feedback to Fritz post a can confirm that there are software bugs when creating large FAT32 (and propably FAT16) partitions.

1) Detection of cluster size is broken for large partitions. A maximum cluster size of 8 sectors (4k) is used resulting in essentially broken partitions.
2) Calculation of FAT table size is not done according to standard resulting in wasted space.

I already fixed these bugs for the FAT32 format routine and they should be gone in the next release.

Further there are display errors (negative numbers) resulting from wrong format specifiers in printf instructions when showing file system information. Fixed in the next release.

I will audit all the FAT related code and bring it to conformance with the standard, so the next release will take some time to prepare.
boeckmann

Aachen, Germany,
14.01.2023, 13:22

@ boeckmann
 

Ranish Partition Manager revived

Post reply

> As a primarly feedback to Fritz post a can confirm that there are
> software bugs when creating large FAT32 (and propably FAT16)
> partitions.
You may track the status of the bugs here:
https://codeberg.org/boeckmann/ranish/issues
rr

Homepage E-mail

Berlin, Germany,
14.01.2023, 14:56

@ boeckmann
 

Ranish Partition Manager revived

Post reply

> > As a primarly feedback to Fritz post a can confirm that there are
> > software bugs when creating large FAT32 (and propably FAT16)
> > partitions.
>
> You may track the status of the bugs here:
> https://codeberg.org/boeckmann/ranish/issues
Thanks for taking care to give people another choice for partitioning a drive.

---
Forum admin
fritz.mueller

Homepage

Munich, Germany,
14.01.2023, 15:25

@ rr
 

Ranish Partition Manager revived

Post reply

Just to complete the oversize tests:

5 TB HD, created with vmware:
fdisk says that total disk space is 925699 Mbytes
allows an active partition 925691 FAT32L
format c: exits with "Invalid Drive! Aborting. Error 61.
fdisk shows a primary c 925699 MB FAT32L 100%
Ranish shows disk1 2097151 MB, total 4,294,967,295 sectors
Part 1 is active, FAT-32 LBA, starting sector 63, number of sectors:
1,895,814,522, ending sector, 1,895,814,584 , size: 903 Gi
CHS cylinder wrap-around detected. Replaced by LBA marker on next save.
Saved with F2. Reboot.
format c: says Invalid drive! Aborting. Error 61.
fdisk still says C, 1, A, Pri DOS, 925699 MB, FAT32L, 100%
boeckmann

Aachen, Germany,
14.01.2023, 22:26

@ fritz.mueller
 

Ranish Partition Manager revived

Post reply

I uploaded version 2.90.08. Several FAT related bugs should be gone. I was able to create a FAT32 partition spanning the whole 2TiB disk.

https://codeberg.org/boeckmann/ranish/releases

I also was able to put FreeDOS onto the partition and boot from it. But it seems that the FreeDOS SYS command puts new boot code into the MBR, which could not successfully boot the partition. I had to uninstall the SYS provided boot code with RANISH using the "uninstall boot manager by loading standard IPL" command. After that FreeDOS successfully booted.
ecm

Homepage E-mail

Düsseldorf, Germany,
15.01.2023, 12:50

@ boeckmann
 

Ranish Partition Manager revived

Post reply

> I also was able to put FreeDOS onto the partition and boot from it. But it
> seems that the FreeDOS SYS command puts new boot code into the MBR, which
> could not successfully boot the partition. I had to uninstall the SYS
> provided boot code with RANISH using the "uninstall boot manager by loading
> standard IPL" command. After that FreeDOS successfully booted.
I don't think that SYS specifically does that. You can read the source code and point out if you find it writes an MBR anywhere, but I don't think it does.

How did you install FreeDOS onto the disk? There probably is a part of the installer that will / can write an MBR loader.

---
l
fritz.mueller

Homepage

Munich, Germany,
15.01.2023, 15:32

@ ecm
 

Ranish Partition Manager revived

Post reply

Hi
I just tested the latest ranish with a virtual 1 GB HDand still have some questions with it.
a) How can I create 3 partitions, an extended partition and more than 4 logical volumes?
b) When creating partitions, it would be helpful to show the maximum number of sectors for this partition in the window. otherwise you always have to calculate the max. value.
c) maybe this is the problem you mentioned before: I created all partitions with ranish, everything was shown green, I rebooted FD and tried to run format,
format c: worked, at all others format reported an error message:
Warning: Resets bad cluster marks, if any.
Drive_IO (write 0 count 1) [FAT12/16] [drive d*]
critical error during disk access
DOS driver error (hex): 01
Description: unknown unit for driver
Program terminated
I started ranish and formatted d - i, rebooted, and now everything works fine.
d) among others I got an interrupt divide by zero, see screenshot.
e) I removed a partition (file system type "unused" and erased it. Ranish still shows the settings of the ("none") partition after a reboot. Means:
Partition entries are still in MBR.
f) The good news: FreeDOS does no longer give out messages about other expected sizes, see screenshot.
[image]
[image]
[image]
[image]

Will continue testing asap.
boeckmann

Aachen, Germany,
15.01.2023, 16:14

@ fritz.mueller
 

Ranish Partition Manager revived

Post reply

A few hints Fritz: If you want to delete a partition press DELETE. You may set the partition type to "unused" via INSERT but that does not alter the size of the partition nor deletes it.

The erase function does not erase the partition itself but fills it with zero. That is primarly usefull if you want to erase all data in that partition. Therefore the menu entry is called "Erase partition DATA". Perhaps it should be named differently to avoid confusion.

The crash in the picture results in trying to display FAT file system information on a partition previously filled with zeros. That is a software bug. I will implement some checks to make sure that screen is only accessible on a formatted FAT partition.

If there is interest I can try to make a screencast or something in a tutorial style howto create complex partition layouts.
fritz.mueller

Homepage

Munich, Germany,
15.01.2023, 16:49

@ boeckmann
 

Ranish Partition Manager revived

Post reply

> A few hints Fritz: If you want to delete a partition press DELETE. You may
> set the partition type to "unused" via INSERT but that does not alter the
> size of the partition nor deletes it.
Thanks for this explanation. Would it be possible to add a "d) delete" text to make this easier to see as it is not self explaining. Thx.

One more thing that I just noticed:
I changed the extended partition (number 4 of first GUI) where 4 logical volumes were inside to FAT32-LBA and saved it. There was NO warning that these volumes will go to hell which is very dangerous.
When I go to this new FAT32-LBA and press ENTER I also get an interrupt divide by zero (machine was booted before).


>
> The erase function does not erase the partition itself but fills it with
> zero. That is primarly usefull if you want to erase all data in that
> partition. Therefore the menu entry is called "Erase partition DATA".
> Perhaps it should be named differently to avoid confusion.
Yes, I have seen that everything including FATs was deleted. This is a good
feature (of course also dangerous).
But sometimes you want to delete everything (making the HD a "virgin", means: even MBR should disappear). Is there a way to do this with ranish? Something like shred /dev/sda in Linux.
>
> The crash in the picture results in trying to display FAT file system
> information on a partition previously filled with zeros. That is a software
> bug. I will implement some checks to make sure that screen is only
> accessible on a formatted FAT partition.
>
> If there is interest I can try to make a screencast or something in a
> tutorial style howto create complex partition layouts.
This should be done under all circumstances when the final version is out, yes.
boeckmann

Aachen, Germany,
15.01.2023, 18:37

@ fritz.mueller
 

Ranish Partition Manager revived

Post reply

> Thanks for this explanation. Would it be possible to add a "d) delete" text
> to make this easier to see as it is not self explaining. Thx.
 I changed the menu slightly to hopefully make it clearer, but it is only an interim solution until I had time to streamline the interface a little bit more. But at the moment I am concentrating on the more technical items. I will update the inferface when the software is stable enough to call it beta.

> One more thing that I just noticed:
> I changed the extended partition (number 4 of first GUI) where 4 logical
> volumes were inside to FAT32-LBA and saved it. There was NO warning that
> these volumes will go to hell which is very dangerous.
> When I go to this new FAT32-LBA and press ENTER I also get an interrupt
> divide by zero (machine was booted before).
 Both problems should be fixed. I uploaded version 2.90.09.

> But sometimes you want to delete everything (making the HD a "virgin",
> means: even MBR should disappear). Is there a way to do this with ranish?
 At the moment there is no way to fill the first sector completely with 0. The partition table may be zeroed out but the boot code is not touched.

All other sectors beside the MBR may be deleted with a trick: delete all partitions and create a new one, make sure partition starts at sector 2 and spans the whole disk, then erase the partition content.
boeckmann

Aachen, Germany,
15.01.2023, 18:49

@ ecm
 

Ranish Partition Manager revived

Post reply

> > I also was able to put FreeDOS onto the partition and boot from it. But
> it
> > seems that the FreeDOS SYS command puts new boot code into the MBR,
> which
> > could not successfully boot the partition. I had to uninstall the SYS
> > provided boot code with RANISH using the "uninstall boot manager by
> loading
> > standard IPL" command. After that FreeDOS successfully booted.
>
> I don't think that SYS specifically does that. You can
> read
> the source code and point out if you find it writes an MBR anywhere,
> but I don't think it does.
>
> How did you install FreeDOS onto the disk? There probably is a part of
> the installer that will / can write an MBR loader.
My bad, there wasn't boot code at all in the MBR. It was a fresh virtual disk image. So I think it was a BIOS message.

I simply did a SYS to put the OS onto the disk.
fritz.mueller

Homepage

Munich, Germany,
15.01.2023, 20:40

@ boeckmann
 

Ranish Partition Manager revived

Post reply

One more thing I just noticed: It seems to be possible to format extended and lba extended. At least it behaves like this. But it seems to write only the necessary information?

Referring to GUI: Do not worry, I can understand that you cannot do everything at once. But it is better for me to report it instead of keeping it in mind and forgetting it later. Ranish is not the first program that I test. At FD you sometimes have a longer time...;-)
boeckmann

Aachen, Germany,
16.01.2023, 13:44

@ fritz.mueller
 

Ranish Partition Manager revived

Post reply

> One more thing I just noticed: It seems to be possible to format extended
> and lba extended. At least it behaves like this. But it seems to write only
> the necessary information?
It writes boot code into the extended MBR and deletes all paritions within.
fritz.mueller

Homepage

Munich, Germany,
17.01.2023, 22:44

@ boeckmann
 

Ranish Partition Manager revived

Post reply

Hi,
I just wanted to report two more things with the actual version of ranish:

a) when I have a HD with 3x primary partitions and an extended partition with several logical volumes inside: I deleted one of the logical volumes, no problem, but: I found no way to come out of this window again, I always get a message that I have to fix problems at the main window (primary partitions) first. How do I come back there?
Message: You cannot save object if one of its parents is invalid or is not saved.

b) I have an old DELL laptop, in about built 2009, last bios 2010, that has no options to change BIOS settings from UEFI, AHCI etc. back to old system. Nevertheless it is able to handle SATA SSDs in FreeDOS, means, I can boot from external USB diskette, run ranish and create partitions, write a system on the SATA etc.
The interesting thing is:
I tested with a 256 GB SATA SSD, created partitions etc, formatted, checked surface of SSD, nothing special happened.
When doing this test with a 1 TB SSD from WD, creating partitions etc. works, but when I come to formatting or checking surface, there is a stop at 13 percent I get a message "One or more bad sectors found on disk" - OK - clicking on OK gives a message "BIOS error code BB".
Format says: Cluster Stats: 30516378 used. 0 bad, 943 items, 30516379 last.
Safe format: have to trash 242.193 used data sectors!
Mirror map is 1908136 bytes long, 238506 sectors mirrored.
Warning: Each FAT is 238410 sectors, > 16 MB-64K, Win9x incompatible.
100 % completed.
Optimized initia Root Directory size : 1 clusters.
Safe Quick format complete.
976.762.552,5 kb total disk space
976.524.064 kb available on disk
32 kb in each allocation unit.
30.516.377 allocation units on
disk.
After this I tested another 1 TB SSD from Samsung, disk1, 953869 mb, partition 1 NO GPT guard partition and ran verify disk surface (v).
At 13 % verifying I get the following message:
One or more bad sectors found on disk. OK
This time: BIOS error code 04.

Forgot to mention: There is no error message at virtualbox 2TB HD.

I am not sure if the two HDs really have a bad sector there.
The two HDs work fine on the same machine with Linux.

There is another thing that I noticed with a 2 TB HD. dosfsck reports an error message with 77 FATs, I could not test it once again at the moment, but I assume it is reproducable. But I think this is a dosfsck bug.
boeckmann

Aachen, Germany,
17.01.2023, 23:11
(edited by boeckmann, 17.01.2023, 23:30)

@ fritz.mueller
 

Ranish Partition Manager revived

Post reply

> b) I have an old DELL laptop, in about built 2009, last bios 2010, that has
> no options to change BIOS settings from UEFI, AHCI etc. back to old
> system. Nevertheless it is able to handle SATA SSDs in FreeDOS, means, I
> can boot from external USB diskette, run ranish and create partitions,
> write a system on the SATA etc.
> The interesting thing is:
> I tested with a 256 GB SATA SSD, created partitions etc, formatted, checked
> surface of SSD, nothing special happened.
> When doing this test with a 1 TB SSD from WD, creating partitions etc.
> works, but when I come to formatting or checking surface, there is a stop
> at 13 percent I get a message "One or more bad sectors found on disk" - OK
> - clicking on OK gives a message "BIOS error code BB".
> Format says: Cluster Stats: 30516378 used. 0 bad, 943 items, 30516379
> last.
> Safe format: have to trash 242.193 used data sectors!
> Mirror map is 1908136 bytes long, 238506 sectors mirrored.
> Warning: Each FAT is 238410 sectors, > 16 MB-64K, Win9x incompatible.
> 100 % completed.
> Optimized initia Root Directory size : 1 clusters.
> Safe Quick format complete.
> 976.762.552,5 kb total disk space
> 976.524.064 kb available on disk
> 32 kb in each allocation unit.
> 30.516.377 allocation units on
> disk.
> After this I tested another 1 TB SSD from Samsung, disk1, 953869 mb,
> partition 1 NO GPT guard partition and ran verify disk surface (v).
> At 13 % verifying I get the following message:
> One or more bad sectors found on disk. OK
> This time: BIOS error code 04.
>
> Forgot to mention: There is no error message at virtualbox 2TB HD.
>
> I am not sure if the two HDs really have a bad sector there.
> The two HDs work fine on the same machine with Linux.
>
> There is another thing that I noticed with a 2 TB HD. dosfsck reports an
> error message with 77 FATs, I could not test it once again at the moment,
> but I assume it is reproducable. But I think this is a dosfsck bug.
I think the disks are fine. 13% of 1TB could be the 137GB 28-bit LBA barrier. If you are interested what it is, the details are explained in this document: https://www.seagate.com/support/kb/disc/tp/137gb.pdf

Everything above, 48-bit LBA access is needed. Perhaps there is some incompatibility with the older BIOS and the larger disks regarding that. Perhaps someone here in the forum knows what is going on. I am not deep in this BIOS / ATA related stuff.

You have partitioned the disk with ranish and formatted with FreeDOS format? Have you used LBA partition types? Not sure if FreeDOS makes use of it.

But your problem is a good indicator that I should change the format routines in ranish, such that the last block of the partition is always written to read from?. Then this problem should be detected, even if doing a quick format (which is the default as of yesterday).
boeckmann

Aachen, Germany,
17.01.2023, 23:14

@ fritz.mueller
 

Ranish Partition Manager revived

Post reply

> Hi,
> I just wanted to report two more things with the actual version of ranish:
>
> a) when I have a HD with 3x primary partitions and an extended partition
> with several logical volumes inside: I deleted one of the logical volumes,
> no problem, but: I found no way to come out of this window again, I always
> get a message that I have to fix problems at the main window (primary
> partitions) first. How do I come back there?
> Message: You cannot save object if one of its parents is invalid or is not
> saved.
I will check that, but you should always be able to press F3 - Undo, go to the primary via ESC, save that with F2 and redo the changes you did to the logical volumes.
fritz.mueller

Homepage

Munich, Germany,
18.01.2023, 12:50

@ boeckmann
 

Ranish Partition Manager revived

Post reply

Hi,
referring to dosfsck: It works with 32 MB, but only run it with lh dosfsck c:

At the enclosed picture you can see the 2 TB HD (I created it in virtualbox with ranish, rebooted, formatted it with ranish, executed a sys c: from FDT2301 and installed FD basic + dosfsck.
The result is that it reports that only 1 or 2 FATs are supported, not 77.
Should I report this at Gitlab too?[image]
As FD format has problems with this size I do not test it with FD format.
Thanks for reading
boeckmann

Aachen, Germany,
18.01.2023, 14:05

@ fritz.mueller
 

Ranish Partition Manager revived

Post reply

> Hi,
> referring to dosfsck: It works with 32 MB, but only run it with lh dosfsck
> c:
I tested it with a 137 GiB disk (thats all my BIOS supports), a single primary FAT32 LBA partition formatted with RANISH. Dosfsck reports no error here. Perhaps it is the 2 TiB size that makes problems with dosfsck. Have you tried it with a 2TiB - 1 GiB partition?
fritz.mueller

Homepage

Munich, Germany,
18.01.2023, 14:48

@ boeckmann
 

Ranish Partition Manager revived

Post reply

> > Hi,
> > referring to dosfsck: It works with 32 MB, but only run it with lh
> dosfsck
> > c:
>
> I tested it with a 137 GiB disk (thats all my BIOS supports), a single
> primary FAT32 LBA partition formatted with RANISH. Dosfsck reports no error
> here. Perhaps it is the 2 TiB size that makes problems with dosfsck. Have
> you tried it with a 2TiB - 1 GiB partition?
I tested it again todaywith the 256 GB SSD HD and got the value 67 instead of 77.
I made a lot of screenshots, available, in case you should need it. Simply send a mail to fritz.mueller near mail.com.

Ranish worked fine with this 256 GB SSD HD, there were no 13% messages at ranish version of format and verify. At the moment the machine runs with FD fdisk c: (already done) and now with format c: /U /S - but this is at 6 GB now, will last a while. For documentation purposes and easier understanding it would be helpful to see different words at the format / verify progress bar window, although it is cosmetic.

I ran some tests with "coming back from logical volume" but could not yet reproduce it. Will report when I know more.
boeckmann

Aachen, Germany,
18.01.2023, 21:15

@ fritz.mueller
 

Ranish Partition Manager revived

Post reply

> I tested it again todaywith the 256 GB SSD HD and got the value 67 instead
> of 77.
I also tried it now under VirtualBox with a 2 TiB disk. dosfsck tries to allocate ~268MB memory, which it fails to. That is exactly the size of the FAT table, so my guess is it tries to load the table into memory as a whole. Interestingly memory allocation also fails if the VM has enough RAM. There seems to be some maximum memory allocation size. Skimming over the XMS specification my guess is that there is an block allocation limit of 64MB (maximum DX value, function number 09h).

I had success in checking a 256GB disk.

If loading dosfsck via LH as you suggested I immediately get the same error code as you. I think it is not working correctly if running with LH.
Japheth

Homepage

Germany (South),
19.01.2023, 16:39

@ boeckmann
 

Ranish Partition Manager revived

Post reply

> Interestingly memory allocation also fails if the VM has enough RAM.
> There seems to be some maximum memory allocation size. Skimming over the
> XMS specification my guess is that there is an block allocation limit of
> 64MB (maximum DX value, function number 09h).
This might actually be a problem of the dpmi host you're using. If it's cwsdpmi, try a newer version or try hdpmi32.

---
MS-DOS forever!
rr

Homepage E-mail

Berlin, Germany,
12.11.2022, 11:10

@ boeckmann
 

Ranish Partition Manager revived

Post reply

> What this software sets apart from other software is its advanced boot
> manager capability. Especially in a multi OS environment you can
> effectively shield the different OS partitions from each other while
> letting the OSes share some common data partitions. So if you want to setup
> DOS / Windows / OS2 etc. on the same host its worth giving it a try. Be
> aware that it is not the simplest software to use and there is some
> learning curve.
In which ways does the feature differ from BOOTMGR?

---
Forum admin
boeckmann

Aachen, Germany,
12.11.2022, 12:09

@ rr
 

Ranish Partition Manager revived

Post reply

> > What this software sets apart from other software is its advanced boot
> > manager capability. Especially in a multi OS environment you can
> > effectively shield the different OS partitions from each other while
> > letting the OSes share some common data partitions. So if you want to
> setup
> > DOS / Windows / OS2 etc. on the same host its worth giving it a try. Be
> > aware that it is not the simplest software to use and there is some
> > learning curve.
>
> In which ways does the feature differ from
> BOOTMGR?
Well that first sentence of your quote was perhaps a bit imprecise of me. That was primarily targeted at the tools provided with the operating systems.

But in comparison to BOOTMGR I would consider the tight integration between the partition manager and the boot loader a benefit. Correct me if I am wrong, I think BOOTMGR operates on the MBR and thus is limited to 4 primary partitions? In RANISH up to 32 partitions can be created and theoretically you can install DOS on 16 of them (boot menu fits 16 entries), because in its advanced mode RANISH is using its own partition table (and patches MBR upon boot).

If using only two primary partitions or so I think there is no much difference between the two and using an ordinary fdisk + BOOTMGR would yield a similar result?
rr

Homepage E-mail

Berlin, Germany,
12.11.2022, 21:08

@ boeckmann
 

Ranish Partition Manager revived

Post reply

> But in comparison to BOOTMGR I would consider the tight integration between
> the partition manager and the boot loader a benefit. Correct me if I am
> wrong, I think BOOTMGR operates on the MBR and thus is limited to 4 primary
> partitions? In RANISH up to 32 partitions can be created and theoretically
> you can install DOS on 16 of them (boot menu fits 16 entries), because in
> its advanced mode RANISH is using its own partition table (and patches MBR
> upon boot).
Ok, I see now.

---
Forum admin
roytam

16.11.2022, 16:39

@ boeckmann
 

Ranish Partition Manager revived

Post reply

Glad to see this is revived.
I wonder if it can create partitions in size-aligned(4KB/1MB) position, besides GPT?
boeckmann

Aachen, Germany,
17.11.2022, 12:04

@ roytam
 

Ranish Partition Manager revived

Post reply

> Glad to see this is revived.
> I wonder if it can create partitions in size-aligned(4KB/1MB) position,
> besides GPT?
You may align the partitions in any kind you want to, only overlapping partitions are not allowed. The start and end of a partition can be specified sector-exact. But Ranish is currently not supporting you by enforcing a particular alignment, if it is that you were referring to. But such a thing could be implemented relatively easy if demand is existing.

GPT support is sadly not one of my short- to medium-term goals. At least not with the current Ranish code base.

But with the things I learned the last weeks by reading the ranish code ideas are coming into my mind of how a partition manager which supports GPT while retaining DOS compatibility could work. Momentarly Ranish uses its own partitioning scheme in advanced mode and patches MBR on every boot. I see no reason why it should not be possible to replace that custom partitioning scheme with the now-standard GPT.
usotsuki

27.11.2022, 10:00

@ boeckmann
 

Ranish Partition Manager revived

Post reply

> Hi,
>
> does anyone remember Ranish Partition Manager, a software from the end of
> the 90s?
I still actually use it.

My 286 has the bootloader installed, and many of the VMs I've spun up use it too so I don't have to use "-boot a" in QEMU.
Back to index page
Thread view  Board view
22049 Postings in 2034 Threads, 396 registered users, 198 users online (0 registered, 198 guests)
DOS ain't dead | Admin contact
RSS Feed
powered by my little forum