Free SparkFS

TBAFS – remind me what it does

It is an image filing system the front end of which offers the following settings:

if that is of any help.

There isn’t a Help file.

needs to be downloaded with a real world browser.

Or to remember the Google Drive hack…

“Open” the link. It, obviously, won’t work with NetSurf but you’ll get a load of rubbish in the URL bar.

Delete all the junk after the file ID. That’s the “/view?usp=drive_open&hl=en_GB” bit.

Replace the “/file/d/” at the beginning of the path with “uc?export=download&id=”.

That is to say, change this:
https://drive.google.com/file/d/0B1cBDcf_WpHLN2Y1Yjc2NmUtMjc4Ny00NWZhLThkMGItMGQ0ODBlNDQ3YTc3/view?usp=drive_open&hl=en_GB

To this:
https://drive.google.com/uc?export=download&id=0B1cBDcf_WpHLN2Y1Yjc2NmUtMjc4Ny00NWZhLThkMGItMGQ0ODBlNDQ3YTc3

Then hit Enter and NetSurf will fetch it.
(or cheat and click the convenient link above ;-) )

needs to be downloaded with a real world browser.

Or to remember the Google Drive hack

Do note that it generates an error message in the page even when using a totally up-to-date PC browser¹, but anyone familiar with Google bits and bobs would expect no less.

¹ I believe the current state of affairs is that Mozilla leads and the others play catch up.

I think that access to archive contents should be transparent – don’t know if (Richard W above) that would be more confusing or less for users from other systems.

Something transparent, yes. Also the funny business about the icon on the left of the icon bar. Should it really be there? Maybe the Filer should have some sort of ‘New file XYZ…’ option? (a bit like Windows does). ArchiveFS could hook into that and add ‘Zip archive’, or ‘Spark archive’, or ‘Tape archive’ etc.?

Maybe, like OmniClient or CloudFS, it could offer hooks for different algorithms to be added? Someone keen could add a TBAFS algorithm, for example.

I’m only thinking out loud here. Obviously, you do whatever you want, and I’m free to be ignored. I don’t want to be the guy who lectures someone on what they should do with their hobby! If I had particularly strong feelings, I’d submit a PR one day. :)

I tried using TBAFS with no compression as a transferable archive and it was very fast. Unfortunately it was not reliable. Without much documentation and apparently no likely development I did not pursue this. I found the one reliable image filing system was StrongHelp, but that is not really a candidate for all sorts of reasons. ;-)

Also the funny business about the icon on the left of the icon bar. Should it really be there?

I believe the logic is signalled in the name:- FS —> File System

I think you will find that ArcFS drops into a similar location on the iconbar

Richard – I agree – if there is an archive FS which everyone has then it can be better integrated into the OS. SparkFS is modular anyone can write modules for new archive types.

Hmm, clearing cookies first allows This old chromebook to stay logged in.
Re Archiving,
I think everyone is agreed on a non compression method being desirable for several reasons, maybe even as the default for the system.
Further considerations may be readability from another platform, and partial recovery if disc corruption in one area where the archive is stored.
Tar would be cross-platform, but could it be used interactively the way Spark works, maybe I mean here ‘as an imagefs’?

Edit: I guess file types are always going to cause problems when writing on other platforms, unless they can leave existing files untouched by appending changes.
Maybe a port of the RISC OS archiver to allow reading would be sufficient for most purposes.

SparkFS has a tar module – read only – but making it writeable would be trivial. I can’t recall what happened with filetypes – I think they do work on tar, even beyond the the thing with sticking the type after a comma in the file name. Tar does not have the problem with writing a directory after all the updates are done – but it does have the problem of having to scan the entire file on loading. Tar is good for recovery – its just a string of files with easy to spot headers between them.

Cookies – one has to visit the home page before doing anything else for the day – login – then go to the forum. Dunno what newbies seeking help do.

Cookies – one has to visit the home page before doing anything else for the day – login – then go to the forum. Dunno what newbies seeking help do.

I use NetSurf but have a tiny app (RunImage 19 lines of BASIC) which quits NetSurf, manipulates the Choices file to remove anything ROOL related, then presents an HTML log-in page with my credentials already entered – a form someone else devised and posted. So I click Log-in, NetSurf loads, and there we are!

People have scoffed – but a coconut every time!

I’ve found a very long thread about the cookie issue – I understand, I’m now sorry I mentioned it.

https://www.riscosopen.com/forum/forums/11/topics/14494?page=1

SparkFS has a tar module – read only

No, it’s read/write. I used to use it way-back-when for website updates from a computer at a local library, back when updates were a matter of “upload a tar file, untar it”.

but making it writeable would be trivial.

Apparently so, it’s worked for me for donkeys’.

I can’t recall what happened with filetypes – I think they do work on tar,

Not really in a sort-of-weirdly-maybe-not way.
You can apply all the file types you want, and they’ll stick for as long as the tar is open. However since there’s no modification of the tar header (I’m not sure there legitimately can be) all of the types are forgotten when the archive is closed. All files revert to ‘Data’ type when the tar is reopened.

Cookies – one has to visit the home page before doing anything else for the day

Not necessarily. I go to /forum/posts first, because there’s no point logging in if there’s no need to. The solution is easy:

• Firefox with Cookie Manager (Rob Wu)
  Go to login page when you want. Call up Cookie Manager. Show the cookies for the current tab. Select them, tap delete, confirm it. Then close Cookie Manager, back to the login tab, can log in without a problem.
  Works on Firefox desktop, tablet, and phone.
• NetSurf
  Go to login page when you want. Call up the menu. Utilities → Cookies → Show cookies. Choose the ROOL entry and press Delete. Then log in. It’s even quicker to do with NetSurf than Firefox. :-)
  Note: Don’t “Delete all cookies”, I find that crashes NetSurf, but maybe that just my system?

Dunno what newbies seeking help do.

That’s a concern, yes. There probably ought to be some instructions in the wiki that are linked to from the login page. Because you’re pretty much stuck in a loop where you can never log in unless you spot messages referring to the nuke-the-cookies trick and know how to do it.

Just looked at Android’s Chrome. Wow, they really don’t want you messing with cookies do they? It also looks like it is extremely easy to delete all site data in addition to the cookies, so you’ll throw away saved login information. So, uh… another reason to not use Chrome. As if the horrible forced text resizing wasn’t enough. :-/

All files revert to ‘Data’ type when the tar is reopened.

That’s with Archive Type/Method set to Tar/Unix. If you use Tar/Arctar the filetypes are preserved.

I use Chrome (on Win, Linux) and I had a struggle to find how to delete individual cookies.

Glad you’ve sorted out tar files – read the manual eh – great software to have covered all the problems so well.

Seriously if I had remembered this, I might have used tar files for my data back up activities.

With the read only archives, it lets you update the catalogue in memory, hopefully useful, but someone could argue it should immediately raise “read only errors”.