« first day (1753 days earlier)      last day (2067 days later) » 

12:06 AM
[ charcoal-se.github.io ] continuous-integration/travis-ci/push success on 93b68b7: The Travis CI build passed
 
12:18 AM
[ SmokeDetector | MS ] Toxic answer detected: What did Balin see in the Mirrormere? by not big gay on scifi.SE (@Mithrandir)
 
General heads-up: Hacktoberfest's coming up again - if you can think of anything that first-time contributors could do, feel free to throw an issue for it (and/or take ownership of helping first-timers through it)
10
 
CI on 951430d succeeded.
 
J F
@ArtOfCode ooh, they upped the ante — you need 5 PRs this time instead of 4
 
12:37 AM
@ArtOfCode I thought about dev docs some time ago and I heard statements that 30% of dev time are writing docs (missing details).
 
yup, docs are good, but unless it's a relatively simple system, we should be the people to write them
you can have great intentions, but without intimate knowledge of the system you're documenting you won't get awesome docs
 
I'd be happy to help with the docs because when school work starts I won't have as much time for debugging code.
just for clarification, I didn't mean "make the docs hactoberfests", but that "we create the docs to help hacktoberfests"
 
Ah
then yes, that would be awesome :)
 
Thanks for the affirmation, I'll throw up a gh-pages branch for that.
 
@iBug There is a fair amount of documentation already in the Smokey wiki, but it's a bit patchy and varied (mostly because it's been written in parts by multiple different people). If you're thinking GH pages, converting and standardising that too would be :D
 
12:46 AM
@ArtOfCode Isn't that documentation for using instead of developing?
 
@thesecretmaster Here you go...
# Logfile created on 2018-09-04 22:50:34 +0000 by logger.rb/61378
I, [2018-09-04T22:50:34.546817 #7252]  INFO -- : ================= BEFORE =================
I, [2018-09-04T22:50:34.546864 #7252]  INFO -- : #<Redis client v3.3.5 for redis://127.0.0.1:6379/0>
I, [2018-09-04T22:50:34.546905 #7252]  INFO -- : Context: ci/circleci: rails-tests
I, [2018-09-04T22:50:34.546920 #7252]  INFO -- : State: success
I, [2018-09-04T22:50:34.546930 #7252]  INFO -- : SHA: 5248bf4fa905cfb539d6bd85e94aad6ac3cd8da3
 
Some of it is, but the "Developer Information" section is for people working on Smokey
 
Then I'd just stick to the wiki.
 
@Undo Oooohhhhh I can explain where all the keys went
 
Oh yeah, I see it.
 
12:48 AM
I set the TTL to 10 minutes not 24 hours. Whoopsie.
 
basically the same ;P
 
Oh wait gawd darnit
typing
Redis returns strings
 
[ charcoal-se.github.io ] continuous-integration/travis-ci/push success on 93b68b7: The Travis CI build passed
[ metasmoke ] ci/circleci: rails-tests failure on a78e531: Your tests failed on CircleCI
[ metasmoke-deploy ] deploy started by thesecretmaster
 
1:07 AM
haah... reason weight for old posts can change??
 
Yep
 
Was this implemented recently?
 
Since forever; byproduct of system design
It's not awesome, but my reviews have said it doesn't impact history much
 
I recall that some time (~2mo) ago I saw old reason weights on old posts.
 
nope
 
1:11 AM
Bleh tests got stuck again :'(
 
[ SmokeDetector | MS ] Bad keyword in answer, bad keyword with email in answer, pattern-matching email in answer: Explaining Horizontal Shifting and Scaling by standard roll on math.SE
[ metasmoke ] ci/circleci: rails-tests failure on 0d8b63a: Your tests failed on CircleCI
[ metasmoke ] ci/circleci success on 9ceebfb: Your tests passed on CircleCI!
 
e.g., I recall seeing 4 for "misleading link" instead of 9 on years-old posts, in June or July (2018)
 
@SmokeDetector @Undo Now CI will only post when either ALL tests pass or on each failing test. Less spaaaam!
 
The highest-weight post has grown from 1280 to 1286
 
1:15 AM
@thesecretmaster \o/
Now to do away with the failing tests
 
What failing tests?
 
The mysterious too-long-with-no-output ones
afk
 
Oh
I hate those
 
[ SmokeDetector | MS ] Bad keyword in body: How to install the EPSON L210 utility by user3689783 on askubuntu.com
[ metasmoke ] ci/circleci success on d346230: Your tests passed on CircleCI!
 
that was because MySQL container exploded?
@SmokeDetector f
 
1:18 AM
[ metasmoke ] ci/circleci success on d017af4: Your tests passed on CircleCI!
 
Ugh I hate it when I push stuff accidentally
@Undo (not urgent) I accidentally pushed some changes to package.json and environments/dev.rb. I'm reverting the dev.rb ones because they matter, but the package.json changes seem to be an objective improvement. Mind if I keep them?
 
CI on 63e9ab4 succeeded.
[ SmokeDetector | MS ] Body starts with title and ends in url: cheap Avon Shampoo Film In Roll by qieb111 on unix.SE
 
@SmokeDetector k
!!/watch chengdetechnology\.com
 
@iBug Added chengdetechnology\.com to watchlist
Restart: API quota is 11455.
 
1:43 AM
@tripleee Domain records/tags seems fixed now.
 
CI on c6003ec succeeded.
 
@Andy You can find roof repair services on metasmoke :)
 
2:00 AM
[ metasmoke ] ci/circleci: rails-tests failure on 5646d1c: Your tests failed on CircleCI
 
@SmokeDetector Hiss
How does mysql stop existing for half of the tests
 
@thesecretmaster Yeah, that's easier to read
 
[ metasmoke-deploy ] deploy started by thesecretmaster
 
Once this deploy finishes, feel free to delete log/circleci.log
 
!!/unwatch termsfeed\.com
 
@iBug Removed termsfeed\.com from watchlist
 
2/19 TP/All rate
 
!!/watch responsivefilemanager\.com
 
2:19 AM
@iBug Added responsivefilemanager\.com to watchlist
[ metasmoke ] ci/circleci success on cbe4116: Your tests passed on CircleCI!
 
Deploying again for testing
Dunno where the "I'm deploying now" message went, but @Undo notice anything different about this travis run?
 
Nice
That'll be convenient
 
I've really used way to many commits to do anything recently... but at least I've been doing stuff.
 
@thesecretmaster Doesn't really matter to me
 
2:33 AM
Eeeh... I like my git history to be neat and I've made a bit of a mess. No tangible problem, but ideally I wouldn't have done that.
 
@thesecretmaster checkout -b -> rebase -> merge
 
I like to cap off big commit streaks with a note in a message showing which commits are in a logical group. Makes it easy to diff by feature changes if someone wants to
 
that's what branches are meant for
 
@iBug Yes, but that doesn't work when you need to test things that impact CI or other prod-only bits.
 
[ SmokeDetector | MS ] Potentially bad keyword in title: vue nuxt, passing API value on header title SEO by Reza Briliantama on stackoverflow.com
fp- by iBug
 
2:51 AM
[ SmokeDetector | MS ] Bad keyword with a link in answer, potentially bad keyword in answer: How long is the longest path in the game tree of chess? by mayduavongts on mathoverflow.net (@FrenzyLi)
tpu- by iBug
 
!!/watch anthaison\.vn
 
@iBug That pattern looks like it's already caught by Potentially bad keyword in answer and Potentially bad keyword in body; append -force if you really want to do that.
 
3:03 AM
[ SmokeDetector | MS ] Offensive answer detected, toxic answer detected: What are the differences between types of anime? by JO and kian on anime.SE
tpu- by Tetsuya Yamamoto
tpu- by Tetsuya Yamamoto on Explaining Horizontal Shifting and Scaling [MS]
 
!!/test-a . .
 
> Would not be caught as a post, title or username.
:46588160 > Would not be caught as a post, title or username.
> Mostly dots in body, mostly dots in title
----------
Title - Post contains 2 dots out of 3 characters
Body - Post contains 2 dots out of 3 characters
> Mostly dots in answer, non-latin link in answer
----------
Body - Position 58-65: >. .</a>
Body - Post contains 2 dots out of 3 characters
 
3:21 AM
[ charcoal-se.github.io ] continuous-integration/travis-ci/push success on 93b68b7: The Travis CI build passed
Recovered from json.decoder.JSONDecodeError: Expecting value: line 1 column 1 (char 0)
 
!!/errorlogs
 
File "/usr/lib/python3.6/threading.py", line 864, in run
self._target(*self._args, **self._kwargs)

File "/home/smokey/SmokeDetector/deletionwatcher.py", line 54, in _start
self.socket.send("hb")

File "/usr/local/lib/python3.6/dist-packages/websocket/_core.py", line 234, in send
return self.send_frame(frame)

File "/usr/local/lib/python3.6/dist-packages/websocket/_core.py", line 259, in send_frame
l = self._send(data)

File "/usr/local/lib/python3.6/dist-packages/websocket/_core.py", line 423, in _send
 
not our fault, Google Perspective again.
 
3:45 AM
[ charcoal-se.github.io ] continuous-integration/travis-ci/push success on 93b68b7: The Travis CI build passed
CI on c78eaff failed.
 
!!/pull
 
[ SmokeDetector ] Blacklists reloaded at rev b0027bc (iBug: Flake8 --autopull) (running on teward/Solar Flare)
 
4:06 AM
!!/blacklist-website-force healthsuppliment4diet\.com
 
@tripleee That pattern looks like it's already caught by Pattern-matching website in body and Pattern-matching website in answer; append -force if you really want to do that.
@tripleee Blacklisted healthsuppliment4diet\.com
[ SmokeDetector | MS ] Bad keyword in body, blacklisted website in body, pattern-matching product name in body, pattern-matching website in body, potentially bad ns for domain in body, +1 more: Think how it would feel to try sailing by Samahuman on drupal.SE
tpu- by Makyen
 
@iBug hmm, I can still repro the traceback? did something change?
!!/blacklist-website-force healthyorderzone\.com
 
@tripleee That pattern looks like it's already caught by Blacklisted website in answer, Pattern-matching website in body, Pattern-matching website in answer, and Blacklisted website in body; append -force if you really want to do that.
@tripleee Blacklisted healthyorderzone\.com
 
4:25 AM
[ SmokeDetector | MS ] Bad keyword in body, bad keyword in title, bad keyword in username, bad pattern in url body, blacklisted website in body, +3 more: Phenterage Garcinia - Improved Energy Level For Weight Loss by PhenterageGarciniaReview on meta.SE
tpu- by Makyen
 
4:38 AM
[ SmokeDetector | MS ] Link at end of answer, potentially bad keyword in answer: Is there a quick way I can download all these images? by 카지노사이트 on stackoverflow.com
tpu- by tripleee
 
4/4, one more and we blacklist
@JoErNanO do you remember why b-o-s-s-c-y-b-e-r was watched? The watch entry had Unicode dashes, I assume that was a mistake?
 
[ SmokeDetector ] Blacklists reloaded at rev f473d0a (tripleee: Oh duh, two more) (running on teward/Solar Flare)
 
5:07 AM
[ SmokeDetector | MS ] Url in title, bad keyword in body, bad keyword in title, blacklisted website in body, blacklisted website in title, +3 more: curlycrayy.com/phendora-garcinia/ by user48233 on rpg.SE
tpu- by tripleee
[ SmokeDetector | MS ] Url in title, bad keyword in body, bad keyword in title, blacklisted website in body, blacklisted website in title, +3 more: curlycrayy.com/phendora-garcinia/ by Tamsea49 on rpg.SE
tpu- by Makyen
 
5:28 AM
[ SmokeDetector | MS ] Bad keyword in link text in body: How To Change Backgroud color with color gardients on my website by Anita Saini on stackoverflow.com
tpu- by suraj
fp- by suraj
 
@SmokeDetector handled by a mod
didnt even get time to retract it :( fast
 
[ SmokeDetector | MS ] Bad keyword in body, bad keyword in title, blacklisted website in body, pattern-matching product name in body, pattern-matching website in body, +1 more: Praltrix South Africa : Male Enhancement Reviews by user48235 on rpg.SE
[ SmokeDetector | MS ] Potentially bad ns for domain in body: Where I can find the Updated (2018) list of NASDAQ 100 Stocks market watch? by user940801 on superuser.com
[ SmokeDetector | MS ] Shortened url in answer: Algorithms and techniques for spell checking by naveen ghanathay on datascience.SE
[ SmokeDetector | MS ] Url in title, bad keyword in body, bad keyword in title, blacklisted website in body, blacklisted website in title, +6 more: healthyskylane.com/ambrosina-skin-cream/ by Lere1940 on drupal.SE
tpu- by suraj
 
5:48 AM
@quartata were you ever thinking of using p values to refine the luncheon meat rule
 
!!/watch apnacoupon\.in
 
that would mean figuring out the mean and standard deviation of the probabilities across all SE posts
 
@tripleee Added apnacoupon\.in to watchlist
 
and then using statistics to determine how likely a certain paragraph is in respect to it's length
 
!!/watch- capitalone\.onl
!!/watch- natwest\.onl
!!/watch- turbotax\.onl
 
5:56 AM
@Makyen That pattern looks like it's already caught by Potentially bad keyword in answer and Potentially bad keyword in body; append -force if you really want to do that.
 
!!/watch- amzn\.to/(?:2Ck80by|2wHBOtq|2NiXtBN|2Q5Stif)
 
!!/watch-force turbotax\.onl
 
@Makyen Added turbotax\.onl to watchlist
[ SmokeDetector ] Blacklists reloaded at rev 0fff2c3 (SmokeDetector: Auto watch of turbotax\.onl by Makyen --autopull) (running on teward/Solar Flare)
tpu- by Glorfindel on Why is mysql such a piece of shit? [MS]
 
@Makyen Watching this mostly to keep a record that we've seen the domain. "turbotax" by itself is watched.
 
5:59 AM
[ SmokeDetector | MS ] Potentially bad ns for domain in answer: More reliable alternative to Samba on Linux, speed is most important ✏️ by athena on softwarerecs.SE
[ SmokeDetector | MS ] Bad ns for domain in body, bad keyword in body, bad keyword in title, blacklisted website in body, link at end of body, +4 more: Purefit Keto best weight burn by user401351 on meta.SE
tpu- by tripleee
[ SmokeDetector | MS ] Pattern-matching website in body: Is Taking a Loan Against Property a Good Idea? by retailloans on superuser.com
tpu- by tripleee
 
!!/watch-force retailloans\.in
 
@iBug Added retailloans\.in to watchlist
[ SmokeDetector ] Blacklists reloaded at rev 39d2266 (SmokeDetector: Auto watch of retailloans\.in by iBug --autopull) (running on teward/Solar Flare)
 
@micsthepick p values?
 
@iBug statistical term
In statistical hypothesis testing, the p-value or probability value or asymptotic significance is the probability for a given statistical model that, when the null hypothesis is true, the statistical summary (such as the sample mean difference between two compared groups) would be the same as or of greater magnitude than the actual observed results. The use of p-values in statistical hypothesis testing is common in many fields of research such as physics, economics, finance, political science, psychology, biology, criminal justice, criminology, and sociology. Their misuse has been a matter of...
 
[ SmokeDetector | MS ] Bad keyword with a link in answer, pattern-matching website in answer: Is there a way to have High Contrast mode and Full Screen Zoom enabled simultaneously? by dam ponting on superuser.com
 
6:08 AM
So how to put that into practice?
 
tpu- by suraj
[ SmokeDetector | MS ] Pattern-matching website in answer, potentially bad keyword in answer, username similar to website in answer, blacklisted user: Is Taking a Loan Against Property a Good Idea? by retailloans on superuser.com
[ SmokeDetector | MS ] Bad keyword in body, bad keyword in title, blacklisted website in body, body starts with title and ends in url, link at end of body, +2 more: Phendora Garcinia The tips by mtfwkusrc on drupal.SE
tpu- by tripleee
 
@iBug not really sure what @micsthepick is driving for but the statistics are not really the problem, it's that the luncheon meat it finds is often not actually a spam indication
 
!!/watch stockmaster\.in
 
@TetsuyaYamamoto Added stockmaster\.in to watchlist
 
I don't see how we can calculate the p value for a specific input
 
!!/blacklist-website-force mobisoftinfotech\.com
 
@TetsuyaYamamoto That pattern looks like it's already caught by Pattern-matching website in body and Pattern-matching website in answer; append -force if you really want to do that.
@TetsuyaYamamoto Blacklisted mobisoftinfotech\.com
[ SmokeDetector | MS ] Bad keyword in body, bad keyword in title, body starts with title and ends in url, link at end of body, no whitespace in title, +1 more: Stendoragarciniasouthafrica by Getede Dere on meta.SE
tpu- by Tetsuya Yamamoto
 
@iBug the p value would be in comparison with some expectation, that's by and large what the function does so I'm guessing the proposal would be to refine the model it uses
 
!!/blacklist-website-force supplementssouthafrica\.co\.za
 
6:20 AM
@tripleee That pattern looks like it's already caught by Pattern-matching website in body and Pattern-matching website in answer; append -force if you really want to do that.
@tripleee Blacklisted supplementssouthafrica\.co\.za
 
!!/test capitalone.onl
 
> Potentially bad keyword in body, potentially bad keyword in title, potentially bad keyword in username
----------
Title - Position 1-14: capitalone.onl
Body - Position 1-14: capitalone.onl
Username - Position 1-14: capitalone.onl
 
well, we can find a t value
I think
an approximation
we should already have the ability to make a comparison
so we just divide the difference by the square root of the length of the string
and that should give a value adjusted for different lengths
now, we might instead need to divide by log(n)/2 if @quartata s values are log adjusted
(which I think they are)
no
subtract?
it's a bit more complicated
 
[ SmokeDetector | MS ] Blacklisted website in body, body starts with title and ends in url, link at end of body, pattern-matching product name in body, pattern-matching website in body: Dr. Harry Preuss, M.D., a systematic by voinlscq on askubuntu.com
tpu- by Makyen
 
13
A: How to determine if character sequence is English word or noise

Brandon LoudermilkDuring NLP and text analytics, several varieties of features can be extracted from a document of words to use for predictive modeling. These include the following. ngrams Take a random sample of words from words.txt. For each word in sample, extract every possible bi-gram of letters. For exampl...

Found this, and I think it could be easier to adapt
 
6:31 AM
[ SmokeDetector | MS ] Bad keyword in body, bad keyword in username, potentially bad keyword in username: Peoples mind is to get rich in every chance you have to be by Apollo Hospital on workplace.SE
tpu- by Glorfindel
 
@iBug but again, luncheon meat already does what it's supposed to do; it's just not a very good indicator of spamminess
@SmokeDetector do we really need #organ-trafficking?
it already exists so easy to apply, but this is crazy
!!/watch [a-z_]*(?:91_)?9343518612[a-z_]
 
@tripleee That pattern looks like it's already caught by Bad keyword in body and Bad keyword in answer; append -force if you really want to do that.
[ SmokeDetector | MS ] Bad keyword in body, bad keyword in title, blacklisted website in body, body starts with title and ends in url, link following arrow in body, +2 more: Rose Diamond Australia : obviously, everyone needs to keep up the first by smither on graphicdesign.SE
tpu- by tripleee
 
!!/unwatch apollo\W*(?:kidney\W*)?hospital\d*
 
@tripleee Removed apollo\W*(?:kidney\W*)?hospital\d* from watchlist
 
duplicated by a somewhat more complex blacklist
 
6:38 AM
[ SmokeDetector | MS ] Url in title, bad keyword in body, bad keyword in title, blacklisted website in body, blacklisted website in title, +3 more: curlycrayy.com/phendora-garcinia/ by user48239 on rpg.SE
tpu- by John Dvorak
[ SmokeDetector | MS ] Url in title, bad keyword in body, bad keyword in title, blacklisted website in body, blacklisted website in title, +5 more: www.nutritionfit.org/retro-lean-forskolin/ by Notee1960 on drupal.SE
tpu- by Glorfindel
[ SmokeDetector | MS ] Username similar to website in answer: www and non www resolution issues in my htaccess file by Book Rivers on stackoverflow.com
naa- by Tetsuya Yamamoto
 
!!/watch bookrivers\.com
 
@Glorfindel Added bookrivers\.com to watchlist
 
That user created an account on 79 sites. I'm actually quite sure it's a spammer.
 
not on aviation :'(
 
6:52 AM
Google results for this domain don't look like they are into illicit promotion, though of course it's possible that they are just starting
but creating a lot of accounts could just be stup... uninformed behavior
 
@Federico Fortunately not... I just sneaked here.
 
@TetsuyaYamamoto yeah, saw that, that's why I posted what I posted :P
 
7:12 AM
[ SmokeDetector | MS ] Bad keyword in body, bad keyword in title, blacklisted website in body, blacklisted website in title, link following arrow in title, +4 more: read more supplementforus.com/retro-slim-forskolin/ by guri zondal on askubuntu.com
tpu- by Glorfindel
[ SmokeDetector | MS ] Bad keyword in body, bad keyword in title, blacklisted website in body, body starts with title and ends in url, pattern-matching website in body, +1 more: Patriot Power Greens But this does not by user10318716 on stackoverflow.com
tpu- by Glorfindel
[ SmokeDetector | MS ] Link at end of answer: Solder paste/flux when desoldering by heat gun? by funnyjokes on electronics.SE
tpu- by Glorfindel
 
!!/watch subway-surfers\.online
 
@TetsuyaYamamoto Added subway-surfers\.online to watchlist
 
the actual way to accommodate for changes of size due to length is just to divide by n
since the log probabilities are added together
it would be interesting to run a n-gram checker on sections of each post, and raise concern if there are no sections that look like english
 
[ SmokeDetector | MS ] Url-only title, bad keyword in body, bad keyword in title, blacklisted website in body, blacklisted website in title, +6 more: www.nutritionfit.org/retro-lean-forskolin/ by Histlenrer on askubuntu.com
tpu- by Tetsuya Yamamoto
 
or, important sections that don't look like it
 
7:27 AM
[ SmokeDetector | MS ] Url-only title, bad keyword in body, bad keyword in title, blacklisted website in body, blacklisted website in title, +5 more: greenlifenutrition.org/praltrix/ by vierer201 on askubuntu.com
tpu- by Tetsuya Yamamoto
 
so we would be able to catch gibberish that looks like this: asdbj vnioqdn bowixng qpenxogn
 
API quota rolled over with 9290 requests remaining. Current quota: 19999.
stackoverflow: 425
math: 356
askubuntu: 125
superuser: 119
physics: 71
unix: 61
english: 52
serverfault: 51
apple: 51
electronics: 34
security: 31
drupal: 26
salesforce: 26
gaming: 22
mathoverflow.net: 21
android: 21
magento: 21
academia: 19
worldbuilding: 18
es.stackoverflow: 18
ja.stackoverflow: 18
dba: 18
gis: 17
wordpress: 17
meta: 16
ell: 16
webapps: 16
diy: 15
sharepoint: 14
ethereum: 14
interpersonal: 13
travel: 13
bitcoin: 12
graphicdesign: 12
movies: 11
datascience: 11
puzzling: 11
hinduism: 11
 
instead of just looking for repeated characters and low amounts of unique characters
of course, things like rot-n would be caught
 
[ SmokeDetector | MS ] Bad keyword in body, bad keyword in title, blacklisted website in body, blacklisted website in title, potentially bad ns for domain in body, +2 more: biorocketblasttry.com/phendora-garcinia/ by user48240 on rpg.SE
 
maybe that should be a separate rule then
 
7:30 AM
tpu- by Tetsuya Yamamoto
Restart: API quota is 19982.
 
I think so
 
[ SmokeDetector | MS ] Bad keyword in body, bad keyword in title, blacklisted website in body, link at end of body, no whitespace in title, +1 more: lutragencreamSkin by user401363 on meta.SE
tpu- by Glorfindel
 
@micsthepick rot-n?
While that could be useful, performance would be an issue, then.
It's not that practical.
Writing a Python C plugin, of course, would solve the performance issue, but
 
7:47 AM
why is that a performance issue?
 
We're running n-gram check against a lot of posts
That's much more calculation than simply summing up 1-gram
 
[ SmokeDetector | MS ] Pattern-matching website in body: Get all state phone code in a country ✏️ by Bunthai Deng on stackoverflow.com
 
our regexes probably rescan strings many more times than this
 
fp- by Tetsuya Yamamoto
 
it is O(n) for whatever you want to check
 
7:53 AM
Am I correct in believing that the issue with AIM reporting multiple feedbacks from the same person when that person quickly gives feedback has become significantly more common in the last couple/few days? To me it seems to be happening about an order of magnitude more often (or more) than it used to. I'm specifically asking about the duplicated feedbacks seen when someone provides feedback shortly after the report was posted.
I ask because it's caused by a race condition. The change indicates that either people are giving feedback faster, or MS is taking longer to send the WebSocket data indicating that the autoflags have been placed. Hmmm... maybe caused by the recent changes in SD to decouple reporting to MS.
 
regex is usually O(n^2) or worse
at at best, it is also O(n)
basically, you just add integers
figure out the log probability, multiply by some number and cast to an int
and you only have to figure out each probability once
then it is just a simple lookup from a table
@iBug actually, it is better than 1-gram because there are less of them
(in a given string)
the lookup table of course grows exponentially with n-gram size
@iBug rot-n is a cipher, I suppose any encoded text would also be picked up
 

« first day (1753 days earlier)      last day (2067 days later) »