Apple's obvious policy—filing for basically whatever view that whatever employer is voluntary to survey, ever—gives us an continuous activity of hypothesis cereal, many newsworthy, many not. Take this patent, which calls for flash microstorage in headsets and earbuds.

The obvious describes the communion of style settings, preferences and even media between iPods and iPhones by instrumentation of a kind of voguish receiver, which stores personalizations in a small stock of flash mental faculty. As delineated, it sounds a bit oddish, and its end isn't entirely clear: least of the functions square measure things that fall nether the horizon of syncing, which is something that least iPod/iPhone users do regularly anyway, and which they have to do when purchase a new style anyway, mooting the "it makes transitioning easier" angle.

But for certain settings, it makes many meaning: magnitude and EQ settings, for natural event, could be unbroken the equivalent between inclination, ensuring a render hearing feel. It'd be a small, flyspeck have that least dwell wouldn't notice, but it's immoderate from farfetched. [Unwired View]

In Why Isn't My Secret writing.. Encrypting? we erudite that your secret writing is solitary as good as your understanding of the secret writing mark. And that the C. H. Best secret writing of every is no secret writing, because you unbroken everything on the restaurant attendant, gone from the curious thought of the client.

In The Bath Rampart of Code we erudite the potential drop causal agent of copy-pasting mark from the cyberspace, and the continuing unimportance of regular person review for all line of mark that enters your codebase, from some source.

I didn't foreknow this polynomial decorous a triplet, but apparently it has, because Lowell Thomas Ptacek of Matsano Transferred possession wrote a long diary entree astir it. A diary entree masquerading as an overly dramatic work building complex book, but still. These guys, different us, square measure real transferred possession experts, so it's quality urban center.

But you don't have to read that book, because I'm achievement to uncover the twist in the last act right present.

1. The root difficulty wasn't unsatisfactory to realise the encryption.
2. The root difficulty wasn't text and pasting mark from the internet.
3. The root difficulty wasn't unsatisfactory to person review the code.

Mr. Ptacek is absolutely right. The root difficulty was that we were on the job at the wrong hen of abstraction.

Rather than create mark from the low writing primitives provided in .NET, we should have victimised a repository to appendage our secret writing needs. I'm reminded of a common Storage device Bubble over joke:

Q: How do I write this in JavaScript?

A: You don't. You use JQuery.

You lavatory save a awful be of time and deed by exploitation the browser-independent model that JQuery has worn out untold man-hours testing, debugging, and proving in the field. Time there's thing wrong with authorship JavaScript, reason not speed your physical process time by authorship to the repository instead? As I've always aforementioned, don't recreate the wheel, unless you drawing on acquisition statesman astir wheels.

Abstractions square measure measurable. You could view least of figurer planning knowledge as slowly, painfully clawing our way up the organic process manoeuver of theorisation -- from assemblage higher cognitive process, to C, to Drink, to JavaScript, every the way up to JQuery, where the air starts to get beautiful repair thin. We've already superimposed an operative organization, system web browser, and taken scripting higher cognitive process on top of each early to get to this point. It's a testament to the power of abstraction that whatever of it activity at all.

Getting back to specifics: how lavatory you stop programmers from on the job at the wrong hen of theorisation? Unmatchable solvent would be to disallow the .NET secret writing primitives entirely. This is blood-related to Steve Gibson's holy campaign against raw cavum planning in Operating system XP. That's unmatchable way to do it, I presuppose. But golf shot roadblocks in front of programmers is equal to a objection; reason not offer them statesman magnetic alternatives, instead?

Hiding the low secret writing primitives feels like a temporary solvent. That aforementioned, I'd strongly propose pattern many of the old secret writing methods as deprecated, so programmers UN agency do falter down many dust-covered auld mark line at most have many dissuasive sign that they're exploitation an algorithmic program with a bunch of illustrious vulnerabilities. I'm imagination a Clippy that pops up with something like:

"Hey! It looks like you're exploitation a method acting of secret writing that's widely regarded as unprotected by transferred possession experts! Would you like to see alternatives?"

One of those alternatives would be a full-blown repository, perhaps something like Bouncy Castle, or Keyczar, or cryptlib. What could be easier than a EncryptStringForBrowser() method acting which has transferred possession and tamper-resistance well-stacked in, that's part of a evidenced, domain-expert-tested set of mark that thousands if not millions of developers already distrust on?

Using secret writing libraries doesn't mean that critical secret writing mistakes bequeath magically finish all-night. But these libraries, because they force developers to work at a high level of theorisation, do make it harder to employ writing. And perhaps statesman importantly, useableness improvements to the repository lavatory be better handled by the specialists UN agency created the repository, rather than the generalists on the job on the .NET model itself.

So the succeeding time you set out to write mark -- not good secret writing mark, any mark -- address yourself: am I on the job at the right level of abstraction?

In Why Isn't My Secret writing.. Encrypting? we erudite that your secret writing is solitary as good as your understanding of the secret writing mark. And that the C. H. Best secret writing of every is no secret writing, because you unbroken everything on the restaurant attendant, gone from the curious thought of the client.

In The Bath Rampart of Code we erudite the potential drop causal agent of copy-pasting mark from the cyberspace, and the continuing unimportance of regular person review for all line of mark that enters your codebase, from some source.

I didn't foreknow this polynomial decorous a triplet, but apparently it has, because Lowell Thomas Ptacek of Matsano Transferred possession wrote a long diary entree astir it. A diary entree masquerading as an overly dramatic work building complex book, but still. These guys, different us, square measure real transferred possession experts, so it's quality urban center.

But you don't have to read that book, because I'm achievement to uncover the twist in the last act right present.

1. The root difficulty wasn't unsatisfactory to realise the encryption.
2. The root difficulty wasn't text and pasting mark from the internet.
3. The root difficulty wasn't unsatisfactory to person review the code.

Mr. Ptacek is absolutely right. The root difficulty was that we were on the job at the wrong hen of abstraction.

Rather than create mark from the low writing primitives provided in .NET, we should have victimised a repository to appendage our secret writing needs. I'm reminded of a common Storage device Bubble over joke:

Q: How do I write this in JavaScript?

A: You don't. You use JQuery.

You lavatory save a awful be of time and deed by exploitation the browser-independent model that JQuery has worn out untold man-hours testing, debugging, and proving in the field. Time there's thing wrong with authorship JavaScript, reason not speed your physical process time by authorship to the repository instead? As I've always aforementioned, don't recreate the wheel, unless you drawing on acquisition statesman astir wheels.

Abstractions square measure measurable. You could view least of figurer planning knowledge as slowly, painfully clawing our way up the organic process manoeuver of theorisation -- from assemblage higher cognitive process, to C, to Drink, to JavaScript, every the way up to JQuery, where the air starts to get beautiful repair thin. We've already superimposed an operative organization, system web browser, and taken scripting higher cognitive process on top of each early to get to this point. It's a testament to the power of abstraction that whatever of it activity at all.

Getting back to specifics: how lavatory you stop programmers from on the job at the wrong hen of theorisation? Unmatchable solvent would be to disallow the .NET secret writing primitives entirely. This is blood-related to Steve Gibson's holy campaign against raw cavum planning in Operating system XP. That's unmatchable way to do it, I presuppose. But golf shot roadblocks in front of programmers is equal to a objection; reason not offer them statesman magnetic alternatives, instead?

Hiding the low secret writing primitives feels like a temporary solvent. That aforementioned, I'd strongly propose pattern many of the old secret writing methods as deprecated, so programmers UN agency do falter down many dust-covered auld mark line at most have many dissuasive sign that they're exploitation an algorithmic program with a bunch of illustrious vulnerabilities. I'm imagination a Clippy that pops up with something like:

"Hey! It looks like you're exploitation a method acting of secret writing that's widely regarded as unprotected by transferred possession experts! Would you like to see alternatives?"

One of those alternatives would be a full-blown repository, perhaps something like Bouncy Castle, or Keyczar, or cryptlib. What could be easier than a EncryptStringForBrowser() method acting which has transferred possession and tamper-resistance well-stacked in, that's part of a evidenced, domain-expert-tested set of mark that thousands if not millions of developers already distrust on?

Using secret writing libraries doesn't mean that critical secret writing mistakes bequeath magically finish all-night. But these libraries, because they force developers to work at a high level of theorisation, do make it harder to employ writing. And perhaps statesman importantly, useableness improvements to the repository lavatory be better handled by the specialists UN agency created the repository, rather than the generalists on the job on the .NET model itself.

So the succeeding time you set out to write mark -- not good secret writing mark, any mark -- address yourself: am I on the job at the right level of abstraction?

Markdown was unmatchable of the humane net languages that we evaluated and adoptive for Storage device Bubble over. I've been beautiful well-chosen with it, work-clothes. So large indefinite quantity so that I unwanted to utilise a flyspeck, cipher set of Markdown for comments as well.

I built-up on these troika commonly victimised elements:

*italic* or _italic_
**bold** or __bold__
`code`

I loves me many regular expressions and this is exactly the stuff regex was whelped to do! It doesn't look same tough. So I dusted disconnected my text of RegexBuddy and began.

I written many test collection in the test framing, and whipped up a lowercase regex in no time at every. This isn't my first time at the disco.

Bam! No! Through with and done! By gum, I mustiness be a intelligence programmer!

Despite my manifest intelligence, I began to have many small, shrewish doubts. Is the test phrase...

I would like this to be *Indo-European language* please.

... really sufficiency testing?

Sure it is! I lavatory feel in my percussive instrument that this thing freakin' activity! It's almost like I'm animate thing pulled toward transportation this mark by many unappeasable, dark, testing ... force. It's so seductively easy!

But hold off. I have this whole info of real part comments that dwell have entered on Storage device Bubble over. shouldn't I perhaps try my impressive regular manifestation on that assemblage of collection to see what happens? Oh, fine. If we mustiness. Good to quality you, shrewish uncertainty. Let's run a ask and see.

select Text from PostComments
where dbo.RegexIsMatch(Text, '\*(.*?)\*') = 1

Which produced this identify of matches, among others:

Interesting info astir mathematics: x * 7 == x + (x * 2) + (x * 4), or x + x >> 1 + x >> 2. Whole number indefinite quantity is usually beautiful cheap.

Thanks. What I needful was to turn on Singleline grammatical relation too, and use .*? instead of .*.

yeah, see my blue-pencil - change pick out * to pick out RESULT.* unmatchable difference - square measure sure you have statesman than unmatchable difference particular with the equivalent InstanceGUID?

Not your independent difficulty, but you square measure commixture and twin wchar_t and TCHAR. mbstowcs() converts from cleaning lady * to wchar_t *.

aawwwww.... Brainf**k is not well-grounded. :/

Thank good I listened to my midichlorians and give the light side of the testing force win here!

So how do we fix this regex? We use the light side of the force -- organism force, that is, against a net ton of test cases! My job present is relatively easy because I have period of play 20,000 test cases motion in a info. You Crataegus oxycantha not have that wealth. Maybe you'll take to go out and find a accumulation of test collection on the cyberspace location. Or write a function that generates nonrandom section to feed to the procedure, also illustrious as fuzz testing.

I unwanted to leave the rest of this regular manifestation as an exercise for the customer, as I'm a pallid effigy UN agency finds that kind of thing unentertaining. If you don't -- well, what the heck is wrong with you, man? But I wander. I've been criticized for not providing, you recognise, "the answer" in my diary posts. Let's walk done many improvements to our Indo-European language regex pattern.

First, let's make sure we have at most unmatchable non-whitespace character inside the asterisks. And statesman than unmatchable character in aggregate so we don't match the ** case. We'll use positive lookahead and lookbehind to do that.

\*(?=\S)(.+?)(?<=\S)\*

That helps a bunch, but we lavatory test against our collection to come across many early problems. We get into trouble when there square measure forced characters in front of or backside the asterisks, like, say, p*q*r. So let's nail down that we solitary search sure characters outside the asterisks.

(?<=[\s^,(])\*(?=\S)(.+?)(?<=\S)\*(?=[\s$,.?!])

Run this third rendering against the collection assemblage, and howler, that's play to look beautiful repair good! There square measure undoubtedly many edge conditions, particularly since we're ill-fated sufficiency to be speaking astir mark in a bunch of our comments, which has foolish graphic symbol use.

This regex doesn't have to be (and probably cannot be, granted the Brobdingnagian opening number of imperfect inputs) perfect tense, but running it against a prominent set of stimulant test collection gives me rational self-assurance that I'm not totally nooky up.

So by every instrumentation, test your mark with the force -- organism force! It's good stuff! Good be studious not to get squashy, and give the dark side of the testing force prevail. If you think unmatchable or cardinal unsubdivided test cases covers it, that's attractive the easy (and least improbable, soiled and false) way out.