English Language & Usage: Multi-Layered Discourse Room

KitFox

15:00

Oh, thanks. Just a min.

Mr. Shiny and New 安宇

The value you hope to see is "AL32UTF8"

tchrist

@Mr.ShinyandNew安宇 If the data is in Western European, or if it is a web page even in an Asian script, UTF-8 wins.

KitFox

Table does not exist.

Mr. Shiny and New 安宇

@KitFox hm.

tchrist

If (and only if) the data is in purely Asian script, not with XML/HTML/SGML markup, then UTF-16 wins.

If you are a programmer, once you get it inside, you probably want it converted to UTF-32/UCS-4 for convenience’ sake. But sometimes not.

Mr. Shiny and New 安宇

15:01

@tchrist What do you mean? We are storing our data in UTF-8, but the max-length in the db is expressed in bytes. So we need some way to tell the user how much more they can type.

tchrist

@Mr.ShinyandNew安宇 You can’t know until they type it.

Mr. Shiny and New 安宇

@tchrist exactly.

@KitFox SELECT * FROM NLS_DATABASE_PARAMETERS

tchrist

Worst case is bytes = 4 * UTF-8 chars, which is virtually impossible to ever actually occur.

KitFox

NLS_NCHAR_CHARACTERSET AL16UTF16

NLS_CHARACTERSET WE8MSWIN1252

tchrist

Well.

Wait.

Matt Эллен

15:04

1252

tchrist

Those are different.

Matt Эллен

isn't on the code page and the other the encoding

so it's fine

oh

tchrist

You can map 1252 to UTF16, but you can only go back the other way if you started with 1252, because UTF16 has byzintillions more characters than 1252 has.

Matt Эллен

I guess not

Mr. Shiny and New 安宇

@KitFox That's not UTF-8.

tchrist

15:05

That’s sad.

But UTF-16 is also variable width, one notes.

However, all 1252 codepoints map to exactly 2 bytes in UTF16.

Well, legal ones.

There are illegal 1252 code points.

The map is not full.

I wonder what kind of Indians we're talking here?

UTF16 does not per se specify such.

And I’d hate to think of BOMming every field.

Mr. Shiny and New 安宇

so, @Kit, you are using a non-unicode character encoding for VARCHAR2. It is one-byte-per-char

Matt Эллен

We ate MS Win 1252. I think they just missed out an h

Andy

but there are empty regions devoted to characters unique to charsets

tchrist

What do you mean? The holes in the 1252 map?

Andy

wait, i confused myself. ignore me today, i'm not all here

KitFox

15:08

@Mr.ShinyandNew安宇 No, it's not.

Mr. Shiny and New 安宇

Also if means if you put UTF-8 data in there, strange things will happen. For one thing, you won't be able to query the data properly. For another thing, you will run into strange encoding issues if you ever have two different clients connected to the DB and one follows the rules and one doesn't.

KitFox

cries

tchrist

These are the cp1252 holes:

cp1252       81  ⇒  U+FFFD  ＜ � ＞  \N{REPLACEMENT CHARACTER}
cp1252       8D  ⇒  U+FFFD  ＜ � ＞  \N{REPLACEMENT CHARACTER}
cp1252       8F  ⇒  U+FFFD  ＜ � ＞  \N{REPLACEMENT CHARACTER}
cp1252       90  ⇒  U+FFFD  ＜ � ＞  \N{REPLACEMENT CHARACTER}
cp1252       9D  ⇒  U+FFFD  ＜ � ＞  \N{REPLACEMENT CHARACTER}

KitFox

It's not my database.

It's a horrible piece of work.

tchrist

Those 5 code points are illegal in cp1252, and thus have no Unicode mapping.

I don’t know why they exist.

I just know that they do.

KitFox

15:10

There. I've said it. I didn't want to bad-mouth my developer—my friend—but this work is crap.

Anyway, it's nothing I have to do anything about. I'll just advocate for client-side control of the string lengths and be done with it.

Mr. Shiny and New 安宇

@KitFox It will be fine as long as you don't try to put in any characters that aren't in Windows-1252 code page.

KitFox

@Mr.ShinyandNew安宇 And what are the chances of that, really.

tchrist

@MattЭллен Not counting the 5 holes, it has 27 code points that do not map to Latin-1.

Whereas MacRoman has 48.

KitFox

It's annoying that she just moved it to a new server, but we didn't update it at all.

tchrist

@KitFox That depends on the person entering the data.

I have had people where one uses a Mac to enter stuff with and another uses Windows, and both encodings show up.

KitFox

15:14

We do have Mac and PC users entering data.

tchrist

Ug.

KitFox

And for some reason, everyone here thinks doing stuff manually is really important.

And that visitors to the site want to search through pages and pages of text to find the thing they are looking for.

And also that web visitors really want to read paragraphs of text.

tchrist

@Mr.ShinyandNew安宇 I’ve been out of the web world too long to remember whether there is some elegant solution to this. I am used to specifying the page’s charset on output, but I don’t know what to do about specifying or at least learning the input charset they use to fire back in the POST data.

KitFox

"Click here for a pdf of this report." Me: Um. You know you could reduce that to an icon.

tchrist

Most stuff I use now seems to accept UTF-8 input, but I don’t actually know “why”.

KitFox

15:17

Inbred Strain;
Additional information on Inbred Strains.

Why do we do stuff like this?

Mr. Shiny and New 安宇

@tchrist The browser sends the character encoding in the request. You can force a certain encoding in a form attribute. On the server side, you need to handle the possibility of wrong data.

KitFox

In 11 pt font, on a page where you are absolutely required to scroll down.

tchrist

@Mr.ShinyandNew安宇 So the charset is part of the HTTP POST response, but may be wrong?

Mr. Shiny and New 安宇

@tchrist No, I mean it may be something other than what you want.

You can't really control what a browser will send you.

tchrist

But can you get it to tell you what it sent you?

Mr. Shiny and New 安宇

15:19

But usually browsers default to iso-8859-1 or to whatever the page's encoding is.

KitFox

I am complaining here.

tchrist

Now, this is a different matter.

KitFox

pouts

Mr. Shiny and New 安宇

@tchrist It sends the encoding with the request, otherwise you have to assume it's iso-8859-1

tchrist

The web standard says that if you do not specify an encoding, it is ISO-8859-1.

jinx

I just don’t remember ever coding to detect the post encoding and decode/convert as appropriate.

Mr. Shiny and New 安宇

15:21

But I could make a site that serves iso-8859-2 pages and creates forms with iso-8859-2 forms and some browser sends me a UTF-8 POST. I can't stop that from happening. I should attempt to do something intelligent with the request. like reject it, or translate it

tchrist

That’s what I was wondering whether happened.

When they send you the UTF8 POST, does the HTTP header admit that, or not? I bet not.

Mr. Shiny and New 安宇

@tchrist if it doesn't, then it's an iso-8859-1 post.

With (potentially) corrupt data.

tchrist

I see.

Ok, so what is happening “here”?

I'm using curly quotes. What is SE receiving?

Mr. Shiny and New 安宇

马

tchrist

macbook# byte2uni -e MacRoman | grep QUOT | grep DOUB
MacRoman     C7  ⇒  U+00AB  ＜ « ＞  \N{LEFT-POINTING DOUBLE ANGLE QUOTATION MARK}
MacRoman     C8  ⇒  U+00BB  ＜ » ＞  \N{RIGHT-POINTING DOUBLE ANGLE QUOTATION MARK}
MacRoman     D2  ⇒  U+201C  ＜ “ ＞  \N{LEFT DOUBLE QUOTATION MARK}
MacRoman     D3  ⇒  U+201D  ＜ ” ＞  \N{RIGHT DOUBLE QUOTATION MARK}
MacRoman     E3  ⇒  U+201E  ＜ „ ＞  \N{DOUBLE LOW-9 QUOTATION MARK}
macbook# byte2uni -e CP1252 | grep QUOT | grep DOUB
cp1252       84  ⇒  U+201E  ＜ „ ＞  \N{DOUBLE LOW-9 QUOTATION MARK}

Mr. Shiny and New 安宇

15:25

Content-Type： application/json; charset=utf-8

That's what my browser POSTed to SE when I typed that horse

tchrist

Ahah.

So I am typing the Mac escape that makes a “ and a ”, but it sends not 0xD2 and 0xD3 but U+201C and U+201D encoded as UTF-8?

Mr. Shiny and New 安宇

@tchrist yes.

tchrist

Wouldn’t that solve @Kit’s dual-input-system issue, too?

Mr. Shiny and New 安宇

The only valid character set for "HTML" is Unicode. But on transfer over HTTP, or storage to disk, etc, any encoding is possible.

@tchrist Not really.

tchrist

How come?

Not enough control?

Mr. Shiny and New 安宇

15:28

In your browser, the HTML and JS are running in a unicode world. But to transfer the data to the server, some encoding must be specified. The default is (oddly) iso-8859-1.

tchrist

@Mr.ShinyandNew安宇 I’m not sure I understand that statement.

Awal Garg

Hmm, room title should be changed to "Programming Languages and Usage" :P

tchrist

I know when they did the SIP protocol, they explicitly used the HTTP protocol as a base but assumed UTF-8 not ISO-8859-1 by default. I think XML may default to UTF-8, also. A default of ISO-8859-1 is really not a good idea.

Mr. Shiny and New 安宇

The browser takes the form data and does its best to encode that data in the requested encoding. So if you have Chinese text in a text field, but the form is set up to send iso-8859-1, who knows what it will send? I've seen it send html escapes: &nnnnn; style

tchrist

But even so, our data entry people kept sending us CP1252 and MacRoman in the SIP data.

It’s usually caused by when you do customer data entry for businesses.

You have no control over the upstream form data entry folks.

Mr. Shiny and New 安宇

15:31

The server gets the HTTP request and has to parse it using the encoding specified (or the default). Then it's up to the server to deal with those bytes/chars. I don't know what server Kit's using; various platforms do it better or worse. In Java, it gets automatically parsed and then stored in memory using UCS-2, which is at least some kind of unicode.

But then when you want to store it in the database, you need a database connection and those are all vendor-specific. You pass a string in and the combination of your application platform and your DB library and the DB itself conspire to store the data, as bytes, on a disk somewhere.

By using "unicode" throughout, you can ensure that everything works.

But Kit's VARCHAR2 columns are Windows codepage 1252.

tchrist

So like let’s just hypothetically consider a UK customer named Prêt-à-Manger. Someone entered that customer name in the database, and that name is sent for SIP transactions. But whether they entered it in CP-1252 or MacRoman or UTF-8 is like playing Unicode Roulette.

Mr. Shiny and New 安宇

So I'm not sure what will happen if she tries to store, say, Chinese in there.

tchrist

@Mr.ShinyandNew安宇 That would be a problem.

With long text, you can virtually always figure out the encoding. But not for short snippets.

I mean, programmatically/deterministically.

Mr. Shiny and New 安宇

@tchrist I'm not familiar with SIP. but with regular web sites, the browser takes the inputted characters from the OS and enters them into the DOM and does whatever tricks it needs to to handle all of unicode, because "HTML" is unicode.

KitFox

I don't remember what server it was moved to. It was on a Sun system, moved to a Linux system, I think.

Mr. Shiny and New 安宇

15:34

Then it transfers the data to the server using some encoding.

If you have some application that builds on top of HTTP and demands a particular encoding, it is probably a bug in the client software if it sends a different encoding.

tchrist

If they enter on a Mac, and it just sticks them in as is without converting to 1252, you will have trouble.

Three of the five "illegal" 1252 code points turn out to be not to rare in MacRoman:

macbook# byte2uni -e cp1252 | grep REPL cp1252 81 ⇒ U+FFFD ＜ � ＞ \N{REPLACEMENT CHARACTER}
cp1252 8D ⇒ U+FFFD ＜ � ＞ \N{REPLACEMENT CHARACTER}
cp1252 8F ⇒ U+FFFD ＜ � ＞ \N{REPLACEMENT CHARACTER}
cp1252 90 ⇒ U+FFFD ＜ � ＞ \N{REPLACEMENT CHARACTER}
cp1252 9D ⇒ U+FFFD ＜ � ＞ \N{REPLACEMENT CHARACTER}
macbook# byte2uni -e MacRoman | egrep -w '8[1DF]|9[0D]'

KitFox

We wouldn't use those, probably.

tchrist

The worst ones are when they differ, though.

Mr. Shiny and New 安宇

@tchrist For websites, the OS's native encoding is never an issue, because the website can control the page's encoding. I don't know about SIP clients. Maybe you have some that are broken.

tchrist

@Mr.ShinyandNew安宇 It’s the low-level telecom wholesaler operators that are the problem.

Mr. Shiny and New 安宇

15:38

The best thing for Kit to do is to convert the entire stack from 1252 to UTF-8.

oerkelens

morning :)

KitFox

> These mutant mice may be used as a Cre-reporter strain; to test the tissue/cellular expression pattern of cre transgenic mice.

See what I have to deal with?

@Mr.ShinyandNew安宇 I have no control over that.

Mr. Shiny and New 安宇

@tchrist according to RFC 3261 you can have non-UTF-8 requests in SIP, if they provide the character set in a header.

tchrist

Eek!!!

I missed that!

I coded it to assume UTF-8.

Mr. Shiny and New 安宇

@KitFox Without using unicode everywhere, you can't ensure that any characters not in 1252 will work. The results are basically undefined.

tchrist

15:44

The problem is, by the time it gets to us, some feeds did not have the full SIP headers, they had been condensed.

Mr. Shiny and New 安宇

@tchrist We now come to the conclusing of Mr. Shiny and New's SIP knowledge. If you enjoyed this broadcast, please join our pledge drive and become a member.

KitFox

@Mr.ShinyandNew安宇 I still have no control over it.

Mr. Shiny and New 安宇

@KitFox I'm just pointing it out. It is a major limitation and there is only one solution. But until you fully unicodeize everything, some illegal stuff may appear to work in some cases and data might get stored in an illegal way. Then when you convert later, you'll find that some of your data is actually garbage. And there will be no way to tell the difference.

tchrist

I have good news and bad news.

KitFox

@Mr.ShinyandNew安宇 Well, it's really the least of our problems.

@tchrist What is it?

Mr. Shiny and New 安宇

15:48

@KitFox yay! hidden data-corruption is the least of your problems! ;)

tchrist

The good news is that CP1252 and MacRoman both represent the same 231/256 possible code points.

The bad news is that 124 of them are in different positions.

KitFox

We have so much duplication and crappy code and rickety data structures and poor practices...

@tchrist Ugh.

Mr. Shiny and New 安宇

@KitFox Just fix things up a bit at a time.

KitFox

Well, that would be nice except it's not up to me.

I just have to design around it.

tchrist

For example, the EN DASH is the most common non-ASCII character used in English-language papers submitted to PubMed and to Elsevier.

But watch:

cp1252 96  ⇒  U+2013 ＜–＞ \N{EN DASH}                     MacRoman 96  ⇒  U+00F1 ＜ñ＞ \N{LATIN SMALL LETTER N WITH TILDE}
cp1252 D0  ⇔  U+00D0 ＜Ð＞ \N{LATIN CAPITAL LETTER ETH}    MacRoman D0  ⇒  U+2013 ＜–＞ \N{EN DASH}

KitFox

15:51

Which is also funny because I'm not a designer either.

tchrist

It's 96 in cp1252 and D0 in MacRoman.

That's why you'll go insane if you don't upconvert everything to Unicode.

In fact, MacRoman and cp1252 share only 5 different 8-bit code points in common.

Mr. Shiny and New 安宇

It's not so bad, as long as any particular bit of text has a known, correct, tag identifying its encoding.

@tchrist Because why wouldn't they.

tchrist

@Mr.ShinyandNew安宇 True.

cp1252 A2  ⇔  U+00A2 ＜¢＞ \N{CENT SIGN}            MacRoman A2  ⇔  U+00A2 ＜¢＞ \N{CENT SIGN}
cp1252 A3  ⇔  U+00A3 ＜£＞ \N{POUND SIGN}           MacRoman A3  ⇔  U+00A3 ＜£＞ \N{POUND SIGN}
cp1252 A9  ⇔  U+00A9 ＜©＞ \N{COPYRIGHT SIGN}       MacRoman A9  ⇔  U+00A9 ＜©＞ \N{COPYRIGHT SIGN}
cp1252 B1  ⇔  U+00B1 ＜±＞ \N{PLUS-MINUS SIGN}      MacRoman B1  ⇔  U+00B1 ＜±＞ \N{PLUS-MINUS SIGN}
cp1252 B5  ⇔  U+00B5 ＜µ＞ \N{MICRO SIGN}           MacRoman B5  ⇔  U+00B5 ＜µ＞ \N{MICRO SIGN}

Those are the overlappers.

Mr. Shiny and New 安宇

lol

tchrist

So in text with all 7-bit ASCII plus only those 8-bit code points, it is indeterminable -- and immaterial -- which of the two encodings it was in.

Mr. Shiny and New 安宇

15:54

I recently had a great issue with encoding, and transferring text across the Java/Ruby boundary (using jruby).

tchrist

Ick.

Ruby is troublesome about that.

Java can do it right, if you’re not uncareful.

Mr. Shiny and New 安宇

The beauty of it is: Ruby doesn't use just one encoding internally (I learned).

tchrist

That is not beauty.

It is sin.

And I have a talk about it.

Mr. Shiny and New 安宇

But Java DOES use just one encoding. So WTF doesn't jRuby just use Unicode?

I mean, it should be completely transparent that every fucking string coming from Java and passed into any jruby anything anywhere is unicode.

It cannot be anything else.

tchrist

Because it was written by Japanese people who wanted to be able to have rare non-Unicode characters.

It is a super asspain.

I know I know I know.

Mr. Shiny and New 安宇

15:57

And. And And. And the fucking HTML parser I was using was defaulting to US_ASCII encoding.

@tchrist That's Ruby's excuse. jRuby is a Java Ruby. It has no excuse for not "Just Working"(tm)

cornbread ninja 麵包忍者

i.imgur.com/PDiajRI.jpg?1

tchrist

I have to run, but here for @Kit's benefit is the full 8-bit mapping for 8-bit data between CP1252 and MacRoman. It should tell her why not unifying on Unicode will hurt hurt hurt.

cp1252 80  ⇒  U+20AC ＜€＞ \N{EURO SIGN}                              MacRoman 80  ⇒  U+00C4 ＜Ä＞ \N{LATIN CAPITAL LETTER A WITH DIAERESIS}
cp1252 81  ⇒  U+FFFD ＜�＞ \N{REPLACEMENT CHARACTER}                  MacRoman 81  ⇒  U+00C5 ＜Å＞ \N{LATIN CAPITAL LETTER A WITH RING ABOVE}
cp1252 82  ⇒  U+201A ＜‚＞ \N{SINGLE LOW-9 QUOTATION MARK}            MacRoman 82  ⇒  U+00C7 ＜Ç＞ \N{LATIN CAPITAL LETTER C WITH CEDILLA}
cp1252 83  ⇒  U+0192 ＜ƒ＞ \N{LATIN SMALL LETTER F WITH HOOK}         MacRoman 83  ⇒  U+00C9 ＜É＞ \N{LATIN CAPITAL LETTER E WITH ACUTE}

cornbread ninja 麵包忍者

WTF.

Mr. Shiny and New 安宇

@cornbreadninja麵包忍者 you have ?1

tchrist

@Mr.ShinyandNew安宇 Oh, I missed it was jRuby, not Ruby. That’s sick!

Mr. Shiny and New 安宇

15:59

@tchrist yeah. Oh, and the problem only exists if Java's default code page is UTF-8.

If Java's default is cp-1252, it works fine.

W. T. F.

tchrist

To be honest, the MacRoman set is the more useful of the two.

But they should all be normalized out of their legacy encodings into Unicode.

KitFox

@tchrist It's not up to me!

Mr. Shiny and New 安宇

I'm heading out for a while. ttyl

tchrist

@KitFox I’m so sorry. They should be storing stuff in UTF-8.

Me too, bye.

KitFox

Bye!

cornbread ninja 麵包忍者

16:06

Thanks! @Mr.Shiny

All I did was copy the link. Hmm.

Great, I've killed the chat. goes back to watching Hulu

Stack Exchange

Limerick Machine: I’m not gonna sell you a sonnet That you can display on your bonnet I’ve got fifty...

posted on June 06, 2014 by sgdi

I’m not gonna sell you a sonnet That you can display on your bonnet I’ve got fifty verses To write on to hearses I’ve a soliloquy if you want it

Matt Эллен

I think it's more that everyone left, not that you killed chat.

time for me to go!

cornbread ninja 麵包忍者

D'oh!

Bai! :)

KitFox

16:23

@MattЭллен No!

Oh! Hi.

cornbread ninja 麵包忍者

16:33

@KitFox hello.

KitFox

Is it me you're looking for?

cornbread ninja 麵包忍者

^_^

Are you somewhere feeling lonely, or is someone loving you?

KitFox

16:48

Both, kind of.

cornbread ninja 麵包忍者

Imaginary lover?

You're mine, anytime.

KitFox

Aww.

She's an easy lover.

I'm trying not to be too disappointed that my friend hasn't read my story yet.

Also, I'm trying not to freak out about how much stuff I have to do right now.

cornbread ninja 麵包忍者

Oh! I was going to ask you if you finished your Nose piece yet!

KitFox

I haven't started.

I mean, I had notes from before I thought I would do it for this series.

but nothing new

Are you going to write something?

cornbread ninja 麵包忍者

No, 'cause he wanted it like, today.

KitFox

16:54

I know you can't for the series, but maybe write something anyway?

@cornbreadninja麵包忍者 Well, just prior to the day he would publish it.

cornbread ninja 麵包忍者

I started something in that vein a few weeks ago.

KitFox

Oh, but you're new, so he might have wanted more lead time.

KitSox

@AwalGarg Command !!norris' does not exist.

@<> "!!norris"

Awal Garg

@KitFox can you tell @KitSox to go to sandy?

cornbread ninja 麵包忍者

What's the room number?

Awal Garg

16:57

1

cornbread ninja 麵包忍者

!!summon 1

KitSox

@cornbreadninja麵包忍者 Registered; need 0 more to execute

Awal Garg

!!summon 1

KitFox

Did she go? I was distracted.

cornbread ninja 麵包忍者

She did.

@AwalGarg oh, duh.

Awal Garg

16:58

And she is behaving weirdly there...

cornbread ninja 麵包忍者

@Kit may I read your story?

KitFox

@cornbreadninja麵包忍者 Which one?

cornbread ninja 麵包忍者

@KitFox the one your friend hasn't.

KitFox

But yes, to all of them.

Oh, it's Spider. I think you've read it already, haven't you?

cornbread ninja 麵包忍者

I have. I rather liked it.

KitFox

17:13

Me too. It's one of my favorites.

cornbread ninja 麵包忍者

I wish the damn laundry would finish so I could put it away and leave for a while.

KitFox

I thought she'd like it too, but she's also really busy.

So she probably hasn't had time.

But also, it's not like she said she would read it. I just asked her to.

cornbread ninja 麵包忍者

17:32

I think my new maxi shrunk enough that I won't have to hem it!

KitFox

raises eyebrow

cornbread ninja 麵包忍者

Well, maybe.

KitFox

!!define maxi

KitSox

@KitFox maxi having a hemline at ankle length

KitFox

I am listening to The One Lilium asmr videos, getting my tingle on and feeling very, very lesbian today.

I mean, I can't help it. She totally wants me.

cornbread ninja 麵包忍者

17:39

raises eyebrow

This'n here.

I get my tingle on with Elvis Costello via earbuds.

KitFox

Oh. I have a dress like that. To me, maxi always meant menstrual pad.

I like the pattern on that.

cornbread ninja 麵包忍者

It's my first.

I should have said maxi dress.

Yay!

KitFox

It's good to know that's in fashion. I like that style.

Cerberus

Is she better than most?

cornbread ninja 麵包忍者

@KitFox they're on sale now for $25.

Oh, you mean ear tingle.

KitFox

17:43

I do mostly get ear tingles, although I realized last week sometime that I get a kind of eye tingle too.

And sometimes tingles make me sneeze.

cornbread ninja 麵包忍者

I thought you meant tingles in your southern hemisphere.

KitFox

Oh.

Oh.

Like when I see the Southern Cross for the first time.

Elvis Costello, huh? makes note

cornbread ninja 麵包忍者

My god.

Might not be your thing, I dunno.

KitFox

Um...what's a good length for a blurb?

Three lines of text at 12pt.

ish.

cornbread ninja 麵包忍者

Lorem ipsum dolor sit amet et cetera?

KitFox

17:47

How many characters do you suppose that is?

@cornbreadninja麵包忍者 Cute

Is that @Mr.Shiny's kid?

cornbread ninja 麵包忍者

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Nunc at ullamcorper tellus, vel iaculis ligula. Aenean massa magna, sodales vel eleifend id, aliquam ac mauris. Donec dolor risus, aliquet sed faucibus at, egestas convallis ipsum. Sed in dapibus enim. Curabitur convallis, tortor in varius pretium, massa erat cursus urna, sed malesuada purus libero pellentesque felis. Donec tellus quam, porttitor vel mi ornare, aliquet consectetur lorem.

@KitFox I hope so.

Cerberus

@cornbreadninja麵包忍者 Stop that!!

cornbread ninja 麵包忍者

@Cerberus O_O

Cerberus

May 27 at 1:01, by Cerberus

@ThatBrazilianGuy Speak proper Latin in this room please!

cornbread ninja 麵包忍者

I'm sorry!

Oh.

Cerberus

17:50

It's OK.

cornbread ninja 麵包忍者

Hey, I was just trying to help Kit.

Cerberus

Just don't let it rehappen.

KitFox

Thanks.

250 characters.

cornbread ninja 麵包忍者

Sheesh!

Ima go to the grocery store.

KitFox

►

There.

cornbread ninja 麵包忍者

17:55

!_!

KitFox

►

Some of these are good.

Mostly I like his accent.

Transcript for

English Language & Usage: Multi-Layer…