Skip to contentSkip to conversation

paid work: vfat symlinks, attributes and special files

DISCUSSION Tomas M 2 participants 5 messages Jun 18, 2007 – Jun 20, 2007 VFS

lore
Share & tools
lore

Tomas M

author
#0 lore raw

Good morning.
I'd like to kindly offer a paid job to someone interested. I'm willing to pay $500 for it, nevertheless I'm open to all suggestions. I hope it's OK to post this offer here.

I'd like to see some features implemented in VFAT filesystem, which would be available per a mount option, for example:
mount -t vfat -o unix /dev/hda1 /mnt

There has been many attempts to implement some of these features in vfat lately, while no implementation actually made it's way to Kernel. Umsdos was removed from kernel as well (which was a good idea indeed).

I agree that people shouldn't use VFAT at all :) as there are many filesystems for Linux which are more suitable for day-to-day work. But the amount of attempts to enhance vfat shows that many people really need some (even limited) improvements to vfat for some purpose.

So here is what I need:

- special file support (symlink,pipe,fifo,socket,c devs,b devs)
- executable / not executable bit
Nothing more is needed at all.

I believe there is some spare space in the VFAT structure to implement these easily. I believe we just need one single byte, I would use lcase char, while the other infos (eg. major/minor numbers, link target, etc..) would be stored inside the particular special file (so not in the metadata). But your implementation is up to you, of course.

Requirements:

- The code must be clean so it could possibly be included in Kernel, nevertheless it's not required to be accepted by kernel people - they sometimes tend to reject interesting stuff, this is something I can live with :)

- The filesystem must be perfectly readable in Windows.

- All files created in Windows (or in Linux without the mount option) should appear as a regular non-executable files in Linux after the mount option is used.

Payment:

As noted at the top of this message, I'm willing to pay $500 to the developer of these changes. I don't know how easy/hard would it be do implement, so if you say why the price should be bigger, I'll reconsider it. I believe it's a fair offer.

Payments are possible by paypal, wire transfer or money in the envelope.

If you are willing to develop requested changes, let me know please, I'll wait till 19th of January for answers and then I'll post a message to the list who will do it (if somebody is found).

Thank you very much

Tomas M
slax.org

#1 lore raw
Quoted 21 lines · Tomas M

Good morning. I'd like to kindly offer a paid job to someone interested. I'm willing to pay $500 for it, nevertheless I'm open to all suggestions. I hope it's OK to post this offer here.

I'd like to see some features implemented in VFAT filesystem, which would be available per a mount option, for example:

mount -t vfat -o unix /dev/hda1 /mnt

There has been many attempts to implement some of these features in vfat lately, while no implementation actually made it's way to Kernel. Umsdos was removed from kernel as well (which was a good idea indeed).

I agree that people shouldn't use VFAT at all :) as there are many filesystems for Linux which are more suitable for day-to-day work. But the amount of attempts to enhance vfat shows that many people really need some (even limited) improvements to vfat for some purpose.

Quoted 5 lines

So here is what I need:

- special file support (symlink,pipe,fifo,socket,c devs,b devs)
- executable / not executable bit
Nothing more is needed at all.

Does a FUSE fs cut it? Trying to hide Unix fluff inside VFAT structures is an evil thing (“What Would Windows Do?”), so the UMSDOS approach seems fine enough. I've almost got a working implementation, just needs actual testing.

Quoted 27 lines

I believe there is some spare space in the VFAT structure to implement these easily. I believe we just need one single byte, I would use lcase char, while the other infos (eg. major/minor numbers, link target, etc..) would be stored inside the particular special file (so not in the metadata). But your implementation is up to you, of course.

Requirements:

- The code must be clean so it could possibly be included in Kernel, nevertheless it's not required to be accepted by kernel people - they sometimes tend to reject interesting stuff, this is something I can live with :)

- The filesystem must be perfectly readable in Windows.

- All files created in Windows (or in Linux without the mount option) should appear as a regular non-executable files in Linux after the mount option is used.

Payment:

As noted at the top of this message, I'm willing to pay $500 to the developer of these changes. I don't know how easy/hard would it be do implement, so if you say why the price should be bigger, I'll reconsider it. I believe it's a fair offer.

Quoted 4 lines

Payments are possible by paypal, wire transfer or money in the envelope.

If you are willing to develop requested changes, let me know please, I'll wait till 19th of January for answers and then I'll

                         ^^^^^^^^^^^^^^^
Quoted 3 lines

post a message to the list who will do it (if somebody is found).

Thank you very much

Please don't name my name.

Thanks,
Jan

#2 lore raw
Quoted 8 lines · Jan Engelhardt
Quoted 2 lines

If you are willing to develop requested changes, let me know please, I'll wait till 19th of January for answers and then I'll

                        ^^^^^^^^^^^^^^^
Quoted 1 line

post a message to the list who will do it (if somebody is found).

Please don't name my name.

(Today, I'm hating my MUA for cc'ing even though I said "Y" to "reply to all?")

Well, stick to it anyway.

Jan

author
#3 lore raw
In reply to Jan Engelhardt
Quoted 2 lines
Quoted 1 line

I'll wait till 19th of January for answers and then I'll

                         ^^^^^^^^^^^^^^^

Oh that is incorrect, it should say '19th of Jun' of course :)

Quoted 1 line

Please don't name my name.

Ok, no problem :)

Regarding the unix fluff inside VFAT, I agree this is something bad, maybe very bad, but from time to time, some people really need it.

I need it so I'm willing to pay some little money for it, others don't have to use it, of course :)

Regarding the FUSE ... I think it may work for me as well, but is there any code for FUSE to implement what I need for VFAT? I searched google but didn't find anything.

Thank you

Tomas M

author
#4 lore raw
In reply to Tomas M
Quoted 4 lines

If you are willing to develop requested changes,
let me know please, I'll wait till 19th of January for answers and then I'll post a message to the list who will do it (if somebody is found).

So it seems nobody is interested.
Never mind, I give up. Thank you anyway :)

Tomas M