« first day (1366 days earlier)      last day (3571 days later) » 

3:00 PM
So I got a job at a security company
wtf I can't be a security guard
 
Congrats.
 
@KitFox Case?
 
@KitFox All this grepping is getting me all hot and bothered
 
@terdon I copypasted the things I wanted to match.
 
I might be able to help if you show me the input and regex,
 
3:03 PM
I can in a minute.
 
@KitFox what are you using to grep? cmd-line grep tool?
 
ugh… another black person got shot by a white person and it's the end of the world again in America.
 
I think she uses some kind of strange version, not the standard GNU grep.
 
it's Treyvon 2.0
 
!wiki grep
 
3:09 PM
who uses non-gnu grep?
O.o
 
@Mr.ShinyandNew安宇 Yes, and also AstroGrep.
 
!!wiki grep
What the hell is grepping.
 
@GeorgePompidou Cygwin, UNIX, BSD, OSX,
 
@KitFox because if you're using a command line grep you might have an issue with escaping/quoting the pattern so that the shell doesn't screw it up
 
Grep is a command-line utility for searching plain-text data sets for lines matching a regular expression. Grep was originally developed for the Unix operating system, but is available today for all Unix-like systems. Its name comes from the ed command g/re/p (globally search a regular expression and print), which has the same effect: doing a global search with the regular expression and printing all matching lines. == History == Grep was created by Ken Thompson as a standalone application adapted from the regular expression parser he had written for ed (which he also created). In ed, the command...
 
3:10 PM
@Mr.ShinyandNew安宇 Theoretically, putting it in a text file fixed that.
 
@terdon nobody on earth uses original unix grep, and osx is bsd grep
which functions the same as gnu grep.
I also was just being sarcastic.
and what the heck is cygwin?
 
@GnomeSlice like gripping, but with fewer hands
 
is that like smoking cigarettes when you're using windows?
 
@KitFox ... so you're sourcing your pattern from the text file?
 
3:12 PM
Oh it's a computer thing
 
@GeorgePompidou cygwin is a unix environment for windows.
 
you can search through text for a pattern
 
@Mr.ShinyandNew安宇 I know, I was also joking about that.
 
@Mr.ShinyandNew安宇 Yes.
 
@KitFox then hm.
 
3:14 PM
this is genius. I've been trying to get linkedin to stop spamming me for a few days, and just now read: "changes will take effect in a few days"
what's so hard about not emailing me starting today
 
yeah. I've never understood why companies do that
 
@GeorgePompidou No it doesn't unfortunately. It lacks all sorts of options. There are in fact a surprising variety of grep implementations, most of which are not GNU grep. Believe me, we get many questions about that on U&L.
 
hm. I've been using OS X and gnu grep the exact same for years now.
 
Underwear & Lingerie
 
I guess I just got lucky.
 
3:15 PM
@GeorgePompidou It's probably 90% CYA text in case there are emails in transit already. 10% chance that they have so much volume of mail they actually queue it up to send it out in the future, and they can't (or can't be bothered to) stop it once it's queued.
 
@Mr.ShinyandNew安宇 When I'm off the phone, Il post a sample.
 
@KitFox k. Do you have cygwin installed?
 
@Mr.ShinyandNew安宇 I thought of that, but I unsubscribed Tuesday from daily group update emails that get sent every morning at 10. and still got them up to today.
 
@Mr.ShinyandNew安宇 No.
 
@KitFox your problem is you need to uninstall Windows. it's a virus.
 
3:17 PM
@GeorgePompidou It's also possible they have a distributed system that doesn't push changes like that through very quickly. Which would be dumb, or at least annoying.
 
@GeorgePompidou Off the top of my head, OSX grep does not support -o which is among the most useful grep options.
 
it's just dumb. it's 100% possible to stop sending me emails within the day.
@terdon sure it does. I just did it.
 
@GeorgePompidou Huh, not according to its man page. How about -P?
 
one sec, let me switch to my mac and show you.
 
Perhaps you've installed GNU grep on your mac?
 
3:19 PM
> fi:~ george$ grep -V
grep (BSD grep) 2.5.1-FreeBSD
fi:~ george$ grep
usage: grep [-abcDEFGHhIiJLlmnOoPqRSsUVvwxZ] [-A num] [-B num] [-C[num]]
[-e pattern] [-f file] [--binary-files=value] [--color=when]
[--context[=num]] [--directories=action] [--label] [--line-buffered]
[--null] [pattern] [file ...]
fi:~ george$
I can also send you screenshot of manpage. it's nearly identical to gnu.
I may just do a diff on it.
but I have better things to do.
 
@GeorgePompidou Huh, OK. I've been reading the online manpages and they're different. I also answered a question where the OP was lacking some basic grep functionality in OSX. Anyway, you also have the other *nixes, busybox embedded systems etc, all of which have simpler grep implementations.
 
sure.
like I said, I know that there are other greps. I was just making a sarcastic reply to a "probably has some non standard gnu grep" comment
because I myself use freebsd grep more often than not.
and it's the same.
 
nuh uh. they have different version information
 
here's a unix question for you @terdon. how can I redirect a manpage to a file without it coming out all screwy?
 
watch me mace myself on the first day
 
3:24 PM
first day of what?
 
of Christmas, duh.
 
@GeorgePompidou screwy? Just man ls > file should work.
 
25 mins ago, by GnomeSlice
So I got a job at a security company
 
it doesn't. comes out all messed up. maybe it's because manpages don't get sent straight to output
 
@GnomeSlice I see! Good luck
 
3:25 PM
they work like more or less or something.
 
This is the sample: <head><a class="Mice" onclick="_gaq.push(['_trackEvent', 'button', 'searchMice', 'micedb']);" href="http://my.url/query/"></head><body id="top" class="js" style="font-size: 68.75%;"><a title="" target="_self" href="http://my.url/query"></body>
This is the regex: <body.*<[^<>]*href="http://my.url/query/*[^<>]*>
 
@GeorgePompidou No, you have the groff or whatever formatting stuff. You mean you see strange characters like [M or whatever?
 
nope. just same character multiple times a lot.
 
I'm still on the phone.
 
I got it.
man grep | col -b > ~/Desktop/mangrep.txt
^.^
there we go
 
3:27 PM
@GeorgePompidou Weird. man grep > foo works perfectly for me.
 
@KitFox you have links in <a in the head??!
 
@KitFox looks fine to me. It matches <body id="top" class="js" style="font-size: 68.75%;"><a title="" target="_self" href="http://my.url/query"> from your file.
 
Shit.
That's not at all what I meant to do.
 
where's your grepping prowess now Kit
chuckles
 
Well, it still should have matched!
 
3:33 PM
What is it you actually want to match?
 
It matches for me in jEdit. I'm going to try cmd line now. But you have query/*, is that intentional?
 
I just got off the phone. Let me spend a moment orienting.
 
The hone?
 
what's a hone?
 
She was honing here regex skills.
 
3:34 PM
ooh! she meant a phone!
 
I like hone.
 
$ grep -E '<body.*<[^<>]*href="http://my.url/query/*[^<>]*>' test.txt
<head><a class="Mice" onclick="_gaq.push(['_trackEvent', 'button', 'searchMice', 'micedb']);" href="http://my.url/query/"></head><body id="top" class="js" style="font-size: 68.75%;"><a title="" target="_self" href="http://my.url/query"></body>
It matches for me with cygwin grep
my test.txt has the whole text in one line, but still.
if I reformat the html like so:
<head>
  <a class="Mice" onclick="_gaq.push(['_trackEvent', 'button', 'searchMice', 'micedb']);" href="http://my.url/query/">
</head>
<body id="top" class="js" style="font-size: 68.75%;">
  <a title="" target="_self" href="http://my.url/query">
</body>
it no longer matches
 
Hmm. That must be it then.
Something with whitespace or newlines?
Here is the revised test text: <div id="header"><a class="Mice" onclick="_gaq.push(['_trackEvent', 'button', 'searchMice', 'micedb']);" href="http://my.url/query/"></div><div id="envelope"><a title="" target="_self" href="http://my.url/query/f?q"></div>
And revised pattern: <div id="env.*<[^<>]*href="http://my.url/query/*[^<>]*>
It should match the second url and not the first.
 
grep doesn't match across newlines
 
I hope nobody comes by my cubicle because I farted.
I said, as the CEO of the company pops up to talk to the guy next to me.
 
3:40 PM
@Mr.ShinyandNew安宇 I don' t think I knew that.
 
All such tools are line-based.
 
I wonder if I can count on their inconsistency.
 
@KitFox There are variants that support it. or other tools. regexes aren't inherently restricted to single lines
 
If you have access to basic *nix tools, you can simply remove the newlines: tr $'\n' ' ' < file | grep PATTERN
Or perl -pe 's/\n/ /' file | grep PATTERN
 
There's a link in the header section that is to be replaced. I want to find all the other pages where the link appears, but not in the header section (because that is on all pages).
 
3:43 PM
can you count matches?
 
Oh, yeah, I could do that. It's a lot of content though, and I don't want to scrape it again.
 
are there pages where the link ISN'T in the header?
 
@Mr.ShinyandNew安宇 I could, but it returns so large that the file barfs.
@Mr.ShinyandNew安宇 No.
 
because you could just find all the pages where count >= 2
$ grep -E 'http://my.url/query' -c test.txt
2
 
Oh god, it's worse than I thought.
 
3:46 PM
$ cp test.txt test2.text
$ grep -E 'http://my.url/query' -c test.txt test2.text
test.txt:2
test2.text:2
 
did you guys know that grep is a noun and verb in the Oxford dictionaries?
 
It's matching "...query/" when I put ..."query" as the pattern. That's annoying.
This is an impossible frigging task.
 
grep shows matching lines
so besides identifying files that need changing, what do you want to achieve?
 
All this grepping is making me frisky.
 
frisk frisk
 
3:58 PM
Well, after all, frisking does involve some groping.
 
OK, here's the new plan. Tell me if it's stupid.
Step 1. Map the old query url to something else.
Step 2. Put a permanent redirect to the old query's original url to point it to the new search url.
I love how talking it through makes it apparent that I have wasted a couple of hours trying to figure out something that it is not my job to figure out.
 
for Step 1 do you mean "make a new search url" and Step 2 "rewrite the old search url to match the new one"?
That is a reasonable thing to do when changing the incoming urls is impractical.
 
No, I just realized that the dev has to solve this problem, not me.
But the requirement is "replace the old search with the new search, but keep access to the old search".
 
yeah you can just make a requirement that says "all the old pages must still work, somehow"
what does "keep access" mean?
 
So if you have the old site bookmarked, you should redirect to the new search instead, but there is a link from the new search back to a place where you can use the old search if you prefer it.
Is the plan.
 
4:05 PM
ah
does the new search use the same query syntax as the old?
 
No.
 
or rather, will the new search produce sensible results for old queries?
 
The new search is completely different.
 
hm.
I'd leave the old search alone and just update pages to point to the new search. no mapping of anything.
but I don't know how many pages you need to touch to make that happen.
 
The old search used a cobbled together query for an Oracle database. The new one uses facets and things to search against a mongo db.
Updating the pages looks like it could be a massive undertaking.
 
4:07 PM
yay! my boss said I can leave early today for my drive to Boston ^.^
is excited to see his best friend
 
Can you translate the old queries into new ones?
 
jiggles around in office chair excitedly
 
@Mr.ShinyandNew安宇 I don't think so. It uses an Apex form or something. I can understand that it is sending parameters, but it makes no sense to me.
 
I don't know what an Apex form is. But if there were some way to algorithmically take the input to one query and translate it into the input for the new query then you could replace the old search with the new, and on the results page put a link to the (moved) old search page.
Then the old pages don't need to change but they get the new search.
Though I have to wonder what the point is. Is the old search being retired? Deprecated? Is it just too slow to use? etc.
 
No, it's not like that. It's "Want to search for mice? Click this link."
Embedded randomly into various pages.
The old search bites ass.
So we're replacing it with a new search application.
One with business logic!
 
4:11 PM
Bites ass?
 
Anyway, we launched the beta version last night.
 
Are the pages where the search links appear generated, or static?
 
static
 
@KitFox Hey, is your internet working today?
 
user116848
Man, it's very depressing today. I just got my result and got failed in one of the papers :(
 
4:13 PM
You must check out today's music, it's so good. soundcloud.com/nudiscoyourdisco/holic-santiago-wmc-2014
 
user116848
Hi folks
 
Hi @Arrowfar. Sorry to hear it.
@GnomeSlice I'll check it out. Thanks.
 
user116848
Yeah, me too. I hope I clear it next time.
 
@Arrowfar boo! Well, don't give up.
 
You only failed one, right? So that's many you did not fail.
 
user116848
4:15 PM
@KitFox, @Mr.ShinyandNew安宇 Yeah thanks guys.
 
user116848
Failing sucks :(
 
user116848
BTW it was one of the ACCA papers.
 
caca papers? I think I would want to fail those.
 
user116848
Yeah. It is one of those accountancy qualifications (UK one)
 
4:18 PM
tata papers, on the other hand...
 
user116848
They have 14 papers in it and I have cleared 10 of them so far. 4 to go. It has been almost 6 years. I am that slow, huh?
 
user116848
Normal people usually take four to five years to complete it.
 
Eh, so you have a life.
 
Heh
 
user116848
Yeah I do.
 
4:24 PM
normal people take 6-7 hours to drive to Boston. my record is 5.5
 
Oh ffs, astrogrep gives me the option of specifying a number of occurrence anyway.
 
but I did get a speeding ticket when setting the record.
I'm going to try to beat it this time.
 
I thought you lived in Chicago.
 
nope. upstate NY for school.
 
Oh.
 
4:25 PM
Too much time spent on the internet @Arrowfar ;-)
 
I need a more fuel efficient car :<
 
Maybe you know Sara and ... the guy she married back in 2001 or 2 or so.
If you live in Troy anyway.
 
user116848
@skullpatrol Yeah I think so. I will have to watch my time carefully.
 
haha. nope, Rochester.
I almost went to RPI in Troy.
 
Oh, then you know Stephan and Marlene.
And Dayeol.
 
4:26 PM
@Arrowfar time is money.
 
my neighbor's name is Stephen.
Kit, I'm starting to think there are all of these connections because you just know someone everywhere in every situation.
 
Is he French Canadian?
 
he's Korean and Colombian.
close. but not quite.
 
Close enough.
 
jincks
 
4:27 PM
Then it's not him.
 
user116848
@skullpatrol You are right.
 
know anyone at Tufts?
or in the greater Boston area?
in particular people from Newton or Chicago
 
Yes.
Relations.
Friends and former friends.
 
do you know any gay puerto rican black supremacists who are convinced that the "upper middle class white imperialist men" are out to get them?
 
Oh, I forgot Stephan is at URMC.
@GeorgePompidou Just the one, but she's a woman.
 
4:29 PM
sigh
 
And she lives somewhere near Miami Beach.
 
they're everywhere!
 
Oh and looks like Daeyeol is at Yale now.
 
that's a silly name.
and an appropriate name of a school for the name.
does he drive a Huayra too?
 
It's Korean. Pretty common too, iirc.
 
4:31 PM
that'd be funny. Daeyeol drives a Huayra at Yale.
all impossible things to say.
 
I don't get why that's impossible to say.
 
it's literally impossible.
 
user116848
Bye guys. I should go and study
 
Good luck.
 
literally 100% of humans are completely unable to say any of those three words.
 
4:34 PM
This is so goddamned asinine that I can't find anything on this effing site even if I have an idea what the effing directory structure looks like.
 
what is asinine, and what are effing sites and directory structures?
how does one eff?
 
It's fucking without you in it.
 
hm.
update on my position on votes and points: I now officially care about reputation points temporarily until further notice.
I recently found that at a certain reputation, I can edit posts without needing approval. I want this a lot.
goes to answer easy questions for extra rep
 
4:53 PM
Here is a page that I know is in the research directory. I'm on the main page, and I figure, oh, must be this link that says "Research". Nope. That takes me to the research subdomain. Turns out, I wanted the link that said "tell me more".
 
Hmm America is bombing the Islamic crazies.
I guess that is good.
 
My other favorite one is the "Cre Mice by site of expression" button that takes you to a page that tells you about the cre mouse repo with a hyperlink to lists of cre strains, which if you look there, gives you several more hyperlinks of the various kinds of cre tools, and then if you are very nice, you might click through to a list that actually contains the site of expression. But not always.
 

« first day (1366 days earlier)      last day (3571 days later) »