CD images, ISO

[Keith Krehbiel]

I have scratched the first CD in my set to the point where it is mostly
unuseable for installation. I downloaded an image that I am reasonable
certain would be a replacement for the one I ruined. When I try to
extract the image to copy the files and burn a new CD, the files come
out with ISO filenames, ie, all caps and truncated. I feel certain that
the linux installer will not be able to find the files in this format.
Does anyone know how to extract them without changing the names?
Copying the image intact gives exactly that: a single file on the CD.
(Or at least that seems like what will happen, I haven't actually burned
one that way.)
Also, downloading the image in the first place was a bit of a
challenge. First I tried with Mozilla and it kept running out of memory
in the desktop file even though I had redirected the file to a much
larger partition with plenty of space. Next I lowered myself to using
Explorer (5.1). It downloaded the file just fine but it was then that I
realized that the filenames had been changed. I thought maybe it was
Explorer that had somehow 'adjusted' things. I booted from a larger
partition and used Mozilla one more time. This time I got it downloaded
to my desktop with a gibberish name. It opened with Disk Copy but once
again the filenames were all ISO. That is what led me to believe that
it was Disk Copy that was changing things. Thanks in advance for any
expertise that can be lent.
Keith

[Rod Smith]

Post by Keith Krehbiel
I have scratched the first CD in my set to the point where it is mostly
unuseable for installation. I downloaded an image that I am reasonable
certain would be a replacement for the one I ruined. When I try to
extract the image to copy the files and burn a new CD

Don't try it that way. Download the image file and burn it straight to
CD-R *AS AN IMAGE FILE* (that is, do *NOT* burn a CD with a single big
file). You can do this directly with cdrecord, or most GUI CD-R tools have
an option to burn a CD-R from an image file. If you need more help on
this, post with information on the CD-R software you're using.

[Keith Krehbiel]

Post by Rod Smith

Post by Keith Krehbiel
I have scratched the first CD in my set to the point where it is mostly
unuseable for installation. I downloaded an image that I am reasonable
certain would be a replacement for the one I ruined. When I try to
extract the image to copy the files and burn a new CD

Don't try it that way. Download the image file and burn it straight to
CD-R *AS AN IMAGE FILE* (that is, do *NOT* burn a CD with a single big
file). You can do this directly with cdrecord, or most GUI CD-R tools have
an option to burn a CD-R from an image file. If you need more help on
this, post with information on the CD-R software you're using.

Actually burning a single big file was about the only option I had not
tried. The file I have downloaded has a garbage title and a generic
icon but Disk Copy will open it and create an image. The problem is
that all the files and folders in the image are named to ISO standards
rather than something that Linux would understand. I am not sure when
this happens, in Disk Copy, probably. I tried just now to create an
image from the file with my burner software (Discribe) and it wouldn't
read the file. Thanks for your suggestion though. The real filenames
are saved in each directory in a file called TABLE, something or
another. Seems like there should be some way to reconstruct it. I am
guessing that the files themselves have not been altered, only the
names. Renaming a couple thousand files by hand is not something I am
prepared to do.

[Joost Kremers]

Post by Keith Krehbiel
icon but Disk Copy will open it and create an image. The problem is
that all the files and folders in the image are named to ISO standards
rather than something that Linux would understand.

the standard file system on cd's is iso9660. this is a limited filesystem,
with e.g. something similar to the 8.3 naming limit that the original msdos
fs has. to overcome these limitations, several extensions were developed,
the most common ones being joliet (used by MS) and rock ridge (used by
linux). linux can read both extensions, but only when the kernel is
configured to do so.

i suspect that somewhere here lies your problem.

[Rod Smith]

Post by Keith Krehbiel

Post by Rod Smith

Post by Keith Krehbiel
I have scratched the first CD in my set to the point where it is mostly
unuseable for installation. I downloaded an image that I am reasonable
certain would be a replacement for the one I ruined. When I try to
extract the image to copy the files and burn a new CD

Don't try it that way. Download the image file and burn it straight to
CD-R *AS AN IMAGE FILE* (that is, do *NOT* burn a CD with a single big
file). You can do this directly with cdrecord, or most GUI CD-R tools have
an option to burn a CD-R from an image file. If you need more help on
this, post with information on the CD-R software you're using.

Actually burning a single big file was about the only option I had not
tried. The file I have downloaded has a garbage title and a generic
icon but Disk Copy will open it and create an image. The problem is
that all the files and folders in the image are named to ISO standards
rather than something that Linux would understand.

The CD-R image file is a complete ISO-9660 filesystem in a single file.
The Disk Copy tool just accesses that filesystem the way a Mac normally
does, except from the file rather than from an actual CD-ROM. The
filesystem image is almost certainly ISO-9660 with Rock Ridge extensions,
which means that it *DOES* have long filenames, but they're encoded via
Rock Ridge, which pre-X versions of MacOS can't understand. (Mac disc
images will be either HFS or ISO-9660 plus HFS, so Macs will see HFS long
filenames.) All of this is irrelevant once the CD-R is created correctly,
because what you should do is find a way to burn the image to disc
without adding it in a wrapper filesystem. That is, the usual way to
create a CD-R is to take files from a disk directory, wrap them in an
ISO-9660 filesystem (often plus other options), and write that created
filesystem to the CD-R. What you want to do is to write the file to disc
*WITHOUT* first wrapping it in an ISO-9660 filesystem, because the image
file *IS* an ISO-9660 filesystem.

Post by Keith Krehbiel
I tried just now to create an
image from the file with my burner software (Discribe) and it wouldn't
read the file. Thanks for your suggestion though.

I'm afraid I'm not familiar with Discribe, so I can't tell you precisely
how to do what's needed with it, or even if it's possible -- it's
conceivable that Discribe is crippled and lacks the necessary option. Look
for something called "burn from image file," "create CD-R from ISO image,"
or something similar. You don't want to *CREATE* an image file -- you've
already got one. You want to COPY that existing image file to disc.

If Discribe is so crippled that it won't do what you need, you might look
into other software. I'm not familiar with MacOS options, so I can't point
you to specific products, but there's got to be something that'll do the
job. If you've already installed Linux and can access the file from Linux,
just use cdrecord to burn the image file to CD-R from Linux.

Post by Keith Krehbiel
The real filenames
are saved in each directory in a file called TABLE, something or
another.

This is a common practice when using certain Unix/Linux CD-R creation
tools. Placing the long filenames in a special files enables users of OSs
that can't read them to figure out what's what. You can and should ignore
this detail, along with everything you see when accessing the file with
Disk Copy, because using Disk Copy as any part of this process is the
wrong approach -- or at the very least, it's the way-too-hard approach.

Post by Keith Krehbiel
Seems like there should be some way to reconstruct it.

In theory, yes. Reconstruction is the wrong approach, though; it's like
trying to pick all the toppings off of a pizza to make a sandwich from
them, when a sandwich just like the one you want to make is sitting right
next to the pizza. Only multiply that by a thousandfold. ;-)

[Keith Krehbiel]

Post by Rod Smith

Post by Keith Krehbiel

Post by Rod Smith

Post by Keith Krehbiel
I have scratched the first CD in my set to the point where it is mostly
unuseable for installation. I downloaded an image that I am reasonable
certain would be a replacement for the one I ruined. When I try to
extract the image to copy the files and burn a new CD

Don't try it that way. Download the image file and burn it straight to
CD-R *AS AN IMAGE FILE* (that is, do *NOT* burn a CD with a single big
file). You can do this directly with cdrecord, or most GUI CD-R tools have
an option to burn a CD-R from an image file. If you need more help on
this, post with information on the CD-R software you're using.

Actually burning a single big file was about the only option I had not
tried. The file I have downloaded has a garbage title and a generic
icon but Disk Copy will open it and create an image. The problem is
that all the files and folders in the image are named to ISO standards
rather than something that Linux would understand.

The CD-R image file is a complete ISO-9660 filesystem in a single file.
The Disk Copy tool just accesses that filesystem the way a Mac normally
does, except from the file rather than from an actual CD-ROM. The
filesystem image is almost certainly ISO-9660 with Rock Ridge extensions,
which means that it *DOES* have long filenames, but they're encoded via
Rock Ridge, which pre-X versions of MacOS can't understand. (Mac disc
images will be either HFS or ISO-9660 plus HFS, so Macs will see HFS long
filenames.) All of this is irrelevant once the CD-R is created correctly,
because what you should do is find a way to burn the image to disc
without adding it in a wrapper filesystem. That is, the usual way to
create a CD-R is to take files from a disk directory, wrap them in an
ISO-9660 filesystem (often plus other options), and write that created
filesystem to the CD-R. What you want to do is to write the file to disc
*WITHOUT* first wrapping it in an ISO-9660 filesystem, because the image
file *IS* an ISO-9660 filesystem.

Post by Keith Krehbiel
I tried just now to create an
image from the file with my burner software (Discribe) and it wouldn't
read the file. Thanks for your suggestion though.

I'm afraid I'm not familiar with Discribe, so I can't tell you precisely
how to do what's needed with it, or even if it's possible -- it's
conceivable that Discribe is crippled and lacks the necessary option. Look
for something called "burn from image file," "create CD-R from ISO image,"
or something similar. You don't want to *CREATE* an image file -- you've
already got one. You want to COPY that existing image file to disc.
If Discribe is so crippled that it won't do what you need, you might look
into other software. I'm not familiar with MacOS options, so I can't point
you to specific products, but there's got to be something that'll do the
job. If you've already installed Linux and can access the file from Linux,
just use cdrecord to burn the image file to CD-R from Linux.

Post by Keith Krehbiel
The real filenames
are saved in each directory in a file called TABLE, something or
another.

This is a common practice when using certain Unix/Linux CD-R creation
tools. Placing the long filenames in a special files enables users of OSs
that can't read them to figure out what's what. You can and should ignore
this detail, along with everything you see when accessing the file with
Disk Copy, because using Disk Copy as any part of this process is the
wrong approach -- or at the very least, it's the way-too-hard approach.

Post by Keith Krehbiel
Seems like there should be some way to reconstruct it.

In theory, yes. Reconstruction is the wrong approach, though; it's like
trying to pick all the toppings off of a pizza to make a sandwich from
them, when a sandwich just like the one you want to make is sitting right
next to the pizza. Only multiply that by a thousandfold. ;-)

Thanks for your reply. I finally applied some repair goop to my
original disk and was able to get a decent install. From the error
messages I was able to determine that at least part of the corruption
happens in the /x directory in /pool.
I have, so far, not found a way to access disks other than the system
disk from inside linux. I am not sure if I have something wrong with
the system or if it is an operator malfunction. At least this time I
have emacs and am learning to use it. Several of my attempts to install
omitted emacs and that was one of the reasons to reinstall: to get a
tool to examine a few things.
One of the things I have done is to start using a rewriteable disk so I
don't burn a coaster every time.
Don't try it that way. Download the image file and burn it straight to

Post by Rod Smith

Post by Keith Krehbiel

Post by Rod Smith
CD-R *AS AN IMAGE FILE* (that is, do *NOT* burn a CD with a single big
file). You can do this directly with cdrecord, or most GUI CD-R tools have
an option to burn a CD-R from an image file.

Discribe has an option to copy CD/DVD. I will try that next. It sounds
like what you were describing. I am able to copy the image into the
buffer in almost any format I have tried but I don't remember if I have
burnt a disc that way. I am sure I have not tried the copy CD/DVD
function yet.
Any suggestions on reading 'foreign' devices from within linux?
Thanks again for your help. It was most informative.
Keith

[Rod Smith]

Post by Keith Krehbiel
I have, so far, not found a way to access disks other than the system
disk from inside linux. I am not sure if I have something wrong with
the system or if it is an operator malfunction.

You might try posting details of what you've tried and what error messages
you get under an appropriately-titled thread. My initial guess is that
you're trying to mount HFS+ partitions, and Linux lacks support for HFS+.
(So far; I believe it's being worked on, although I've not looked into
this recently.) It could also be that you're just trying the wrong
partition numbers.

Post by Keith Krehbiel
At least this time I
have emacs and am learning to use it. Several of my attempts to install
omitted emacs and that was one of the reasons to reinstall: to get a
tool to examine a few things.

That's overkill. If you omit a package during the initial install, you can
simply install that package later. I went back and skimmed your earlier
posts, but didn't see mention of what Linux distribution you're using, so
I can't provide specific instructions. In any event, it's usually easier
to fix an existing Linux installation than to re-install it.

Post by Keith Krehbiel
Discribe has an option to copy CD/DVD. I will try that next. It sounds
like what you were describing.

CD-R software "copy" options usually expect to copy from one physical
medium to another, rather than from an image file to CD-R. I'm not
familiar with Discribe specifically, though, so I can't promise that this
isn't where the image copy option is located.

[Keith Krehbiel]

I am using Debian from a set of disks from Edmunds Enterprises of
America. I am sure that I am not trying to access HFS+. It could very
well be that I am trying to access the wrong partition numbers. I have
had very limited success trying read the directory structure from within
linux. I know that part of it is ignorance on my part. I have seen in
threads that the error output can be directed to a file and the file
then can be read and posted but I have not been able to complete the
circle there. At this point I have not tried to install any sort of
internet access onto linux. I figured to try to get the os working
first. Maybe that is a mistake. Chalk it up to the simple mind.

snip>That's overkill. If you omit a package during the initial install,
you can
simply install that package later. <clip
I agree wholeheartedly that it is overkill. The reason I have found it
necessary is that the only time I can access the installation library is
from the initial installer and from the first boot following
installation. In subsequent sessions I cannot access anything that is
not on the boot partition and that includes the distribution CDs which I
am relatively certain are not HFS+. The error I get at that point is
that the file/directory I am trying to access does not exist. I have
even tried to access another linux hard drive and that fails to work. Go
figure. The first appears to be HFS or at least partially HFS and the
remaining disks appear to be ISO9660, at least from MacOS.
I tried the CD Copy function but, as you said, it expects a CD to be
read and then written to a blank. I did, however, burn a disk with the
image I had downloaded. I have not yet tried to read it from linux but
the image opens from MacOS to look like the original disk that I had
damaged. This is after opening the image. Opening the disk shows the
image, not the directory structure as on the original.
I tore down my system to move it out of the way after I installed the
last time. Upon reassembling it in a different location I have not been
able to boot again. This is a Mac problem not linux. I just have to
remember which spell to cast to get it going again.
Thank you for your help. I am sure that a lot of my trouble is caused
by plain ignorance on my part.
Keith

Post by Rod Smith

Post by Keith Krehbiel
I have, so far, not found a way to access disks other than the system
disk from inside linux. I am not sure if I have something wrong with
the system or if it is an operator malfunction.

You might try posting details of what you've tried and what error messages
you get under an appropriately-titled thread. My initial guess is that
you're trying to mount HFS+ partitions, and Linux lacks support for HFS+.
(So far; I believe it's being worked on, although I've not looked into
this recently.) It could also be that you're just trying the wrong
partition numbers.

Post by Keith Krehbiel
At least this time I
have emacs and am learning to use it. Several of my attempts to install
omitted emacs and that was one of the reasons to reinstall: to get a
tool to examine a few things.

That's overkill. If you omit a package during the initial install, you can
simply install that package later. I went back and skimmed your earlier
posts, but didn't see mention of what Linux distribution you're using, so
I can't provide specific instructions. In any event, it's usually easier
to fix an existing Linux installation than to re-install it.

Post by Keith Krehbiel
Discribe has an option to copy CD/DVD. I will try that next. It sounds
like what you were describing.

CD-R software "copy" options usually expect to copy from one physical
medium to another, rather than from an image file to CD-R. I'm not
familiar with Discribe specifically, though, so I can't promise that this
isn't where the image copy option is located.

[Bill Unruh]

Again you gave us very little information.
exactly what command did you run to try to mount the disks?
Exactly what is the content of the /etc/fstab file?

What happens when you try
mount -a
after you have logged on as root?
Give all output from that command.

What is the output when you type
df

What happens if you type
mount /mnt/cdrom
?

If that does not work, what happens when you type
mount /dev/hdc /mnt/cdrom

and then, if there is no error messages
ls /mnt/cdrom



Keith Krehbiel <***@pldi.net> writes:

]I am using Debian from a set of disks from Edmunds Enterprises of
]America. I am sure that I am not trying to access HFS+. It could very
]well be that I am trying to access the wrong partition numbers. I have
]had very limited success trying read the directory structure from within
]linux. I know that part of it is ignorance on my part. I have seen in
]threads that the error output can be directed to a file and the file
]then can be read and posted but I have not been able to complete the
]circle there. At this point I have not tried to install any sort of
]internet access onto linux. I figured to try to get the os working
]first. Maybe that is a mistake. Chalk it up to the simple mind.

]snip>That's overkill. If you omit a package during the initial install,
]you can
]simply install that package later. <clip
]I agree wholeheartedly that it is overkill. The reason I have found it
]necessary is that the only time I can access the installation library is
]from the initial installer and from the first boot following
]installation. In subsequent sessions I cannot access anything that is
]not on the boot partition and that includes the distribution CDs which I
]am relatively certain are not HFS+. The error I get at that point is
]that the file/directory I am trying to access does not exist. I have
]even tried to access another linux hard drive and that fails to work. Go
]figure. The first appears to be HFS or at least partially HFS and the
]remaining disks appear to be ISO9660, at least from MacOS.
]I tried the CD Copy function but, as you said, it expects a CD to be
]read and then written to a blank. I did, however, burn a disk with the
]image I had downloaded. I have not yet tried to read it from linux but
]the image opens from MacOS to look like the original disk that I had
]damaged. This is after opening the image. Opening the disk shows the
]image, not the directory structure as on the original.
]I tore down my system to move it out of the way after I installed the
]last time. Upon reassembling it in a different location I have not been
]able to boot again. This is a Mac problem not linux. I just have to
]remember which spell to cast to get it going again.
]Thank you for your help. I am sure that a lot of my trouble is caused
]by plain ignorance on my part.
]Keith

]Rod Smith wrote:
]> In article <***@corp.supernews.com>,
]> Keith Krehbiel <***@pldi.net> writes:
]>
]>>I have, so far, not found a way to access disks other than the system
]>>disk from inside linux. I am not sure if I have something wrong with
]>>the system or if it is an operator malfunction.
]>
]>
]> You might try posting details of what you've tried and what error messages
]> you get under an appropriately-titled thread. My initial guess is that
]> you're trying to mount HFS+ partitions, and Linux lacks support for HFS+.
]> (So far; I believe it's being worked on, although I've not looked into
]> this recently.) It could also be that you're just trying the wrong
]> partition numbers.
]>
]>
]>>At least this time I
]>>have emacs and am learning to use it. Several of my attempts to install
]>>omitted emacs and that was one of the reasons to reinstall: to get a
]>>tool to examine a few things.
]>
]>
]> That's overkill. If you omit a package during the initial install, you can
]> simply install that package later. I went back and skimmed your earlier
]> posts, but didn't see mention of what Linux distribution you're using, so
]> I can't provide specific instructions. In any event, it's usually easier
]> to fix an existing Linux installation than to re-install it.
]>
]>
]>>Discribe has an option to copy CD/DVD. I will try that next. It sounds
]>>like what you were describing.
]>
]>
]> CD-R software "copy" options usually expect to copy from one physical
]> medium to another, rather than from an image file to CD-R. I'm not
]> familiar with Discribe specifically, though, so I can't promise that this
]> isn't where the image copy option is located.
]>

[Keith Krehbiel]

Post by Bill Unruh
Again you gave us very little information.
exactly what command did you run to try to mount the disks?
Exactly what is the content of the /etc/fstab file?

I have not looked there yet, didn't know to look.

Post by Bill Unruh
What happens when you try
mount -a
after you have logged on as root?
Give all output from that command.

I have not tried that either. I hope you don't mean for me to type the
output verbatim. That is what it will require to give what you ask for
at this point.

Post by Bill Unruh
What is the output when you type
df

see above.

Post by Bill Unruh
What happens if you type
mount /mnt/cdrom
?

ditto

Post by Bill Unruh
If that does not work, what happens when you type
mount /dev/hdc /mnt/cdrom

same

Post by Bill Unruh
and then, if there is no error messages
ls /mnt/cdrom

ls usually returns no error but merely echoes what the pathlist I had typed.

I will try to get it all running again and do the things you suggested.
I think it is probably all good advice.
Thanks

Post by Bill Unruh
]I am using Debian from a set of disks from Edmunds Enterprises of
]America. I am sure that I am not trying to access HFS+. It could very
]well be that I am trying to access the wrong partition numbers. I have
]had very limited success trying read the directory structure from within
]linux. I know that part of it is ignorance on my part. I have seen in
]threads that the error output can be directed to a file and the file
]then can be read and posted but I have not been able to complete the
]circle there. At this point I have not tried to install any sort of
]internet access onto linux. I figured to try to get the os working
]first. Maybe that is a mistake. Chalk it up to the simple mind.
]snip>That's overkill. If you omit a package during the initial install,
]you can
]simply install that package later. <clip
]I agree wholeheartedly that it is overkill. The reason I have found it
]necessary is that the only time I can access the installation library is
]from the initial installer and from the first boot following
]installation. In subsequent sessions I cannot access anything that is
]not on the boot partition and that includes the distribution CDs which I
]am relatively certain are not HFS+. The error I get at that point is
]that the file/directory I am trying to access does not exist. I have
]even tried to access another linux hard drive and that fails to work. Go
]figure. The first appears to be HFS or at least partially HFS and the
]remaining disks appear to be ISO9660, at least from MacOS.
]I tried the CD Copy function but, as you said, it expects a CD to be
]read and then written to a blank. I did, however, burn a disk with the
]image I had downloaded. I have not yet tried to read it from linux but
]the image opens from MacOS to look like the original disk that I had
]damaged. This is after opening the image. Opening the disk shows the
]image, not the directory structure as on the original.
]I tore down my system to move it out of the way after I installed the
]last time. Upon reassembling it in a different location I have not been
]able to boot again. This is a Mac problem not linux. I just have to
]remember which spell to cast to get it going again.
]Thank you for your help. I am sure that a lot of my trouble is caused
]by plain ignorance on my part.
]Keith
]> ]>>I have, so far, not found a way to access disks other than the system
]>>disk from inside linux. I am not sure if I have something wrong with
]>>the system or if it is an operator malfunction.
]> ]> ]> You might try posting details of what you've tried and what error messages
]> you get under an appropriately-titled thread. My initial guess is that
]> you're trying to mount HFS+ partitions, and Linux lacks support for HFS+.
]> (So far; I believe it's being worked on, although I've not looked into
]> this recently.) It could also be that you're just trying the wrong
]> partition numbers.
]> ]> ]>>At least this time I
]>>have emacs and am learning to use it. Several of my attempts to install
]>>omitted emacs and that was one of the reasons to reinstall: to get a
]>>tool to examine a few things.
]> ]> ]> That's overkill. If you omit a package during the initial install, you can
]> simply install that package later. I went back and skimmed your earlier
]> posts, but didn't see mention of what Linux distribution you're using, so
]> I can't provide specific instructions. In any event, it's usually easier
]> to fix an existing Linux installation than to re-install it.
]> ]> ]>>Discribe has an option to copy CD/DVD. I will try that next. It sounds
]>>like what you were describing.
]> ]> ]> CD-R software "copy" options usually expect to copy from one physical
]> medium to another, rather than from an image file to CD-R. I'm not
]> familiar with Discribe specifically, though, so I can't promise that this
]> isn't where the image copy option is located.
]>

[I R A Darth Aggie]

On Mon, 01 Nov 2004 14:32:58 -0600,

+ > What happens when you try
+ > mount -a
+ > after you have logged on as root?
+ > Give all output from that command.
+
+ I have not tried that either. I hope you don't mean for me to type the
+ output verbatim. That is what it will require to give what you ask for
+ at this point.

No, that's why G*d invented redirection...

# mount -a > ~/mount-output

Copy and paste (or otherwise insert) the contents of the file
mount-output.

You might follow up the "mount -a" with the output of "fdisk -l"
(so we can see your filesystems):

# fdisk -l >> ~/mount-output
^^append to an existing file, create if necessary

Simply trim the resulting file, stick it in a post, and send it on.

James

[Keith Krehbiel]

Post by I R A Darth Aggie
On Mon, 01 Nov 2004 14:32:58 -0600,

+ > What happens when you try
+ > mount -a
+ > after you have logged on as root?
+ > Give all output from that command.
+
+ I have not tried that either. I hope you don't mean for me to type the
+ output verbatim. That is what it will require to give what you ask for
+ at this point.

No, that's why G*d invented redirection...
# mount -a > ~/mount-output
Copy and paste (or otherwise insert) the contents of the file
mount-output.
You might follow up the "mount -a" with the output of "fdisk -l"
# fdisk -l >> ~/mount-output
^^append to an existing file, create if necessary
Simply trim the resulting file, stick it in a post, and send it on.
James

Believe it or not, I actually kinda understand output redirection. I
guess with the current name of the thread it is a little confusing.
From the file I really have no place to paste. As things currently are
I am communicating from another computer running MacOS 9.1. One of the
several problems I am having is accessing disks/files/devices, etc. from
linux. Not sure if I can make text go to MacOS, maybe that would work.
Thanks for the idea and thank you all for your patience.
Keith

[I R A Darth Aggie]

On Mon, 01 Nov 2004 22:55:01 -0600,

+ From the file I really have no place to paste. As things currently are
+ I am communicating from another computer running MacOS 9.1. One of the
+ several problems I am having is accessing disks/files/devices, etc. from
+ linux.

Then the first thing you need to look at is the output of 'fdisk
-l'. It's been a *very*long*time* since I've looked at MacOS9.x, but I
think the odds are that you have an HFS filesystem. fdisk will give up
the pertinent information about the current partitions, and tell us
what it thinks your disk state is. So far as I know, if you have
HFS+, you're hosed.

If not, then it should be a matter of figuring out which device is the
correct partion for MacOS (more than a few kilobytes in size). Then:

mount -t hfs /dev/hfsfilesystem /somemounpoint

where "hfsfilesystem" is like "hda6" or "sda5", and "/somemountpoint"
is an arbritary empty directory. There may be a /mnt directory already
available. If not, make one. "mkdir /somemountpoint" for instance.

+ Not sure if I can make text go to MacOS, maybe that would work.

Fortunately, text is text, even if it looks a little odd due to the
niceties of the end-of-record delimiter.

James

[Keith Krehbiel]

Post by I R A Darth Aggie
On Mon, 01 Nov 2004 22:55:01 -0600,

+ From the file I really have no place to paste. As things currently are
+ I am communicating from another computer running MacOS 9.1. One of the
+ several problems I am having is accessing disks/files/devices, etc. from
+ linux.

Then the first thing you need to look at is the output of 'fdisk
-l'. It's been a *very*long*time* since I've looked at MacOS9.x, but I
think the odds are that you have an HFS filesystem. fdisk will give up
the pertinent information about the current partitions, and tell us
what it thinks your disk state is. So far as I know, if you have
HFS+, you're hosed.
If not, then it should be a matter of figuring out which device is the
mount -t hfs /dev/hfsfilesystem /somemounpoint
where "hfsfilesystem" is like "hda6" or "sda5", and "/somemountpoint"
is an arbritary empty directory. There may be a /mnt directory already
available. If not, make one. "mkdir /somemountpoint" for instance.

+ Not sure if I can make text go to MacOS, maybe that would work.

Fortunately, text is text, even if it looks a little odd due to the
niceties of the end-of-record delimiter.
James

Yes, text is indeed text. One of the problems that is particularly
vexing for me is an inability to access files of any kind especially if
they are located on a medium other than the boot partition. This is why
I have not been able to post the error output, etc. I am online with a
different computer than has the linux installation and have found NO WAY
to transfer files of any kind, including text. I am absolutely sure
that some or most of the problem is my ignorance but some of it may be
quirks in Linux.

I did try mount. mount -a /dev/cdrom0 returns the error that the device
is not found in /etc/fstab or /etc/mtab. mount -a /dev/hda6 returns the
same error and I am running linux from that partition. I will try the
other suggestions asap.

I know it must sound like I'm surly but it is from frustration, not
anger at anyone trying to help.

I am beginning to think that maybe what I need is a book to learn some
of this from. From what I have done so far it seems that there is some
variability in nomenclature of devices, etc. For instance, /dev/cdrom
would seem to be the cdrom drive but during bootup the term /sdr gets
used. I have used both terms to try to access the cdrom drive with
equal lack of success.
Anyway, thanks for all your help
Keith

[Rod Smith]

Post by Keith Krehbiel
I am online with a
different computer than has the linux installation and have found NO WAY
to transfer files of any kind, including text. I am absolutely sure
that some or most of the problem is my ignorance but some of it may be
quirks in Linux.

There are many ways to transfer text files (or binary files, for that
matter). What's easiest for you depends on specifics of your
configuration, your LAN (if you've got one), etc., and you've provided
very little in the way of information on this. Unfortunately, some of the
specifics (as others have requested) relate to information in key Linux
configuration files, which you can't transfer without first overcoming
this hurdle -- a chicken-and-egg problem. A couple of questions can help,
though:

1) Does the system you're trying to debug have a floppy disk? If so,
you could transfer files via it, using a few commands. (If it's a USB
floppy, this might be more complex, depending on whether your
distribution has correctly detected the USB floppy, though.)
2) Do you have access to a network file server, such as an FTP site? If
so, you could transfer the files to the FTP server from Linux and then
download them in another OS.

Post by Keith Krehbiel
I did try mount. mount -a /dev/cdrom0 returns the error that the device
is not found in /etc/fstab or /etc/mtab. mount -a /dev/hda6 returns the
same error and I am running linux from that partition. I will try the
other suggestions asap.

The "-a" parameter to mount tells the command to mount all the
filesystems defined in /etc/fstab; you don't normally use it in
conjunction with a filesystem specification (such as /dev/cdrom0 or
/dev/hda6). The /etc/fstab file, in turn, defines what Linux devices
should be mounted at what points, using what options.

Post by Keith Krehbiel
I am beginning to think that maybe what I need is a book to learn some
of this from.

That or a basic online tutorial. There's lots of Linux documentation at
http://www.tldp.org, but it's highly variable in quality -- some is quite
good, some is not so good or is outdated. I haven't read it in depth, but
glancing at the table of contents, the "Introduction to Linux" guide
(http://www.tldp.org/LDP/intro-linux/html/index.html) might be helpful.
I've got information on books (including several that I've written) at
http://www.rodsbooks.com/books/, but I've not really kept up with the
introductory market. Among those I've written or co-authored, _Linux
System Administration_ (with Vicki Stanfield) and _Linux+ Study Guide_
should be helpful, but you might want something that's a little shorter
and lower-level than these to get started, and I don't have any specific
recommendations in that arena.

Post by Keith Krehbiel
From what I have done so far it seems that there is some
variability in nomenclature of devices, etc. For instance, /dev/cdrom
would seem to be the cdrom drive but during bootup the term /sdr gets
used.

The CD-ROM is weird. SCSI CD-ROMs are usually /dev/scd0. ATAPI CD-ROMs are
identified much like hard disks -- usually /dev/hdc, but sometimes
/dev/hdb, /dev/hdd, or something else; and if SCSI emulation is in use,
they're /dev/scd0. I've seen a few other device names used on rare
occasion, but not very often. /dev/cdrom is a partial exception; it's a
symbolic link that points to the real device name, the idea being to
provide a standardized name for the system's primary CD-ROM device, no
matter what the "real" device filename is.

[Keith Krehbiel]

Post by Rod Smith

Post by Keith Krehbiel
I am online with a
different computer than has the linux installation and have found NO WAY
to transfer files of any kind, including text. I am absolutely sure
that some or most of the problem is my ignorance but some of it may be
quirks in Linux.

There are many ways to transfer text files (or binary files, for that
matter). What's easiest for you depends on specifics of your
configuration, your LAN (if you've got one), etc., and you've provided
very little in the way of information on this. Unfortunately, some of the
specifics (as others have requested) relate to information in key Linux
configuration files, which you can't transfer without first overcoming
this hurdle -- a chicken-and-egg problem. A couple of questions can help,
1) Does the system you're trying to debug have a floppy disk? If so,
you could transfer files via it, using a few commands. (If it's a USB
floppy, this might be more complex, depending on whether your
distribution has correctly detected the USB floppy, though.)
2) Do you have access to a network file server, such as an FTP site? If
so, you could transfer the files to the FTP server from Linux and then
download them in another OS.

Yes I have a floppy, and also a router that I can hook up to. That
still doesn't solve the problem of not being able to find any of those
devices from within the OS. You can see that my ignorance runs deep. I
have been reading some of the pertinant and not so pertinant man pages
on the drive. Slowly we learn. One question I have is: Is there a way
to copy a file? Can it be copied from one medium to another or must it
be duplicated first and then moved? I know that sounds stupid but it no
where says that up front. I think part of the disadvantage of starting
to use linux is that the programmers that have written it know all these
things and assume everyone else does too. Not trying to complain. It
does seem to work flawlessly when you know what to tell it.
I was able to finally get /dev/cdrom into some kind of a device list.
It now shows up when I 'df'. Dselect still will not find the files it
wants to see there however. I saw some mention of 'unmount' in some
manual page but apparantly I don't have it installed. Maybe it doesn't
exist in this version.
As always thanks for your patience and helpful responses.
Keith

Post by Rod Smith

Post by Keith Krehbiel
I did try mount. mount -a /dev/cdrom0 returns the error that the device
is not found in /etc/fstab or /etc/mtab. mount -a /dev/hda6 returns the
same error and I am running linux from that partition. I will try the
other suggestions asap.

The "-a" parameter to mount tells the command to mount all the
filesystems defined in /etc/fstab; you don't normally use it in
conjunction with a filesystem specification (such as /dev/cdrom0 or
/dev/hda6). The /etc/fstab file, in turn, defines what Linux devices
should be mounted at what points, using what options.

Post by Keith Krehbiel
I am beginning to think that maybe what I need is a book to learn some
of this from.

That or a basic online tutorial. There's lots of Linux documentation at
http://www.tldp.org, but it's highly variable in quality -- some is quite
good, some is not so good or is outdated. I haven't read it in depth, but
glancing at the table of contents, the "Introduction to Linux" guide
(http://www.tldp.org/LDP/intro-linux/html/index.html) might be helpful.
I've got information on books (including several that I've written) at
http://www.rodsbooks.com/books/, but I've not really kept up with the
introductory market. Among those I've written or co-authored, _Linux
System Administration_ (with Vicki Stanfield) and _Linux+ Study Guide_
should be helpful, but you might want something that's a little shorter
and lower-level than these to get started, and I don't have any specific
recommendations in that arena.

Post by Keith Krehbiel
From what I have done so far it seems that there is some
variability in nomenclature of devices, etc. For instance, /dev/cdrom
would seem to be the cdrom drive but during bootup the term /sdr gets
used.

The CD-ROM is weird. SCSI CD-ROMs are usually /dev/scd0. ATAPI CD-ROMs are
identified much like hard disks -- usually /dev/hdc, but sometimes
/dev/hdb, /dev/hdd, or something else; and if SCSI emulation is in use,
they're /dev/scd0. I've seen a few other device names used on rare
occasion, but not very often. /dev/cdrom is a partial exception; it's a
symbolic link that points to the real device name, the idea being to
provide a standardized name for the system's primary CD-ROM device, no
matter what the "real" device filename is.

[Bill Unruh]

Keith Krehbiel <***@pldi.net> writes:

]Rod Smith wrote:
]> In article <***@corp.supernews.com>,
]> Keith Krehbiel <***@pldi.net> writes:
]>
]>>I am online with a
]>>different computer than has the linux installation and have found NO WAY
]>>to transfer files of any kind, including text. I am absolutely sure
]>>that some or most of the problem is my ignorance but some of it may be
]>>quirks in Linux.

cp
copies files.

cp filea path/to/file/b

Transfer files to where? Tell us what you want to do.

Since millions transfer many many files using linux every day, yes it is
possible and is not due to any quirks in Linux.

]>
]Yes I have a floppy, and also a router that I can hook up to. That
]still doesn't solve the problem of not being able to find any of those
]devices from within the OS. You can see that my ignorance runs deep. I

What version of Linux are you running?

]I was able to finally get /dev/cdrom into some kind of a device list.
]It now shows up when I 'df'. Dselect still will not find the files it
]wants to see there however. I saw some mention of 'unmount' in some

umount, not unmount.

]manual page but apparantly I don't have it installed. Maybe it doesn't
]exist in this version.

[Keith Krehbiel]

]> ]>>I am online with a
]>>different computer than has the linux installation and have found NO WAY
]>>to transfer files of any kind, including text. I am absolutely sure
]>>that some or most of the problem is my ignorance but some of it may be
]>>quirks in Linux.
cp
copies files.
cp filea path/to/file/b
Transfer files to where? Tell us what you want to do.
Since millions transfer many many files using linux every day, yes it is
possible and is not due to any quirks in Linux.

Thanks, I knew there had to be a way.

]>
]Yes I have a floppy, and also a router that I can hook up to. That
]still doesn't solve the problem of not being able to find any of those
]devices from within the OS. You can see that my ignorance runs deep. I
What version of Linux are you running?

Debian Woody

]I was able to finally get /dev/cdrom into some kind of a device list.
]It now shows up when I 'df'. Dselect still will not find the files it
]wants to see there however. I saw some mention of 'unmount' in some

umount, not unmount.
Ahh, so that's what it is. Thanks again.

]manual pa

ge but apparantly I don't have it installed. Maybe it doesn't

]exist in this version.

[Rod Smith]

Post by Keith Krehbiel
Yes I have a floppy, and also a router that I can hook up to. That
still doesn't solve the problem of not being able to find any of those
devices from within the OS.

Try putting a pre-formatted FAT floppy disk in the floppy drive and then
issuing these commands AS ROOT:

mkdir -p /mnt/floppy
mount /dev/fd0 /mnt/floppy

(After the first time, you won't need the first command; it just creates
the /mnt/floppy directory, which is where the floppy disk is mounted --
that is, where it's grafted into the Linux directory tree.) The result
SHOULD BE that the floppy disk's contents appear at /mnt/floppy. If it's
got files, you should be able to see them with ls. You can then copy files
to the floppy:

cp /etc/fstab /mnt/floppy

When you've copied all the files you need, unmount the floppy:

umount /mnt/floppy

Note that "umount" has just one "n" in it.

If this procedure doesn't work, you could try the mtools package:

apt-get install mtools
mcopy /etc/fstab a:

The apt-get command installs the mtools package on Debian systems. (You
mentioned dselect later, so I assume you're using Debian or one of its
derivatives. Thus, apt-get should be available to you. This command may
ask you to insert a CD-ROM; or it might tell you that mtools is already
installed.) You must type the apt-get command as root. The mcopy command
is part of the mtools package, and it works very much like COPY under
DOS. It lets you move files between the hard disk and a FAT floppy
without actually mounting the floppy. You might or might not need to be
root to use mcopy, depending on your device permissions -- but that's an
fairly advanced topic, so if you try it as an ordinary user and it
doesn't work, just try it as root, at least for the moment.

[Keith Krehbiel]

Post by Rod Smith

Post by Keith Krehbiel
Yes I have a floppy, and also a router that I can hook up to. That
still doesn't solve the problem of not being able to find any of those
devices from within the OS.

Try putting a pre-formatted FAT floppy disk in the floppy drive and then
mkdir -p /mnt/floppy
mount /dev/fd0 /mnt/floppy
(After the first time, you won't need the first command; it just creates
the /mnt/floppy directory, which is where the floppy disk is mounted --
that is, where it's grafted into the Linux directory tree.) The result
SHOULD BE that the floppy disk's contents appear at /mnt/floppy. If it's
got files, you should be able to see them with ls. You can then copy files
cp /etc/fstab /mnt/floppy
umount /mnt/floppy
Note that "umount" has just one "n" in it.
apt-get install mtools
The apt-get command installs the mtools package on Debian systems. (You
mentioned dselect later, so I assume you're using Debian or one of its
derivatives. Thus, apt-get should be available to you. This command may
ask you to insert a CD-ROM; or it might tell you that mtools is already
installed.) You must type the apt-get command as root. The mcopy command
is part of the mtools package, and it works very much like COPY under
DOS. It lets you move files between the hard disk and a FAT floppy
without actually mounting the floppy. You might or might not need to be
root to use mcopy, depending on your device permissions -- but that's an
fairly advanced topic, so if you try it as an ordinary user and it
doesn't work, just try it as root, at least for the moment.

Thanks. That sounds like something to try.
I have Debian woody. Slowly I am learning my way around it. It goes
much better with the help of all on here.
Keith

[Keith Krehbiel]

Post by Keith Krehbiel

Post by Rod Smith

Post by Keith Krehbiel
Yes I have a floppy, and also a router that I can hook up to. That
still doesn't solve the problem of not being able to find any of
those devices from within the OS.

Try putting a pre-formatted FAT floppy disk in the floppy drive and then
mkdir -p /mnt/floppy
mount /dev/fd0 /mnt/floppy
(After the first time, you won't need the first command; it just creates
the /mnt/floppy directory, which is where the floppy disk is mounted --
that is, where it's grafted into the Linux directory tree.) The result
SHOULD BE that the floppy disk's contents appear at /mnt/floppy. If it's
got files, you should be able to see them with ls. You can then copy files
cp /etc/fstab /mnt/floppy
umount /mnt/floppy
Note that "umount" has just one "n" in it.
apt-get install mtools
The apt-get command installs the mtools package on Debian systems. (You
mentioned dselect later, so I assume you're using Debian or one of its
derivatives. Thus, apt-get should be available to you. This command may
ask you to insert a CD-ROM; or it might tell you that mtools is already
installed.) You must type the apt-get command as root. The mcopy command
is part of the mtools package, and it works very much like COPY under
DOS. It lets you move files between the hard disk and a FAT floppy
without actually mounting the floppy. You might or might not need to be
root to use mcopy, depending on your device permissions -- but that's an
fairly advanced topic, so if you try it as an ordinary user and it
doesn't work, just try it as root, at least for the moment.

Thanks. That sounds like something to try.
I have Debian woody. Slowly I am learning my way around it. It goes
much better with the help of all on here.
Keith

I haven't yet gotten a floppy to mount but I did get the hfs partition
on my Linux disk to mount and copied some files to it.
fstab is as follows:

# /etc/fstab: static file system information.
#
# <file system> <mount point> <type> <options> <dump> <pass>
/dev/hda6 / ext2 errors=remount-ro 0 1
/dev/hda7 none swap sw 0 0
proc /proc proc defaults 0 0
/dev/fd0 /floppy auto user,noauto 0 0
/dev/cdrom /cdrom iso9660 ro,user,noauto 0 0
Hmmm. That doesn't show hda5, my hfs partition.

mtab:

/dev/hda6 / ext2 rw,errors=remount-ro 0 0
proc /proc proc rw 0 0
devpts /dev/pts devpts rw,gid=5,mode=620 0 0
/dev/hda5 /mnt/hd hfs rw 0 0

Ahh! there it is.

I was having so much fun that I copied XftConfig too:

# $XFree86: xc/lib/Xft/XftConfig.cpp,v 1.5 2000/12/14 23:03:53 keithp Exp $

dir "/usr/X11R6/lib/X11/fonts/Type1"
dir "/usr/X11R6/lib/X11/fonts/TrueType"

#
# alias 'fixed' for 'mono'
#
match any family == "fixed" edit family =+ "mono";

#
# Check users config file
#
includeif "~/.xftconfig"

#
# Alias between XLFD families and font file family name, prefer local
# fonts
#
match any family == "charter" edit family += "bitstream charter";
match any family == "bitstream charter" edit family =+ "charter";

I tried to install mtools but that seems to want a desktop environment,
another thing that I have not yet gotten to work.

Thanks again for all your help
Keith

[Rod Smith]

Post by Keith Krehbiel
I haven't yet gotten a floppy to mount but I did get the hfs partition
on my Linux disk to mount and copied some files to it.
# /etc/fstab: static file system information.
#
# <file system> <mount point> <type> <options> <dump> <pass>
/dev/hda6 / ext2 errors=remount-ro 0 1
/dev/hda7 none swap sw 0 0
proc /proc proc defaults 0 0
/dev/fd0 /floppy auto user,noauto 0 0
/dev/cdrom /cdrom iso9660 ro,user,noauto 0 0

With this information, I can say that you SHOULD be able to mount a floppy
as either root or an ordinary user by typing:

mount /floppy

If that doesn't work, try reporting the error message, along with
information on the disk (was it FAT or HFS, for instance).

Post by Keith Krehbiel
Hmmm. That doesn't show hda5, my hfs partition.

Try adding this line:

/dev/hda5 /mnt/hd hfs defaults,umask=0 0 0

This will mount your HFS partition automatically whenever the computer
boots. Alternatively, you could replace "defaults" with "user,noauto" to
have the system not mount it automatically, but to enable ordinary users
to mount the partition. The last I checked (quite a while ago), Linux's
HFS support was good, but still posed some risk of disk corruption. Thus,
I'd recommend using it sparingly, unless this is a "scratch" partition
that doesn't hold any really vital data.

The "umask=0" part of the above configuration line gives all files 777
(rwxrwxrwx) permissions, which enables anybody to read or write files on
the partition. You can specify other umask values to give more
restrictive permissions, but you'd probably then want to add "uid=aaa",
where "aaa" is a UID number, and possibly "gid=bbb", where "bbb" is a GID
number, to assign ownership of all files to something appropriate.

Post by Keith Krehbiel
/dev/hda6 / ext2 rw,errors=remount-ro 0 0
proc /proc proc rw 0 0
devpts /dev/pts devpts rw,gid=5,mode=620 0 0
/dev/hda5 /mnt/hd hfs rw 0 0
Ahh! there it is.

/etc/mtab shows the currently mounted filesystems. /etc/fstab shows the
filesystems that are mounted automatically or that are defined for easier
manual mounting. You edit /etc/fstab to change how the system behaves,
but you don't normally edit /etc/mtab; it's maintained automatically by
the system.

Post by Keith Krehbiel
I tried to install mtools but that seems to want a desktop environment,
another thing that I have not yet gotten to work.

You could just let it go ahead and install the desktop environment. It
really shouldn't be necessary, though; I suspect this is an unnecessary
dependency -- or perhaps you're interpreting the name of some dependency
as a desktop environment when in fact it's something else.

[Keith Krehbiel]

Post by Rod Smith

Post by Keith Krehbiel
I haven't yet gotten a floppy to mount but I did get the hfs partition
on my Linux disk to mount and copied some files to it.
# /etc/fstab: static file system information.
#
# <file system> <mount point> <type> <options> <dump> <pass>
/dev/hda6 / ext2 errors=remount-ro 0 1
/dev/hda7 none swap sw 0 0
proc /proc proc defaults 0 0
/dev/fd0 /floppy auto user,noauto 0 0
/dev/cdrom /cdrom iso9660 ro,user,noauto 0 0

With this information, I can say that you SHOULD be able to mount a floppy
mount /floppy
If that doesn't work, try reporting the error message, along with
information on the disk (was it FAT or HFS, for instance).

Post by Keith Krehbiel
Hmmm. That doesn't show hda5, my hfs partition.

/dev/hda5 /mnt/hd hfs defaults,umask=0 0 0
This will mount your HFS partition automatically whenever the computer
boots. Alternatively, you could replace "defaults" with "user,noauto" to
have the system not mount it automatically, but to enable ordinary users
to mount the partition. The last I checked (quite a while ago), Linux's
HFS support was good, but still posed some risk of disk corruption. Thus,
I'd recommend using it sparingly, unless this is a "scratch" partition
that doesn't hold any really vital data.
The "umask=0" part of the above configuration line gives all files 777
(rwxrwxrwx) permissions, which enables anybody to read or write files on
the partition. You can specify other umask values to give more
restrictive permissions, but you'd probably then want to add "uid=aaa",
where "aaa" is a UID number, and possibly "gid=bbb", where "bbb" is a GID
number, to assign ownership of all files to something appropriate.

My reasons for having the hfs partition was to allow installation of
MacOS if I decided to do it that way and to provide space for file
exchanges such as these. I really don't intend to use it for any
important storage. So far no real work has been done from linux anyway.

Post by Rod Smith

Post by Keith Krehbiel
/dev/hda6 / ext2 rw,errors=remount-ro 0 0
proc /proc proc rw 0 0
devpts /dev/pts devpts rw,gid=5,mode=620 0 0
/dev/hda5 /mnt/hd hfs rw 0 0
Ahh! there it is.

/etc/mtab shows the currently mounted filesystems. /etc/fstab shows the
filesystems that are mounted automatically or that are defined for easier
manual mounting. You edit /etc/fstab to change how the system behaves,
but you don't normally edit /etc/mtab; it's maintained automatically by
the system.

Post by Keith Krehbiel
I tried to install mtools but that seems to want a desktop environment,
another thing that I have not yet gotten to work.

You could just let it go ahead and install the desktop environment. It
really shouldn't be necessary, though; I suspect this is an unnecessary
dependency -- or perhaps you're interpreting the name of some dependency
as a desktop environment when in fact it's something else.

I think I have most of Gnome installed but it fails to find my screen.
I am hoping to eventually work through those issues. In the course of
doing that I have ruined a couple installations of linux and had to
start over, hence the wear and tear on my installation cd that started
all this thread. Before I think I had parts of several desktops
installed at once and I suspect they were fighting among themselves.
This time I tried to limit it to Gnome since that seemed to match better
with Debian. It doesn't cry as loud on startup as it did before.
One of the things I discovered is that the /x directory in /pool on the
cd is one that is corrupted so that may have underlain some of my
problems with x-windows. I want to try it again when I can get linux to
access a healthy copy of the software.
Keith

[Rod Smith]

Post by Keith Krehbiel
I think I have most of Gnome installed but it fails to find my screen.

GNOME sits atop X, which is what directly accesses the video hardware. I
suspect you're having X problems, not GNOME problems. All of this is
unrelated to INSTALLATION issues for the mtools package, though, which is
what started this sub-thread -- you can INSTALL most packages without
actually USING them.

Post by Keith Krehbiel
I am hoping to eventually work through those issues. In the course of
doing that I have ruined a couple installations of linux and had to
start over, hence the wear and tear on my installation cd that started
all this thread.

As I said before, re-installing Linux is usually overkill. I *STRONGLY*
recommend that you look up and read some basic Linux documentation -- some
of the stuff at http://tldp.org, a basic Linux book, or something else. As
it is, you're groping around in the dark, with an occasional nudge from a
newsgroup post. Reading some basic documentation will be like turning on
the lights.

Post by Keith Krehbiel
Before I think I had parts of several desktops
installed at once and I suspect they were fighting among themselves.

Unlikely. You can INSTALL GNOME, KDE, XFce, and other desktop
environments on a single computer and the fact that they're installed on
one system won't cause problems. In fact, you can even run several of
them simultaneously on different virtual terminals, over network logins,
or whatnot, without problems. I'm not aware of any desktop environments
that will conflict with any other desktop environment, although sometimes
desktop environment CONFIGURATIONS from different distributions will
conflict with one another because they make assumptions about where files
are stored. If you're repeatedly re-installing a single distribution
(Debian), though, this shouldn't be a problem.

Post by Keith Krehbiel
One of the things I discovered is that the /x directory in /pool on the
cd is one that is corrupted so that may have underlain some of my
problems with x-windows.

That's entirely possible; if the installer couldn't read certain files,
that means that some packages probably weren't installed properly.

Incidentally, "x-windows" is incorrect terminology, albeit a commonly used
incorrect terminology. It's "The X Window System" or "X" for short.

[Keith Krehbiel]

Post by Rod Smith

Post by Keith Krehbiel
I think I have most of Gnome installed but it fails to find my screen.

GNOME sits atop X, which is what directly accesses the video hardware. I
suspect you're having X problems, not GNOME problems. All of this is
unrelated to INSTALLATION issues for the mtools package, though, which is
what started this sub-thread -- you can INSTALL most packages without
actually USING them.

Post by Keith Krehbiel
I am hoping to eventually work through those issues. In the course of
doing that I have ruined a couple installations of linux and had to
start over, hence the wear and tear on my installation cd that started
all this thread.

As I said before, re-installing Linux is usually overkill. I *STRONGLY*
recommend that you look up and read some basic Linux documentation -- some
of the stuff at http://tldp.org, a basic Linux book, or something else. As
it is, you're groping around in the dark, with an occasional nudge from a
newsgroup post. Reading some basic documentation will be like turning on
the lights.

Post by Keith Krehbiel
Before I think I had parts of several desktops
installed at once and I suspect they were fighting among themselves.

Unlikely. You can INSTALL GNOME, KDE, XFce, and other desktop
environments on a single computer and the fact that they're installed on
one system won't cause problems. In fact, you can even run several of
them simultaneously on different virtual terminals, over network logins,
or whatnot, without problems. I'm not aware of any desktop environments
that will conflict with any other desktop environment, although sometimes
desktop environment CONFIGURATIONS from different distributions will
conflict with one another because they make assumptions about where files
are stored. If you're repeatedly re-installing a single distribution
(Debian), though, this shouldn't be a problem.

Post by Keith Krehbiel
One of the things I discovered is that the /x directory in /pool on the
cd is one that is corrupted so that may have underlain some of my
problems with x-windows.

That's entirely possible; if the installer couldn't read certain files,
that means that some packages probably weren't installed properly.
Incidentally, "x-windows" is incorrect terminology, albeit a commonly used
incorrect terminology. It's "The X Window System" or "X" for short.

X it is then, we live and learn. There have been some threads on here
and elsewhere concerning the proper installation/configuration of X on
some Macs. I have read them all with interest since the symptoms seem
similar to the ones I experience. There have been config files posted
that are supposed to remedy the various ailments but none have been for
my specific machine (Umax C600). When Linux boots the way I have it now
it is much more civilized than it had been before. It used to try at
least 3 times to start X and finally give up. Now it calls it quits
after one try. Seems like an improvement to me: shorter boot time.
When X fails I get a message asking if I want to view the configuration
file. If I answer yes I get a blank window with 100% in the lower right
corner so I usually answer no. I have burnt a copy of the first cd but
have yet to try it.
One other dumb question. Is it possible to get on the internet without
a GUI? That is something I have never tried or even heard of really.

[Rod Smith]

Post by Keith Krehbiel
When Linux boots the way I have it now
it is much more civilized than it had been before. It used to try at
least 3 times to start X and finally give up. Now it calls it quits
after one try. Seems like an improvement to me: shorter boot time.

Debian starts X via a SysV startup script. No doubt this is Greek to you,
but it basically means that there's a script in /etc/init.d that starts
X, and a link to that script in /etc/rc3.d (or possibly /etc/rc5.d). The
link is called S99gdm, S99kdm, or S99xdm, depending on how your system is
configured. Delete it or rename it to K01gdm, K01kdm, or K01xdm
(depending on the original name) and X will no longer start. (Your system
may have all three files and links, so you may need to do this for all
three of them.) You can subsequently restore the original name to have X
start when the system boots, if you get X running.

After doing this, you can test X by typing "startx" at a command prompt. A
log file called /var/log/XFree86.0.log should appear (or maybe
/var/log/Xorg.0.log, if you're using X.org-X11 rather than XFree86).
That'll have hints to what's going wrong.

Post by Keith Krehbiel
One other dumb question. Is it possible to get on the internet without
a GUI? That is something I have never tried or even heard of really.

Yes, but precisely how you do this depends on your network connection
type -- PPP dial-up modem, ADSL, cable modem, etc. You might want to
check http://www.tldp.org/HOWTO/ISP-Connectivity.html and
http://www.tldp.org/HOWTO/PPP-HOWTO/index.html for information on PPP,
http://www.tldp.org/HOWTO/DSL-HOWTO/ for Linux DSL information, and
http://www.tldp.org/HOWTO/Cable-Modem/index.html for cable modem
information.

[Keith Krehbiel]

Post by Rod Smith

Post by Keith Krehbiel
When Linux boots the way I have it now
it is much more civilized than it had been before. It used to try at
least 3 times to start X and finally give up. Now it calls it quits
after one try. Seems like an improvement to me: shorter boot time.

Debian starts X via a SysV startup script. No doubt this is Greek to you,
but it basically means that there's a script in /etc/init.d that starts
X, and a link to that script in /etc/rc3.d (or possibly /etc/rc5.d). The
link is called S99gdm, S99kdm, or S99xdm, depending on how your system is
configured. Delete it or rename it to K01gdm, K01kdm, or K01xdm
(depending on the original name) and X will no longer start. (Your system
may have all three files and links, so you may need to do this for all
three of them.) You can subsequently restore the original name to have X
start when the system boots, if you get X running.
After doing this, you can test X by typing "startx" at a command prompt. A
log file called /var/log/XFree86.0.log should appear (or maybe
/var/log/Xorg.0.log, if you're using X.org-X11 rather than XFree86).
That'll have hints to what's going wrong.

Post by Keith Krehbiel
One other dumb question. Is it possible to get on the internet without
a GUI? That is something I have never tried or even heard of really.

Yes, but precisely how you do this depends on your network connection
type -- PPP dial-up modem, ADSL, cable modem, etc. You might want to
check http://www.tldp.org/HOWTO/ISP-Connectivity.html and
http://www.tldp.org/HOWTO/PPP-HOWTO/index.html for information on PPP,
http://www.tldp.org/HOWTO/DSL-HOWTO/ for Linux DSL information, and
http://www.tldp.org/HOWTO/Cable-Modem/index.html for cable modem
information.

One more dumb question: What sort of browser or other software should
be used? Will Mozilla work without X?

[Bill Unruh]

Keith Krehbiel <***@pldi.net> writes:

]Rod Smith wrote:
]> In article <***@corp.supernews.com>,
]> Keith Krehbiel <***@pldi.net> writes:
]>
]>>When Linux boots the way I have it now
]>>it is much more civilized than it had been before. It used to try at
]>>least 3 times to start X and finally give up. Now it calls it quits
]>>after one try. Seems like an improvement to me: shorter boot time.
]>
]>
]> Debian starts X via a SysV startup script. No doubt this is Greek to you,
]> but it basically means that there's a script in /etc/init.d that starts
]> X, and a link to that script in /etc/rc3.d (or possibly /etc/rc5.d). The
]> link is called S99gdm, S99kdm, or S99xdm, depending on how your system is
]> configured. Delete it or rename it to K01gdm, K01kdm, or K01xdm
]> (depending on the original name) and X will no longer start. (Your system
]> may have all three files and links, so you may need to do this for all
]> three of them.) You can subsequently restore the original name to have X
]> start when the system boots, if you get X running.
]>
]> After doing this, you can test X by typing "startx" at a command prompt. A
]> log file called /var/log/XFree86.0.log should appear (or maybe
]> /var/log/Xorg.0.log, if you're using X.org-X11 rather than XFree86).
]> That'll have hints to what's going wrong.
]>
]>
]>>One other dumb question. Is it possible to get on the internet without
]>>a GUI? That is something I have never tried or even heard of really.
]>
]>
]> Yes, but precisely how you do this depends on your network connection
]> type -- PPP dial-up modem, ADSL, cable modem, etc. You might want to
]> check http://www.tldp.org/HOWTO/ISP-Connectivity.html and
]> http://www.tldp.org/HOWTO/PPP-HOWTO/index.html for information on PPP,
]> http://www.tldp.org/HOWTO/DSL-HOWTO/ for Linux DSL information, and
]> http://www.tldp.org/HOWTO/Cable-Modem/index.html for cable modem
]> information.
]>
]One more dumb question: What sort of browser or other software should
]be used? Will Mozilla work without X?

No. It uses the gui. If you want to use a browser without X use lynx. Note
you will not be able to see any of the images-- just the text-- if you use
lynx.

[Keith Krehbiel]

]> ]>>When Linux boots the way I have it now
]>>it is much more civilized than it had been before. It used to try at
]>>least 3 times to start X and finally give up. Now it calls it quits
]>>after one try. Seems like an improvement to me: shorter boot time.
]> ]> ]> Debian starts X via a SysV startup script. No doubt this is Greek to you,
]> but it basically means that there's a script in /etc/init.d that starts
]> X, and a link to that script in /etc/rc3.d (or possibly /etc/rc5.d). The
]> link is called S99gdm, S99kdm, or S99xdm, depending on how your system is
]> configured. Delete it or rename it to K01gdm, K01kdm, or K01xdm
]> (depending on the original name) and X will no longer start. (Your system
]> may have all three files and links, so you may need to do this for all
]> three of them.) You can subsequently restore the original name to have X
]> start when the system boots, if you get X running.
]> ]> After doing this, you can test X by typing "startx" at a command prompt. A
]> log file called /var/log/XFree86.0.log should appear (or maybe
]> /var/log/Xorg.0.log, if you're using X.org-X11 rather than XFree86).
]> That'll have hints to what's going wrong.
]> ]> ]>>One other dumb question. Is it possible to get on the internet without
]>>a GUI? That is something I have never tried or even heard of really.
]> ]> ]> Yes, but precisely how you do this depends on your network connection
]> type -- PPP dial-up modem, ADSL, cable modem, etc. You might want to
]> check http://www.tldp.org/HOWTO/ISP-Connectivity.html and
]> http://www.tldp.org/HOWTO/PPP-HOWTO/index.html for information on PPP,
]> http://www.tldp.org/HOWTO/DSL-HOWTO/ for Linux DSL information, and
]> http://www.tldp.org/HOWTO/Cable-Modem/index.html for cable modem
]> information.
]> ]One more dumb question: What sort of browser or other software should
]be used? Will Mozilla work without X?
No. It uses the gui. If you want to use a browser without X use lynx. Note
you will not be able to see any of the images-- just the text-- if you use
lynx.

That's ok if the images don't show. At this point it would mostly be to
retrieve software. Thanks for the suggestion. I didn't know what to
look for
Keith

[Keith Krehbiel]

Debian starts X via a SysV startup script. No doubt this is Greek to you,
but it basically means that there's a script in /etc/init.d that starts
X, and a link to that script in /etc/rc3.d (or possibly /etc/rc5.d). The
link is called S99gdm, S99kdm, or S99xdm, depending on how your system is
configured. Delete it or rename it to K01gdm, K01kdm, or K01xdm
(depending on the original name) and X will no longer start. (Your system
may have all three files and links, so you may need to do this for all
three of them.) You can subsequently restore the original name to have X
start when the system boots, if you get X running.

After doing this, you can test X by typing "startx" at a command prompt. A
log file called /var/log/XFree86.0.log should appear (or maybe
/var/log/Xorg.0.log, if you're using X.org-X11 rather than XFree86).
That'll have hints to what's going wrong

That showed one thing wrong for sure. X couldn't be found by whatever
was looking for it after 'startx'. I couldn't find it either.
Still no luck getting a cd that linux will read. The last two burns I
put the file on the cd just as I got it from the net and then just
renaming it to something similar to the original. I think I have the /x
directory separatly intact if that will help and also XFree86. I am
pretty sure that is what is being used here.
Haven't tried to connect to the net yet. I need to study that a little
more.