Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

SQL Fiddle

sqlfiddle.com

Post reply on HN
Print view View original

61–70 of 87 posts

Re: SQL Fiddle

#61

Post by jere »

pbreit wrote: post #50
timsayshey wrote: post #32

Earlier quoted context omitted.

I agree that Adobe's ColdFusion is deadend but not Open Source Railo which is more of a Java Framework. At the end of the day there's no logical reason against CFML any more than there is against PHP, Rails, etc.

Sorry, but it is completely logical to choose Ruby/Rails, Python/Django, etc over CFML in 2013 for web projects.

>ColdFusion is... not a particularly good choice these days for a variety of reasons.

Not that I disagree with you, but I wonder if the reasons you have in mind are valid. A lot of the reasons one might assume CF sucks aren't really applicable (i.e. speed, security) any more than most other languages.

Re: SQL Fiddle

#62

Post by rip747 »

pud wrote: post #42
nbevans wrote: post #21

http://i.imgur.com/RHxoW5J.png 1. Security flaw in leaking that much error info. 2. Who uses ColdFusion in 2013?

I use ColdFusion (well, technically I use Railo) and I love it. I agree that the owner of the website here needs a tag, which tells the server what to do instead of showing that big error message.

CFWheels has that already built into it, but I'm convinced that its a problem with the framework itself and not anything the author did.

the issue seems coming from this line:

https://github.com/cfwheels/cfwheels/blob/v1.1.8/wheels/even...

basically when wheels throws an error, it will try to send out an error email containing the environment and framework information. In this case there seems to be no smtp server defined, so CFML itself is spitting up its skull.

The error handling can be reworked so that the $mail() call is wrap in a try/catch so the 500 page will just be shown.

disclosure: part of wheels core.

Re: SQL Fiddle

#63

Post by jsmeaton »

buckbova wrote: post #58
goostavos wrote: post #16

Earlier quoted context omitted.

As a guy about to finish up school and make his way into the work force, would you be willing to release some of the questions you use in interviews. It be neat to see which areas I need to patch up knowledge wise!

Try this quiz. It's based on a quiz I saw posted on HN a while back. http://sqlfiddle.com/#!3/3973f/1 Edit: I'm not allowed to reply just yet, so I'll post here. You need to add the boss first to the employee table with a NULL bossID. Then you add another employee that has a bossID that refers to that last added record.

Aww - no test data? And when trying to add test data, I get errors due to the self-referencing foreign key. Perhaps the schema could do with a

> DEFAULT (IDENT_CURRENT('dbo.Employees'))

on the BossID column. Been awhile (years) since I've worked with SQL Server.

Re: SQL Fiddle

#65

Post by ZoF »

benihana wrote: post #24
nbevans wrote: post #21

http://i.imgur.com/RHxoW5J.png 1. Security flaw in leaking that much error info. 2. Who uses ColdFusion in 2013?

> Who uses ColdFusion in 2013? About 2/3 of the federal government. Also, that isn't ColdFusion, it's Railo, an OS alternative.

He had "railo" highlighted... I think he knows that it's not literally ColdFusion...

Re: SQL Fiddle

#66

Post by jka »

breckinloggins wrote: post #54

One thing that would be great here would be autocompletion. I often forget exactly what the syntax of "CREATE TABLE" is or which options I can use. The same goes for infrequently used SQL functions (now is that date_diff or datediff?). You could combine autocompletion with a doc panel that automatically updates to show the "quick reference" for the keyword you're currently typing.

Slightly off-topic, but one thing that always impressed me about Microsoft's LINQ as an alternative to SQL queries is that they considered autocompletion up-front in the language syntax.

For SQL, it's a big challenge - you may type in 'SELECT', but your IDE still has no idea what relation(s) you're going to access, and so it can't reasonably autocomplete the possible fields you might want to use.

It's an interesting example of a language that was arguably designed for human readability in the small ('give me this from that cross-referenced with this other thing'), but in some sense became a weakness of sorts later on.

Re: SQL Fiddle

#67

Post by radiowave »

jka wrote: post #66
breckinloggins wrote: post #54

One thing that would be great here would be autocompletion. I often forget exactly what the syntax of "CREATE TABLE" is or which options I can use. The same goes for infrequently used SQL functions (now is that date_diff or datediff?). You could combine autocompletion with a doc panel that automatically updates to show the "quick reference" for the keyword you're currently typing.

Slightly off-topic, but one thing that always impressed me about Microsoft's LINQ as an alternative to SQL queries is that they considered autocompletion up-front in the language syntax. For SQL, it's a big challenge - you may type in 'SELECT', but your IDE still has no idea what relation(s) you're going to access, and so it can't reasonably autocomplete the possible fields you might want to use. It's an interesting…

One way to approach it would be to have a grab bag of templates. You drag a template like "SELECT ... FROM ... WHERE ..." out into the editor and then tab through it filling in the blanks.

I've always thought that SQL's notion of readability was somewhere between superficial and disingenuous - you can read the query and it sounds like it ought to make sense, but there's so many occasions where you won't actually be able to understand why it isn't working unless you know how to translate in your head from the order in which the clauses are written, to the order in which they're carried out.

Re: SQL Fiddle

#68

Post by jka »

radiowave wrote: post #67
jka wrote: post #66

Earlier quoted context omitted.

Slightly off-topic, but one thing that always impressed me about Microsoft's LINQ as an alternative to SQL queries is that they considered autocompletion up-front in the language syntax. For SQL, it's a big challenge - you may type in 'SELECT', but your IDE still has no idea what relation(s) you're going to access, and so it can't reasonably autocomplete the possible fields you might want to use. It's an interesting…

One way to approach it would be to have a grab bag of templates. You drag a template like "SELECT ... FROM ... WHERE ..." out into the editor and then tab through it filling in the blanks. I've always thought that SQL's notion of readability was somewhere between superficial and disingenuous - you can read the query and it sounds like it ought to make sense, but there's so many occasions where you won't actually be a…

True - but so much of the value of autocomplete comes from just being able to try things out:

'from customers, orders select ...'

Would allow showing a whole bunch of columns, and you could quickly determine whether those tables had the information you're after, or not.

It might seem strange/lazy that the developer can explore the environment (schema) this way without 'knowing' what's out there, but in practice that's exactly what we all do with software development, web search, etc - we try things out and see what comes back, rather than knowing in the first place. Data development has a way to go, but I think things will improve..

Re: SQL Fiddle

#69

Post by jka »

jka wrote: post #68
radiowave wrote: post #67

Earlier quoted context omitted.

One way to approach it would be to have a grab bag of templates. You drag a template like "SELECT ... FROM ... WHERE ..." out into the editor and then tab through it filling in the blanks. I've always thought that SQL's notion of readability was somewhere between superficial and disingenuous - you can read the query and it sounds like it ought to make sense, but there's so many occasions where you won't actually be a…

True - but so much of the value of autocomplete comes from just being able to try things out: 'from customers, orders select ...' Would allow showing a whole bunch of columns, and you could quickly determine whether those tables had the information you're after, or not. It might seem strange/lazy that the developer can explore the environment (schema) this way without 'knowing' what's out there, but in practice that'…

Actually, one alternative approach would be: use a search index to autocomplete after even a basic 'SELECT' statement (no other context), but then fill the FROM clause based on the selection.

i.e. - type:

'select city'

prompts with:

'city (table airports)'

'city (table populations)'

'city (table companies)'

And then, upon selection, you have a statement: 'select city from '.

There's a lot more here though. Worth pondering for a while I think.

Re: SQL Fiddle

#70

Post by b0z0 »

aimhb wrote: post #35
nbevans wrote: post #21

http://i.imgur.com/RHxoW5J.png 1. Security flaw in leaking that much error info. 2. Who uses ColdFusion in 2013?

Can someone explain how leaking a traceback is a security flaw, considering it's a lot less information than just releasing the source code (which we do all the time and don't consider it a security flaw)?

There are tons of good explanations of why a stacktrace is a security risk. See, for example, this article that appeared on HN: http://www.troyhunt.com/2013/07/everything-you-wanted-to-kno...
Post reply on HN