Mac and iOS Bug Crashes Apps With a Single Indian-Language Character (mashable.com)

A lone Indian-language character is crashing a number of messaging apps on iOS, users are reporting. The problem also extends to the Apple Watch and even Macs, all of which struggle to process the character specific to the Telugu language spoken in India.

Comments

huh

[cascadingstylesheet]

And with all those Telugu-speaking programmers on staff too ... huh.

Re:huh

[cayenne8]

Geez.......

And I thought it was bad enough trying to understand them on support calls......

;)

Re: huh

[Anonymous Coward]

may be revenge in return of such BS comments in office ;)

Re:huh

[Anubis IV]

Ehh, the issue was fixed in the public betas before Mashable had even published their story about it.

For a better writeup, check out MacRumors' reporting on it (which was published prior to Mashable's). They mention that the bug was reported to Apple on Monday and was fixed at some point between then and now, so Apple has had a pretty quick turnaround on this one. Even so, it remains to be seen whether these fixes will be published in a minor update prior to their next intended release, or whether we'll have to wait for iOS 11.3 and macOS 10.13.4 before we get the fixes, which would likely be next month.

Meanwhile, Mashable's reporting is lagging. While they've found the time to update their article to indicate they finally managed to reproduce the issue, they apparently haven't had the time to update it with the fact that the bug was fixed before they ever wrote a word about it.

Re:huh

[Hal_Porter]

As soon as I heard of this I posted an message on Porter Industries internal message board in Telugu inviting everyone to an expensive restaurant for a free meal.

All who turned up were FIRED.

Re:huh

[Anubis IV]

Apple has now confirmed that the fixes they have in the public betas will be rolled out as minor releases for the current versions of iOS and macOS, rather than being kept for the next big release of each.

As for Mashable's reporting? Still lousy. They finally updated their article with the information it was missing at the time it was originally published, but then they went and added that "the fix will slowly roll out to older software", suggesting that the current versions won't get the fix until after the betas are released, even though the iMore article they linked as a source for that info says nothing of the sort. Quite the contrary, iMore's editor-in-chief is saying Apple will roll the fixes out before the betas are released.

Why is Slashdot using Mashable as a source when pretty much anyone else has more timely and accurate reporting, regardless of topic?

Re:huh

[AmiMoJo]

So what are your options if this bug caused a boot loop? If the character gets too your home screen somehow, say via home screen notification, you can't boot the phone to install updates.

https://twitter.com/MalwareTec...

Can you update via USB without wiping the phone?

Re:huh

[Anubis IV]

Can you update via USB without wiping the phone?

I can't say with certainty. I know it's possible to update via USB under normal circumstances, and I know it's possible to use DFU mode when you're stuck in a reboot loop, but DFU mode will leave you with a wiped phone, so far as I know. It might be possible to clear the notification in some other way, but I admit I don't know how that'd be done.

That said, for any normal users encountering this issue, they're likely to be fine. The default behavior once an iPhone user sets up an iCloud account is to enable iCloud Backups for their iPhone. In the few times I've used iCloud backups (e.g. switching my wife to a newer iPhone), it's had the phone back up and running in a matter of minutes. Most people seem to have it enabled, with geeks who prefer using encrypted iTunes backups being the only big exception I've seen.

Re:huh

[vaibhav.dlv]

Here in India you will find a new lingo called Hinglish - where Hindi speaking people type hindi text in English. Maybe there should be Telg-lish too!!! p.s. E.g. "How are you?" in hinglish is "kya haal hai?". In unicode etc it will look like - ?. But the hinglish text works as well. So similar thing in Telugu will be "Enna kudu varudu" or something. (I'm not a Telugu speaker - google translate shows , which I of course cannot read :J)

Re:huh

[Plumpaquatsch]

So what are your options if this bug caused a boot loop?

So what are your options if a bug in your Android CPU caused a boot loop?

Re: huh

[Anonymous Coward]

Typically, you can manually (or get someone who knows how) to flash the previous version back on.

Certain brands can download software that will do that for you without knowing much more than clicking the clearly labeled emergency button.

A UTF8 processing failure?

[Viol8]

In 2018? Apples quality control really is on the slide.

Either it can't process the utf8 code or its crashing when it doesn't have the font installed. Either way, these were solved problems 30 years ago, never mind now.

Re:A UTF8 processing failure?

[Penguinisto]

Solved 30 years ago? try 2017...

Re:A UTF8 processing failure?

[Anonymous Coward]

Seen similar problems in other systems too. You just like cherry picking.

Re:A UTF8 processing failure?

[sjames]

That's not a UTF processing fail, it's just a MS fail.

Re:A UTF8 processing failure?

[Anonymous Coward]

No process involving humans has a defect rate of 0. The important thing is that you fix the issues when they arise.

A years old Apple phone with a bug will be patched.

I can't be sure an Android phone will ever receive a patch at all.

Re:A UTF8 processing failure?

[aaarrrgggh]

um... this is /. after all. UTF is still in its infancy.

Re: A UTF8 processing failure?

[Anonymous Coward]

Name exactly one example...

Re: A UTF8 processing failure?

[Anonymous Coward]

https://github.com/sqlcipher/android-database-sqlcipher/issues/199

Re: A UTF8 processing failure?

[Hal_Porter]

A lot of embedded systems will behave strangely if you feed them a lot of characters like this

https://en.wiktionary.org/wiki...
http://www.unicode.org/cgi-bin...

That character is four bytes in UTF-8 which kills systems that assume a maximum of three - which used to be true for Chinese and Japanese, but isn't now.

It's also two UTF-16 code points, which will mess up systems that assume each character is a single code point.

Now you'll say "Those systems are all buggy". That's true now, but it wasn't true when a lot of them were designed - Unicode used to be limited to 64K characters which meant it was a fixed width encoding for UCS-2. And that three bytes was the maximum encoding for UTF-8.

When it grew those ceased to be true. Which is fine for systems that are maintained - the vendor would find bugs created by the standard change and push an update. Unfortunately a lot of systems - particularly embedded ones - aren't like that. Hell, Android isn't like that. Google push updates out to vendors but if your machine is EOL you're SOL.

Re:A UTF8 processing failure?

[TheFakeTimCook]

In 2018? Apples quality control really is on the slide.

Either it can't process the utf8 code or its crashing when it doesn't have the font installed. Either way, these were solved problems 30 years ago, never mind now.

Posted on Slashdot, where you don't dare even use the "wrong" quote character...

Re:A UTF8 processing failure?

[Hal_Porter]

Android sucks for patching because it's all up to the vendors who don't care about old machines - they want you to buy a new device.

However Apple sucks too - they force everyone to the latest version to get patches, and that version may run so slowly that you need to buy a new device.

Actually Google have a clever idea called Project Treble to solve the update problem

https://android-developers.goo...

The idea is that there's a stable vendor interface to the low level parts so the stuff above that can be swapped out. Bad news is that it will only be in Android O and later. So it won't do anything to fix all the ageing Android devices out there as they slip out of vendor's support window.

Re:A UTF8 processing failure?

[Megol]

30 years ago? UTF-8 didn't exist 30 years ago.

Re:A UTF8 processing failure?

[zifn4b]

In 2018? Apples quality control really is on the slide.

Either it can't process the utf8 code or its crashing when it doesn't have the font installed. Either way, these were solved problems 30 years ago, never mind now.

I don't think it was solved 30 years ago. That would have been 1988 when the C64, Apple IIe and Atari 800XL roamed the earth. But it was awhile ago. For the uninitiated, check this out.

Re: A UTF8 processing failure?

[Anonymous Coward]

/. Always messes up characters like ðY" and Ä

Re: A UTF8 processing failure?

[AmiMoJo]

Unicode is broken, and most Unicode apps are even more broken.

It's time we replaced Unicode. Make 32 bits the only encoding. Ditch all combinational characters. Separate out all merged languages. Create some solid libraries to handle it an convert UTF8/16.

With Unicode you can't even reliably tell how long a string is. Most software that claims to support it is buggy as hell. Programmers can't be expected to become language experts.

Re: A UTF8 processing failure?

[Hal_Porter]

We all need to adopt the Mojibake standard for non ASCII characters, like Slashdot.

Re: A UTF8 processing failure?

[Anonymous Coward]

Unicode is cancer and it's a subset of the cancer of political correctness. It tries to accomodate every language as being valid and equal and then some. It even includes Emoji and Klingon. Ridiculous!

Computing is inherently rooted in the English language, which is why UTF-8 exists in the first place - so all this Unicode cancer can work with mature ASCII systems.

32-bit character encodings still won't fix the problem because Unicode allows characters to be composed from multiple codepoints.

The band "Stormtroopers of Death" said it best with their album "Speak English or Die"

Re:A UTF8 processing failure?

[angel'o'sphere]

How exactly do you quality control random stuff no one is thinking about?

Re: A UTF8 processing failure?

[Ecuador]

It even includes Emoji and Klingon. Ridiculous!

No, it does not include Klingon - there was a proposal, but Klingon was rejected (people use the private area to encode them). In fact, it would be better if it was accepted, because that way it would mean that Unicode is at least fair in its all-inclusiveness, as apart from the various poo emojis, they have also accepted weird constructed alphabets (e.g. Deseret, Shavian) and even two of JRR Tolkien's fictional languages are scheduled to be added!
So it they accept any crap (even literally - remember poo emoji), but will balk at some like Klingon seemingly randomly.

Re:A UTF8 processing failure?

[Viol8]

You do sanity checks on values you're presenting to the display layer.

Re:A UTF8 processing failure?

[Plumpaquatsch]

Either it can't process the utf8 code or its crashing when it doesn't have the font installed. Either way, these were solved problems 30 years ago, never mind now.

Can't handle UTF-8? Now we now were the Slashcode developers went!

Re: A UTF8 processing failure?

[Anonymous Coward]

Unicode is cancer and it's a subset of the cancer of political correctness. It tries to accomodate every language as being valid and equal and then some. It even includes Emoji and Klingon. Ridiculous!

Still pissed it doesn't support alt-right-blather or jockistani?

Re:A UTF8 processing failure?

[lfourrier]

by fuzzing ( https://en.wikipedia.org/wiki/Fuzzing)

Re:A UTF8 processing failure?

[Anonymous Coward]

Either it can't process the utf8 code or its crashing when it doesn't have the font installed. Either way, these were solved problems 30 years ago, never mind now.

It's not a UTF-8 encoding issue. It's a Composite Combining Character issue. Apple has been bitten by this hard in the past, e.g.: when CCCs got split over soft line wraps created by the Messages app.

In Unicode, Telugu relies on many CCC combinations to produce various characters. This Apple bug wasn't affected by a single character but by a series of consonant-consonant-vowel combinations.

Apple probably never tested languages that use three or more CCCs because most western languages get by on single characters or two CCCs at most. At least Telugu, Devanagari and Kannada fall into this category.

Re: A UTF8 processing failure?

[mjwx]

A lot of embedded systems will behave strangely if you feed them a lot of characters like this

This is a good point but IOS and MacOS are general purpose OS's, not embedded systems. They should be expected to gracefully handle unexpected input. Even Windows can manage that, and has been managing it for well over a decade (almost 15 years, when was XP SP2 released).

Re: A UTF8 processing failure?

[Hal_Porter]

To their credit MS spent a lot of time on Unicode support. All Win32 functions have an A version which takes 8 bit chars and a W one which takes 16 bit ones. And Windows has moved smoothly from supporting UCS-2 for the W functions to UTF-16.

Re:A UTF8 processing failure?

[angel'o'sphere]

Why don't you read the article you linked? While 'fuzzing' would be possible, no one likely evver thought about it.

Who in his sane mind would think that a random UTF code can crash a software, even the whole computer?

So it brings us back to square one: why and how would you test something you never thought about?

Re:A UTF8 processing failure?

[angel'o'sphere]

No, you don't.
The display layer, displayes.
For what exactly would it need 'sanity' checks, and what exactly would be a sanity check in case of a unicode display?
Hu? Yes, this UTF-8 code is a valid UTF-8 code: please display it! Yes, this UTF-16 code is a valid UTF-16 code, please display it. Yes, this UTF-32 code is a valid UTF-32 code, please display it.
That would be a pointless excercise as all UTF-8/16/32 ccodes are valid codes, there is nothing to 'sanity' check!

Should we call tech support?

[Anonymous Coward]

Should be call tech support in India about this? I'm sure they have a very mindful and thorough list of steps to run through to a solution.

Re:Should we call tech support?

[Anonymous Coward]

this bug can be fixed thru updation.

Re: Should we call tech support?

[Anonymous Coward]

learn proper english and then call

Re:Should we call tech support?

[Anonymous Coward]

They'll call you. "This is Mac calling. Your computer have virus"

Re: Should we call tech support?

[Anonymous Coward]

How dare you! I have attend University -- I speak the perfect English!

plaintext FTW, eh?

[cellocgw]

Right on the heels of the article and discussion of plaintext vs. rich text vs. whateverthefuck Google is about to unleash on email comes this screwup.
I say it's high time that any and all text-ish messaging systems require just plain ASCII characters and if people insist on using alternative alphabets they dang well should be required to paste them in as images, not Unicode 84E0DDC2834A or however big the field is these days.

Re:plaintext FTW, eh?

[frank_adrian314159]

I think if you have room for a picture of turds in your character set, you have been given too much space.

Re:plaintext FTW, eh?

[Wycliffe]

Right on the heels of the article and discussion of plaintext vs. rich text vs. whateverthefuck Google is about to unleash on email comes this screwup.
I say it's high time that any and all text-ish messaging systems require just plain ASCII characters and if people insist on using alternative alphabets they dang well should be required to paste them in as images, not Unicode 84E0DDC2834A or however big the field is these days.

I agree that the unicode emojis are ridiculous but plain ascii barely works for english let along other languages. You could map other languages onto english characters but even that presents a problem because many languages have more than 26 letters in their alphabet. A two byte character code either DBCS or unicode makes sense but any system that implements it should have a default for all 65k characters including the ones that it doesn't know how to display. The most common way I'm used to seeing it is a little box icon with the 4 digits crammed inside it but because many of those 65k characters are still undefined it should definitely not crash on an undefined character.

Re:plaintext FTW, eh?

[Anonymous Coward]

The world should just adopt English. It has a very simple and efficient representation via ASCII and all the important countries are primarily English speaking countries.

Re:plaintext FTW, eh?

[ShanghaiBill]

I say it's high time that any and all text-ish messaging systems require just plain ASCII characters

Great idea. I totally agree that we should all adopt a single alphabet, but obviously the standard should be based on Chinese hanzi ideograms, not ASCII. Hanzi have a bigger user base, don't depend on a single spoken language, and are more compact since each character represents an entire syllable.

If everyone is in agreement, we can start working on the unification immediately.

Re:plaintext FTW, eh?

[ghoul]

As long as the standard characters are Telugu and you have to paste in English characters I am on board

Re:plaintext FTW, eh?

[mark-t]

There's nothing wrong with using either unicode or UTF8 in the 21st century, but if a particular subsystem isn't prepared to cope with input, it should reject any attempt to submit input it cannot properly handle. Even rendering unknown characters each in a small box with the unicode value written in it is an entirely reasonable coping mechanism. Rendering the wrong thing, or worse, crashing, is a result of either poor QA or lazy developers or both.

Re:plaintext FTW, eh?

[Anonymous Coward]

all the important countries are primarily English speaking countries.

In your dreams. Times have changed.

Re:plaintext FTW, eh?

[Jeremi]

My friends and I communicate exclusively in Telugu, you insensitive clod!

Re:plaintext FTW, eh?

[tlhIngan]

There's nothing wrong with using either unicode or UTF8 in the 21st century, but if a particular subsystem isn't prepared to cope with input, it should reject any attempt to submit input it cannot properly handle. Even rendering unknown characters each in a small box with the unicode value written in it is an entirely reasonable coping mechanism. Rendering the wrong thing, or worse, crashing, is a result of either poor QA or lazy developers or both.

The problem is Unicode is complex. A "character" can be composed of multiple codepoints, which is why they're named such. It's entirely possible to have a character take many kilobytes of space because someone decided to apply many decorations to it.

Or, even, for something composed of say, 5 codepoints to occupy more space on the screen than 6 codepoints because the language says iif A is followed by B, it is a Foo, but if it's A followed by B followed by C, it's a Bar. And bars character is less wide than a Foo character.

For more fun, check out the flag characters. They are designed that if the font lacks those characters, then it would appear in the text as the country code. So you could ask for the US flag, and if the font you chose didn't have it, you'd actually end up with "US" instead. But if you did have it, you'd get the US flag. The strange thing is, you have to realize you have to treat both as a single character for text purposes, so deleting one character doesn't leave you with a U and half a flag, but removes it all.

Squares work only in the simple renderer that treats almost all codepoints as individual characters, but these days that's far too simplistic.

And trolls love making 100K worth of codepoints to become a huge black blob 1000 pixels high, 1000 pixels wide or more that gets rendered as a huge black blob on your application or web page.

Re:plaintext FTW, eh?

[Anonymous Coward]

Are we europeans supposed to stop using our diacritics and convert to The Only True language, coming from our American overlords? Are the Chinese supposed to cut themselves from the Internet because it does't want their languages? Why do you think Unicode has been invented in the first place?

Re:plaintext FTW, eh?

[TheFakeTimCook]

There's nothing wrong with using either unicode or UTF8 in the 21st century, but if a particular subsystem isn't prepared to cope with input, it should reject any attempt to submit input it cannot properly handle. Even rendering unknown characters each in a small box with the unicode value written in it is an entirely reasonable coping mechanism. Rendering the wrong thing, or worse, crashing, is a result of either poor QA or lazy developers or both.

Sez the person who has never coded a modern Operating System, that has to operate flawlessly WORLDWIDE, with a plethora of languages, fonts, even writing direction.

And as I said earlier, Slashdot can't even handle an alternate "Quote" character without displaying gibberish. So I don't want to hear it...

Re:plaintext FTW, eh?

[mark-t]

The problem is Unicode is complex. A "character" can be composed of multiple codepoints, which is why they're named such. It's entirely possible to have a character take many kilobytes of space because someone decided to apply many decorations to it.

A sane way to properly handle that would be to reject code sequences that the system cannot cope with before attempting to display something it cannot handle, or in the case of overly long codepoint sequences, display any excess codepoints beyond what it is designed to cope with as individual glyphs that indicate which decorations could not be rendered as part of the one character. Each codepoint is, after all, only 4 bytes long at worst, so there's no reason that a system can't render each individual one as its own character as a fallback, perhaps, as I said before, for any codepoints it is not going to handle, showing a rectgangular glyph with the hex code for the codepoint inside of it. So when you have 100k worth of codepoints that is not going to be handled, all you get as an output is 100k tiny litle glyphs, showing each individual codepolnt that is not being rendered correctly.

It's far from ideal, but it does not crash, and even better, clearly communicates to the end user where the problems with the input are.

And even then, this does not preclude being able to sanely cope with codepoint sequences that are below some predetermined length, and show them appropriately.

Re:plaintext FTW, eh?

[geek]

I'm switching to smoke signals. Fucking Natives had it right all along.

Re:plaintext FTW, eh?

[Megol]

I propose using Futhark as it's easier to carve into stone making it future proof to the world after WW III.

Re:plaintext FTW, eh?

[T.E.D.]

I'll be back to get started on this after I'm done learning to alphabetize all of CNS 11643's 48,027 glyphs....

Re:plaintext FTW, eh?

[Megol]

I think that illustrates well the point that Unicode is a badly designed mess of a standard. It isn't a replacement for ASCII or other character encoding but something that is so complex to fully support that there are very little software that do.

A better design would IMO be something that divides characters and (necessarily) complexity in 2 or 3 layers so that simple software can handle basic rendering and software that support more complex scripts use the other layers to properly fuse/split text according to the language selected.
And the language is important as there are languages that use the same type of characters but handle them differently - something Unicode can't handle without even more complication.

Instead we get those freaking emojis which aren't well defined in the first place... Insane.

Sorry for ranting.

Re:plaintext FTW, eh?

[Anonymous Coward]

lol its cute that you think the under 40 crowd wants to read when pictures are available. remember, they voted an orangutan into office.

Re:plaintext FTW, eh?

[Anonymous Coward]

Hanzi have a bigger user base, don't depend on a single spoken language, and are more compact since each character represents an entire syllable.

I'd recommend Katakana or similar as they were basically designed precisely to syllable foreign words. More importantly, the actual set that needs learned is very small which would actual further adoption. More, there's enough double meaning in Hanzi and general confusion about pronunciation, it seems like a bad idea to rely upon it. Really, the only thing going for it is its wide adoption. To that end, there's more people who use the Latin alphabet, but that doesn't mean I'd recommend it.

Re:plaintext FTW, eh?

[Solandri]

All of this has happened before. All of this will happen again. The world is returning to ideograms for an alphabet. But instead of calling them characters, we're calling them emojis.

Re: plaintext FTW, eh?

[ArmoredDragon]

In other words, it's not an alphabet at all, rather it's a syllabary.

Re:plaintext FTW, eh?

[Bert64]

Well the language is actually from england, which is part of europe, and the americans adopted it just fine.

See! It just works!

[Chas]

Used to be, for crashing an app, you'd have to put in a bit of actual EFFORT.

Now, these kids today...one character.

POOF!

Computers should be English only!

[Anonymous Coward]

They aren't meant to use other languages. Extended ASCII is as far as it should go. If your language doesn't fit, well, I guess you're screwed.

'Lone' character

[mccalli]

Hmm. I don't know of course, but it sounds more likely that this character will be one of a class of characters which could trigger the bug. Just the most used one. I doubt there's anything unique about that particular character.

Somone found Apple's Brown Noise

[js290]

http://southpark.cc.com/clips/...

You'd think

[bobstreo]

In this day and age, running all possible unicode characters in testing would not be an afterthought.

Talk about things that would be trivial to automate....

Re:You'd think

[Wycliffe]

Unicode has a ton of unallocated character space and frequently adds new ones so testing all possible characters isn't really practical or desirable. What should be tested on the other hand is support for unknown characters. You have defined characters and an associated bitmap and then undefined characters that need to have a default. That default can be either a static image or another way I've seen it is the unicode number crammed inside a small box. Either way, testing for a single valid character and a single invalid character should be sufficient if all the characters are treated the same way as they should be.

Re:You'd think

[Anonymous Coward]

Either way, testing for a single valid character and a single invalid character should be sufficient if all the characters are treated the same way as they should be.

I see you aren't familiar with UTF8. All characters cannot be treated identically, because UTF8 doesn't use a constant character size. Some are 1 byte, some are 2 bytes, some are 4 bytes. Off the top of my head I can't recall if some are three bytes, but it wouldn't surprise me if they are. So you need byte continuation markers similar to varints and such. That ensures that the symbol 'a' cannot be treated identically to 'poopEmoji'.

I'm not defending Apple here, just pointing out that your comment is rather naive.

Re:You'd think

[Jeremi]

Unicode has a ton of unallocated character space and frequently adds new ones so testing all possible characters isn't really practical or desirable.

The theoretical maximum number of unicode characters is "only" 2^32, and in practice the limit is set at 1,111,998 characters. While it might not be practical to test that they all render correctly, it wouldn't be difficult to set up an automated test that at least verifies that they don't cause the renderer to crash when used.

Re: You'd think

[Anonymous Coward]

You donâ(TM)t actually have to formally render themonto a real screen. Just an offscreen bitmap. Do all the work without a screen. This is generally called unit testing, and not only testing 100% of valid characters (that is, those YOU consider valid) you should also throw a bunch of junk at the interpreter to fuzz it.

At some point, random garbage WILL be dumped into the interpreter whether intentionally or not, and itâ(TM)s simply NOT ACCEPTABLE to crash on user-specified input. It can never be trusted to be in any way sane, to sanity-checking is job #1, here.

LOL

[Anonymous Coward]

and apple claims to be a global company. What a joke and embarrassment they are.

ASCII ASCII ASCII

[RightwingNutjob]

If it's not encoded in ASCII, I won't bother reading it.

Re:ASCII ASCII ASCII

[Anonymous Coward]

You certainly won't be reading it here on Slashdot (tm)!

Re:ASCII ASCII ASCII

[PPH]

Welcome to Slashdot.

Good one Apple!

[Anonymous Coward]

Great job Apple! Add all these additional STUPID emojis but can't handle unicode text. Nothing like priorities.

Shame it wasn’t Kannada script

[Anonymous Coward]

Then you could give a literal look of disapproval to iOS users.

How does this happen?

[Galaga88]

Can anybody with experience in programming unicode handling explain how a bug like this happens? It seems weird that a specific character could trigger a crash like this - what is being handled at such an individual level?

Re:How does this happen?

[Wycliffe]

Can anybody with experience in programming unicode handling explain how a bug like this happens? It seems weird that a specific character could trigger a crash like this - what is being handled at such an individual level?

The only way this would make sense to me is either they didn't code for an undefined character or if instead of storing bitmaps for the font they are storing vector graphics for each character and there was something wrong with their renderer code.

Re:How does this happen?

[PPH]

Developer doesn't provide for a branch in the code to handle a case that they think won't ever happen. Or even a catch-all branch that just prints a '?' and carries on. The exception is carried back up the app uncaught until it crashes.

Re:How does this happen?

[jrumney]

It's an Indian character. They need to store a lot more than bitmaps or vectors for each character. The font itself contains processing code, and the font rendering involves much more than just blting a bitmap or tracing some vectors.

Sanitizing input is old fashioned

[innocent_white_lamb]

Sanitizing input is old fashioned and not brave, I guess.

Sheesh. Any programmer with more than 6 months of experience should know these things.

Er, shouldn't they? Or am I living in a world that's somehow passed me by. Again...

Sigh...

Re:Sanitizing input is old fashioned

[DickBreath]

What programmers may already know how to do, and what managers allow them to spend time doing are two different things.

Don't blame programmers when management is an adequate explanation for the problem. Managers are the cause of most problems in the world. They know this, and so they try to CYA on everything. It is because of their managers. It's managers all the way down, until you get into the infernal nether regions.

And, they have the value hierarchy inverted. The "higher" up the org chart you go, the lower you are actually going down the value hierarchy.

No problem on other OSs

[sdsri]

No problems with Telugu characters on Ubuntu and Windows for several years now. Apple is just now starting to support this language?

Re:No problem on other OSs

[Anonymous Coward]

The summary talks about "a number of messaging apps", which translates as: there are a lot of garbage messaging apps in the App store that were built with Apple's SDK, and there are some weak libraries in there that are now causing common problems. Inexperienced developers, poor project management, and Apple's corporate priorities are all to blame.

Re: No problem on other OSs

[Anonymous Coward]

The whole point of OSs and FooKits and stuff like that is that an application doesnâ(TM)t have to provide its own bundled libraries for memory management, networking, graphics, sound, and font rendering.

So no, this is not the fault of âoegarbage apps,â it is Appleâ(TM)s fault for providing inadequate support for Unicode. Worst case should be that the character is garbled (hello, /.!), replaced by a placeholder (?) or simply missing. Crashing is not acceptable.

Please start posting this character on slashdot

[Anonymous Coward]

Please Please Please start posting this character on every slashdot story. Lets rid ourselves of all these mac/ios isheep and their fucked up character posts.

Re:Please start posting this character on slashdot

[fluffernutter]

Some of us use a mac because we have to. I don't like it, but I grin and bear it.

Re:Please start posting this character on slashdot

[Anonymous Coward]

I feel sorry for you. You must work for a millennial startup. Best of luck to you.

Re:Please start posting this character on slashdot

[Megane]

Unicode? On my slashdot? You must be new here?

Re:Please start posting this character on slashdot

[Plumpaquatsch]

Please Please Please start posting this character on every slashdot story. Lets rid ourselves of all these mac/ios isheep and their fucked up character posts.

Apple's evil plan with this "bug" is actually to force Slashdot to support more UTF chars!

Re:See! It just works!

[PolygamousRanchKid]

Now, these kids today...one character.

POOF!

Note that this character was developed by the Indian military industry in a project with similar goals as the "Killer Joke": https://en.wikipedia.org/wiki/...

The intent is that people reading the character die.

Joke warfare is officially banned by the Geneva Convention, but India's archenemy Pakistan has been working on their own form of joke warfare involving a schoolgirl and a Nobel Peace Prize, and India felt threatened.

I bet all the emoji...

[Anonymous Coward]

... are working just fine though.
Who needs languages when we have the pile of turd unicode point?

Common UTF-16 problem

[Anonymous Coward]

When dealing with 16-bit values, you must always ask: Was this Indian character it big-endian or little-endian?

I love isheep

[Anonymous Coward]

eat me sheep!!

How does something like this even happen?

[Qbertino]

Yes, I know, characters have been a mess ever since back in the day. Glyphs are difficult and academic debates over what constitutes an own seperate glyph are probably never going to end ... but how on earth does something like this even happen? Could someone explain?

I don't get it. So the bitrange for UTF8 is widened and processing chars at international scale is a tad more resource hungry - I get it. But crashing a system with a character? What is this? It's not that we're back in punchcard era with killpokes that set systems on fire. How super-bad does one have to screw up for this even to happen? In iOS. ... I seriously can't fathom this.

This is an honest question - maybe someone can explain?

Re:How does something like this even happen?

[Anonymous Coward]

Think about the classic quoting problem: I have a string of information and I want to include it in a message, but need a way to delimit where that string starts and ends. But the character I choose as a delimiter might exist in the string, so the first thing to do is replace that delimiter with an "escaped" delimiter, and of course now you have to escape the escape delimiter as well but eventually you will have a string that you can put in your message. On the receiving end, you have to remember that the message might contain some escaped sequences which must be restored before doing anything else with the message. The problem is that strings can be included in messages and messages can be included as strings in other messages arbitrarily, so there is no way to simply look at a string and determine if is has been properly decoded for the current context. At some point those strings get fed into a state engine, and certain obscure combinations of characters might look like an instruction to screw the pooch, which is the final result.

just works

[Anonymous Coward]

I thought Apple just works? Something must be wrong with the character.

Re:See! It just works!

[ghoul]

You sir just won the internets!!!!

I work for a fortune 500

[Anonymous Coward]

And just hid this character in our home page. Cause fuck you isheep!!!!

To quote Nelson....

[Anonymous Coward]

Haa-Ha!

Fake news!

[fluffernutter]

My mac pro is awesome! Fake news! I don't believe this for a second! Whatthehell does this character look like?? There I found it in goog

Which character

[110010001000]

Which character is it? I want to try it on mys

Re:Which character

[Anonymous Coward]

https://pastebin.com/9Tr8ytTr

Killer Glyph

[T.E.D.]

Back in the day we used to have killer pokes, that would do things like cause the electron gun to stop sweeping and put a hole in your monitor with a single BASIC ("POKE") instruction. The assembly equivalent of course is the Halt-and-Catch Fire instruction (eg: 0xDD on a M6800).

Nice to see kids with their fancy new fonts can get in on the act with Killer Glyphs.

Re:Killer Glyph

[SuiteSisterMary]

Nice to see kids with their fancy new fonts can get in on the act with Killer Glyphs.

Power Word: Crash?

Shady Indian Characters

[IhateMonkeys]

We had our apps crash once because of an Indian character. Pretty sure his name was "Steve" or Rajesh, not Telugu.

Re:Shady Indian Characters

[nnet]

You mean it was Peggy.

What to call this character?

[Provocateur]

A lone Indian-language character Shall we call this "Tonto"? No,not the Jonny Depp movie