rant: filenames

Nov 08, 2021 Last reply: 3 years ago 88 Replies

Yeah, you can have 512 files named Panteltie using different cases. Very advanced, how stupid people have been to stick to the Latin alphabet for millenia.

Yeah, good thing you figured out an exception to cling to, see my comment above.

Then the naming is not the only shortcoming of the unix filesystem, it (like that of windows) makes worst fit allocation impractical. The latter can be fixed by writing/adopting a well designed filesystem; the naming flaw cannot be fixed.

On a sunny day (Mon, 15 Nov 2021 19:03:50 +0200) it happened Dimiter_Popoff snipped-for-privacy@tgi-sci.com wrote in <smu3tn$8jl$ snipped-for-privacy@dont-email.me:

Cool!

Read:

formatting link
list files ignoring case

You really need to read up on this stuff if you want to use it. I have NEVER encoutered a problem with Linux file systems, and tried many and use many.

But MS windows crap filesystem as it comes with some USB sticks or is used by my Chinese digital TV receivers even limit file size to 4293402624 bytes. 4293402624 Jul 8 2020 /mnt/sda2/video/satellite/magnum_3x_part_1.ts 41383680 Jul 8 2020 /mnt/sda2/video/satellite/magnum_3x_part_2.ts

-> 2^32 = 4294967296

You are just blaming the car that you do not know how to drive.

Now, now Dimiter, just because you can does not mean it's a good idea to do so. But I would resent any filesystem *preventing* me from doing so. Nothing in UNIX prevents you from naming your all your files in all upper case, should you want to. Choose your evil.

Jeroen Belleman

You don't have to convince me a solution can be programmed to just about any flaw, including the file naming in unix.

Oh I am sure it works, in its flawed by design way. Not using "worst fit" for allocation brings just more fragmentation than there has to be which you probably have not even noticed - or did not know there was life without it. And you have learned to live with the flawed use of text.

I never intended to compare these two filesystems. The MS approach with file names is the correct one though, and the unix' use of text for file names is the flawed one, that much is obvious.

Yeah, what do I know about filesystems. How many filesystems have you designed.

I know you cannot switch off from fuming mode - your entering of which I predicted - so I suggest you just let it go, you don't really know what you are talking about.

Have a nice evening.

On a sunny day (Mon, 15 Nov 2021 20:03:14 +0200) it happened Dimiter_Popoff snipped-for-privacy@tgi-sci.com wrote in <smu7d3$emg$ snipped-for-privacy@dont-email.me:

I am not in fuming mode, just trying to teach you, but you seemingly just post insults to get a response as you feel so lonely (your own words in previous posts) Maybe get a cat or dog ? Or go to bar and try it there?

So you almost qualify for the filter now.

File names are text entities and as such they are for human consumption. And how humans use text has been pretty well established for a long time. The Unix filesystem breaks these rules. Indeed you can use only upper or only lower case to work around that so you don't inadvertently access one of two very similar - in both name and contents - files; but then you lose the use of capitalization which humans are used to. So the correct way of naming files is obviously to preserve the case information while treating the names uniquely regardless of case. It even costs nearly nothing in terms of CPU power for search if data are organized in a practical way. E.g. in dps the file names in a directory entry are stored all upper (or was it lower, I did it years ago) case and the case information is stored separately, a bit per character. So dps can search in a machine friendly way - just comparing the text part - and humans can read the name with cases as it is. Even case dependent search does not cost more than a single longword compare more for every 32 characters of the name.

Insults? Teach me? How to use the ali-express toys you are tinkering with, sure.

Feel free to filter me, I won't respond any further to your fuming posts anyway.

perhaps: JW series relays from Panasonic/Nais

something downloaded from "alldatasheet.com"

some problem with your PDF reader.

I think I have a copy here.

Some are created from individual pages, others are created as a single, long file. If the file isn't protected, you should be able to select page or continuous display.

And that's the default for MacOS, which is after all just a Unix. Mind you I turn it off, because I prefer the Unix way.

However, with Unicode, it's worse. Do you want to consider comparable all accented and non-accented characters? How about composed vs decomposed characters? Should all filenames suffer UTF-8 normalisation?

It's a slippery slope.

Clifford Heath.

Fuming? Why would anyone want a case-insensitive file system?

There's nothing special about ASCII after all--if you like case insensitivity, how about allowing all Unicode glyphs that look alike to sort the same way too?

No thanks.

Cheers

Phil Hobbs

You prefer it out of habit or for some other reason? (My habits are case-independent for naming and the case-dependent unix style has been a nuisance to me, far from the worst nuisance I have had really but still). Same for my symbol naming habits when programming.

It is slippery slope indeed and I have not considered other alphabets than Latin and Cyrillic. However having a separate bit per byte of name (9 bits per character) - the way file names are stored in a dps directory entry - allows case change for any byte. Now what happens when the characters are two byte etc. is up to the higher (than search) level, I have neither done that nor thought of it. And how it goes with the Chinese hieroglyphs... I am not sure I even want to think of that. But I believe text should remain text as we have had it for ages, I can't see a need to have case dependent naming. Or passwords (rant mode on), I really hate when they enforce that on me. I can make up unique enough words case independently and I memorize

*words* (rant mode off).

Why do you need the case sensitivity? To me, it is normal to say copy *.jpg and this to copy all .jpg and all .JPG files, for instance.

Why use both extensions? Pick one and stick with it. If you have some good reason to use both, you can always go

cp $(find . -iname "*jpg") destdir

or something equivalent. If you need do that a lot, a 1- or 2-line script will automate it for you.

No biggie either way.h

Cheers

Phil Hobbs

you are asking for 0x6a,0x70,0x67 and 0x4a,0x50,0x47 to be the same

To Rob: Are you booted from a live disk? Those generally try to be polite and not let you scribble on your HDD by accident. If you know what you're doing, you can use sudo to do whatever you like.

Cheers

Phil Hobbs

Not at all. I am asking .JPG and .jpg to be treated the same. How you encode these is up to a very small proportion of the human population, it has been to me for example

[Here is how it goes: .jpg becomes 04 2e 6a 70 67 20 20 20 00 00 00 00 .JPG becomes 04 2e 6a 70 67 20 20 20 38 00 00 00

(in case you wonder about the excessive bytes - they are to keep things longword aligned). ]

Humans read text, not bytes. File names (any names, really) are text.

if you change it to jpg instead of JPG...

j is not the same as J

j is not the same as J indeed. Now please read out loud John john JOHN JOhn and listen to what came out.

Join the Discussion

Have something to add? Share your thoughts — no account required.

Didn't find your answer?

Ask the community — no account required