Comment by TacticalCoder | Hacker News Reader

Comment by TacticalCoder | Hacker News Reader

Define "space". Is the Hangul filler we talked about yesterday a spacing character? Is the zero-width non-breaking space a spacing character? What about the typographic spacing characters?

You should better be very afraid of using spaces in filenames.

You should do everything you can to support them but you have to know you'll invariably encounter countless cases where you'll have this or that tool that won't work properly with them.

I still live in a world where I cannot name a song from the french group L'impératrice with an eacute in the filename or my car's media system will display garbage (it's running QNX and I don't know which filesystem).

FWIW, and it should be food for thought, every single Git repository in the world contains a pre-commit hook sample (disabled by default but it's there) that enforces that every committed file in the repo is named using a subset of ASCII characters.

Every Git repository in the world has that example: let that sink in.

> FWIW, and it should be food for thought, every single Git repository in the world contains a pre-commit hook sample (disabled by default but it's there) that enforces that every committed file in the repo is named using a subset of ASCII characters.

I use Git for documents too, not only code. Why shouldn't I use my native language?

Tab completion don’t work well for languages that require IME. That is one reason why I don’t.

That's actually a good point. On the other hand, not all languages use IMEs. Mine just uses the AltGr modifier key, but is otherwise just a standard QWERTY layout without any features.

IME == Input method editor?

https://en.wikipedia.org/wiki/Input_method

yup, I type in pronunciation and let it guess what I'm trying to say. Works okay in editors but don't work great with shells in a terminal emulator, so I just prefer not having to use it in shell operations.

Tab completion works just fine for me with a Japanese IME.

non-ascii characters cause annoying hard to fix problems. If you're willing to deal with that - kudos. Personally I don't find it worthwhile

I haven't had problems yet. Spaces, punctuation, and quotes are the main offenders, most of the time.

> I still live in a world where I cannot name a song from the french group L'impératrice with an eacute in the filename or my car's media system will display garbage (it's running QNX and I don't know which filesystem).

I have an Android phone and I tell MusicBrainz Picard to save all files with ASCII-only names and Windows-compatible names for the ones that get sent over to the phone. Basically for this reason. Sometimes it's players on Android itself, but even more frequently, whatever bluetooth radio I'm connected to freaking out with non-ASCII characters.

What do you mean, display garbage?

L'imp?ratrice? L'imp�ratrice? L'impÃ©ratrice? L'imp‚ratrice? L'impÚratrice?

You get all those space characters working and then some jerk comes along and uploads a file like this: ŗ̶̧̢͓̳͍͙͔̳̻̥͉̭͓̫̟͍̞̭͉͓͉̮̹͍͚̳̹̬͉͚̰͈̘̐̊̾̈̀̒͒̀͛̓̋̔͊̏͘̚ę̴̨̛̣͙̤̟̬̩̟͙͖̥̹̱̱̊͑͗̇̇͛̆̈́̃͋̓̀̔̍̍̌̐͊̎̓̅̀̕ͅģ̴̹̜̘͍̱̑͐̉̌̐̄̊͛̎́̐̌̅̈́͂͑̈́̋̔͂̊̊̒̒̔͛͆̚͘̕͠e̶̙͕̫̳̘͐̾́̑͆̓͂̿͊̊̍͛͐̌̆͗̌̅̅̔͊̂͛͗̅̕͝͝͝͝x̵̢̧̦̫͖̝̥̹͓̬͖̤̩͚̝̫̋̃̅̈́̆͋̌͑́̎̈́̊̾͒̀̒̎̓͛͊̿̓͊̀̍͐̆̚͝͝-̴̨̮̯͖͖̠̜̲̪͕̘͈͖̮̈́̓̐̃́̅̄̏́̍̉̐͌́̔̓̄͋͗̐̕͜͝ţ̴̢̧̖̗͖̞̮̫̦̼̝̺̼̱̳͓͉̜̟̤̲͖̻͙́̌̈̌̈͆̾̄͊̿̏̓͗̈́̕͜ͅh̶̢̧̨̥̭̼̟̣͖̯̗̤̖̙͉͕̙͎̰̠̝̖͈̻͙̪̮̘̯̻̼͕͓̖̣͈̽́͊̎͐͌̆̍̎̏̿͐̒́͋͑̍̿̎͆̑͆̄͂̀͐̄͑̀͗̿̽̎̾̊̕͝͝͝͝͝ͅi̴͚͈͍̫̮̝̣͖͉͓̯̠̙̭̟̖̘̾̓̄̈́̒̏̽̆̉̿͛̀́̃̋̒̈́͋̂̇̈́͛̕͜͠͠͝ͅs̶͇̖̳̞͉̱̞͓̖͔͔͍̗͇̖̮̹̅͊̔͋͊̈́̎̐̆̋̒̀̍̕͜ͅ.̴̧͎͇̰͉̼̱̰̦̟̑̋̏͌̍͊͑̄̀͌́̆̓͛̒̆̾̉͐̄̂̈́͆̒̃͗̐̂̎̈́̈͛̿́͛̾̚͘͜͝͝ͅȩ̷̡̲̪̱̪̥̳͍̼̰̘̗̹͙͙͓̣̟̩̥̥̖̠̪̮̹̞̥̻͎͖͍̯̂͑̏̑̆̍͋̎͛̅̑̑̏̎̓̀̓̒̈́͊͌̀̈́̒̌͐͂͛̊̍̐͂́̔̌̾͐̈́̋̇̏̚͜͝͝͝͠ͅx̶̧̛͚̗̜̪͍͖̘̙͎͚͇͙̬̱̟̭͓̺̙͍̖̱͚̣̘̪̭͔͔̮͎̬̪̤̹̟͔̩͍̬͕͔̩͐̈́̒̂͛̂̈̀̿̍̔̓̓̀̃̍͆̈́̍̓̌͐̈́̾̇̎̑͌͒̄̆̿̍͆̅͗͆͘͠͝͝ͅͅͅe̷̢̡̡̨̧̛͕͚̬̮̞̥̼͍͔̝̟̝̯͈̟̥͖̱̹̣̩̼̩̅̌͌̑̎̐̀̽̏́͐̋̏̎̎͛͌̀̊͊͒̑͌̎̎̑͊̌̉͆̾̚͘̚͜͠͠͠͝͝ͅͅͅ

Honest question - what the heck are those characters?

Zalgo text: https://zalgo.org/

It was a great joke for a couple weeks two internets ago.

> two internets ago

It's been like three internets since I heard someone using "internet" as a measurement of time.

It's actually interesting to think about "generations" of internet, just like generations of people, and how the culture shifted between them.

There was a time in the early '00s when broadband was catching on, yet YouTube didn't exist. A time when Ebaumsworld and Newgrounds ruled the internet. When Homestar Runner was pop internet culture. Weebls Stuff. The frog blender.

Combining diacritic marks.

It corrupted text or "Zalgo" text, it relies on diacritics.

See this answer on stackoverflow:

https://stackoverflow.com/questions/1732348/regex-match-open...

I disagree with calling it "corrupted." We're not tricking the browser into trying to render garbage bytes that are actually the middle of a jpeg or something. It's actually valid Unicode. It's an edge-case which is not seen in regular usage, but it's technically following all of the rules.

It's just the way it's called, not a statement of fact.

In digital typography, combining characters are characters that are intended to modify other characters. The most common combining characters in the Latin script are the combining diacritical marks (including combining accents).

https://en.wikipedia.org/wiki/Combining_character

Specifically _Vietnamese_ combining characters. The Vietnamese writing system uses multiple combining characters at a time, and stacks them vertically. Throw in a few that wrap around the character like t҉his, some 𝑎lternatᵉ lꬲttꬲr fᵒrms, disturbing imagery, and perhaps a few other tricks, and you have zalgo. See also https://stackoverflow.com/a/1732454/823846

This legitimately made me laugh out loud in my office.

The characters reach up off the screen as I reply to this. They overlay the comment above you. Amazing. How?

It's usually called Zalgo text, and it's what you get when you start stacking all kinds of Unicode diacritics on poor unsuspecting characters.

https://en.wikipedia.org/wiki/Zalgo_text

Interestingly I get different behaviour per browser/OS. Firefox/Linux clips it to the bounding box of the parent element, Firefox/Mac and Safari/Mac clip it to the line height, and only Chrome/Mac lets it extended further.

Huh, I tried it in Chrome to see how it reacted here and it maintained about the same position as it did in my usual browser, Firefox.

Firefox and Safari on iOS 15 both render all the glyphs attached to the base character. Vivaldi, Chrome and Firefox on Win10 all render them stacked and overlapping the parent and child comments.

This is the best generator I found: https://lingojam.com/GlitchTextGenerator

I find http://animalswithinanimals.com/generator/generator.html much more controllable.

For anyone who is curious (and acolytes of Zalgo): "In Unicode, character rendering does not use a simple character cell model where each glyph fits into a box with given height. Combining marks may be rendered above, below, or inside a base character. So you can easily construct a character sequence, consisting of a base character and “combining above” marks, of any length, to reach any desired visual height, assuming that the rendering software conforms to the Unicode rendering model."

[https://stackoverflow.com/questions/6579844/how-does-zalgo-t...]

Hah, lucky for me Chrome on Ubuntu didn't implement the spec correctly. ;)

768 characters is too long for macOS it seems. (References online say HFS+ has a limit of 255 UTF-16 characters. Didn't find anything for APFS immediately... edit: same for APFS)

Glad you didn’t choose a sequence that crashes my browser.

regex this, bravo

Please don't Zalgo on HN. It's enough to speak its name.

It would be one thing if it was making other comments difficult to read or causing browser issues, but I appreciated the demonstration that both would presumably be possible on certain browsers

Until now, I haven't actually thought of what would happen if zalgotext occurred anywhere other than a web browser. Looking forward to the five minutes of fun with the file manager and whatnot.

Reply on news.ycombinator.com