中国DOS联盟

Brother Budian, if the C: drive has Windows 98 installed and the D: drive has Windows XP installed, and you put grldr and menu.lst on the C: drive and add "c:\grldr=xxxx" at the end of boot.ini, but after selecting find /menu.lst it restarts automatically. Does it have to require Windows 2000/XP/2003 to be installed on the C: drive?

Brother Budian, I've tried the FTP you provided many times, but it always says it can't connect or has no access. Could it be related to me being abroad?

Esme:

All four versions have been sent to you by Email, please check it.

blue777:

What you mentioned is probably a big bug? Can you further test for a longer period, consider various possibilities, and then determine the cause of the bug? If you already have a judgment or a guess now, please write it out.

Okay, the above situation is reported by my several users. I will install 98 and XP for testing in person. Also, from which version did grub remove the RUN command? Because I want to add a boot ISO item to the floppy boot disk.

.title Boot from iso file (hd0,0)/iso_file.iso
run /ieldr
The floppy disk has ieldr and isoemu.ini. Put isoemu.ini and iso_file.iso in (hd0,0), and edit isoemu iso_file.iso to boot from the ISO. But the pre4 and even the previous version to grub 0.90 have already removed the RUN command. How to handle it?

Since it's not the problem you encountered, then I guess it might be a false positive, because the occurrence of such a problem in pre4 should be said to be unlikely. The user may have confused pre4 with other previous versions, because they are all called GRLDR, and it becomes invalid after renaming.

You may have made a mistake yourself, because the RUN command was added by GANDALF, and only the versions he released have the RUN command. The versions released by me do not have the RUN command, nor do they have support for ISO, CDROM.

I don't know that you posted without the RUN command. Because the Grub command index http://grub4dos.sourceforge.net/grubcmd_ch.htm has the RUN command. scdrom is compiled in by ganfalf, which I know. I've always used cd. bs+BCDL or Smart Boot Manager to boot the CD. Can you add a function to call grldr, ieldr, etc. in the command line?

Brother Budian, you still sent me software during your busy schedule. How can a single "thank you" suffice? I've tested all versions, and the results are optimistic. All versions have no conflicts with ntfspro, so the conflict with ntfspro was also a false positive, which wasted everyone's time. I checked the previous reason for the conflict and found that the problem was with the pre4 version of grldr I was using. Because I couldn't open the ftp, I extracted grldr from the compressed package provided by blue777 at the bottom of the fourth page. But he modified it according to his own functions, and some code was modified inappropriately, so there were bugs. However, this also revealed another problem, that is, the ftp or the grub download page of the new DOS need to be improved, and it is best to be able to provide files of various versions in a timely and comprehensive manner. Also, I have some suggestions for blue777. I think it is not necessary to implant the run command in grub 4 dos to run ieldr. We can also consider treating ieldr and grldr as parallel during startup, although it may seem one more step in the startup process.

Correct again. Maybe blue777 didn't modify that grldr. I also studied the io.sys in his img and found that the encoding is completely incorrect. Also, I have read some articles about UltroISO compression distortion before. Maybe the problem is with the img file itself. It is suggested that you re-compress the img with UltroISO, and maybe the CD-ROM can be booted. Just a guess, heh heh

I didn't change any code in that img's grldr at all, and I don't have that level of expertise. I just followed the installation method introduced by tinybit and installed it on a floppy disk. But so far, it still can't be used as a bootable CD. Because I can't keep bothering Brother Budian, so I gave up using grldr to boot from the floppy disk and still use the previous stage1, stage2 to boot. (So unlucky!) I only make floppy disk boot disks, so what you said about parallel booting grldr and ieldr in boot.ini is not meaningful to me. I hope Brother Budian can help!

The 55AA in io.sys has been modified to remove the hard disk logical lock restriction. It has nothing to do with not being able to make a bootable CD.

Download the latest pre4, and there are two ways to install it to a floppy disk:
1. Use stage1, stage2: root (fd0), setup (fd0), and the resulting floppy disk image is assumed to be called a.img.
2. Use grldr, the method is omitted, and the resulting floppy disk image is assumed to be called b.img.

After testing in a virtual machine, testing with the virtual floppy drive vFloppy, and testing with a real floppy disk in a real floppy drive, the conclusion is drawn:
Both a.img and b.img can be booted.
↓
Now use a.img as the bootable file to make a test bootable CD, assumed to be called a.iso.
Use b.img as the bootable file to make a test bootable CD, assumed to be called b.iso.

After testing in a virtual machine and burning into a real CD, the conclusion is drawn:
a.iso can be booted, but b.iso cannot be booted.
Please upload the bootable b.iso if you can solve this problem, thank you!

blue777：

After the problem occurs, the cause should be identified before it can be concluded. If you give up without identifying the cause, I'm afraid it will be difficult for you to improve from now on.

The GRLDR ISO issue you mentioned has actually been almost completed and is approaching the stage of "reaching a conclusion". Just when you were about to succeed, you gave up, which is called功亏一篑.

The GRLDR.IMG file you sent me is a 1.44M ordinary floppy image. I directly replaced the 1.44M boot image on the genuine Win98 installation CD with it, and the result proved that it could successfully boot GRUB. However, the CD you made yourself with xxx software could not boot. What conclusion can you draw from this? You probably concluded that there was something wrong with the CD made with GRLDR, and it was not as good as the one made with stage1 and stage2. But you didn't ask why others could boot with the CD made with GRLDR? And it was even made with the GRLDR.img you provided yourself.

It can't be said that the conclusion you drew is completely unreasonable, but you seem to be able to draw other conclusions. This is what is called "multiple solutions". For x squared equals 1, there are two solutions, x = 1 and x = -1. These are two different solutions to the same problem.

Imagine, are the ISO making software you used all reliable? How reliable are they? Is there absolutely no chance of an error? If your software is really like this, you have completely ruled out the possibility of it making a mistake. But the problem is that you can't prove that it's not wrong unless you test whether the ISO file it generates is completely correct according to the CD boot specification. But you didn't do this test. In short, you don't want to spend more time here, so you chose to give up the CD made by the GRLDR method, and also avoided further thinking about this problem, or gave up a very good opportunity to improve. Understanding a problem will definitely improve you, me, and even everyone; if you continue to be vague, then you will stay where you are on this problem, I will stay where I am, and everyone may stay where they are.

Of course, what I'm saying is my view or habit towards problems, and I don't mean to blame the brothers. We are just talking about principles here. This is a forum after all, and we are talking about in-depth technology, not a general chat room.

> Can you add a function to call grldr, ieldr, etc. in the command line?

To boot grldr, you can do this:

chainloader (hd?,?)/grldr

Therefore, I guess that to boot IELDR, it can be similar:

chainloader (hd?,?)/ieldr

Since I haven't studied the structure of IELDR, the above is just a guess and may not be successful. If anyone knows the structure of IELDR, please explain it in detail here. I believe that after knowing its structure, we can use a suitable chainloader command line parameter to boot it. The chainloader in pre4 has this function, and the chainloader in previous versions does not. Everyone can also study the new chainloader command by themselves. You can use the help chainloader command to see the short English description. It's a pity that there aren't enough documents to explain these details now.

In x86 RealMod, after the code runs, it first copies itself from 0x0000:0x7c00 to 0x0000:0x8200 for execution, and then performs such an operation again, with the destination being 0x0000:0x0800, which is the memory space where the partition management program runs. Now, we will enumerate the system partitions, find a FAT32/NTFS partition, that is, start to look for the configuration file (isoemu.ini). If found, we start to analyze it; otherwise, we look for the next partition until all partitions are analyzed, and then push the user into the built-in Shell. When the isoemu line in the configuration file is found, we start to look for the specified iso file in the root directory of this partition and obtain its starting cluster number. Once failed, enter Shell/. When we find this file, we start to install the new INT13h interrupt of our mini-fs, and then initialize it with the starting cluster number of the iso file. Next, wait for the user's instruction and transfer to booting the system from the virtual CD-ROM. If the specified iso file is not valid or not bootable, then transfer to Shell. If bootable, then load the boot sector and transfer the control of the system to it - done! Hehe.

Correct the test results again

First, apologize. Because when I did the test of 4 versions last time, it was almost time for class, in a hurry, and I didn't want to test with a virtual machine, so I took a shortcut, resulting in inaccurate results. The process is as follows:

Add the following to boot.ini:
c:\grldr=Pre
c:\grldr02=Pre2
c:\grldr03=Pre3
c:\grldr04=Pre4

Then boot and test one by one. The result is that there is no conflict between the four versions and ntfspro. But after I came back from school and tested again, I found that the last three lines added in boot.ini are unreliable. For example, booting with Pre2 only goes through the initial of the grldr02 file, and then continues to execute grldr. Checking the grldr code shows that this is inevitable. So I had to start each version of grldr alternately, and this time I found the problem.

Still only the Pre version can run ntfs4dos pro normally.

The general results are as follows:
Pre is normal, Pre2 and Pre4 can execute ntfs4dos pro, but the machine freezes after any command. Pre3 freezes immediately when executing ntfs4dos pro.

I checked the memory and found that Pre3 seriously interferes with other programs' calls to HMA and XMS memory, and even DOS cannot load the kernel into HMA.

From this, I guess that there are bugs in the memory handling of versions Pre2/Pre3/Pre4.

blue777, if you only study floppy disk booting, Ieldr seems not useful to you? If the CD image iso file can be small enough to fit on a floppy disk, then it has no function, and it is better to boot with the floppy disk img, while booting the real mode CD grub can be competent. If you want Ieldr to call the iso in the hard disk, this function is not universal, because the files on everyone's hard disk are different