« first day (1398 days earlier)      last day (3539 days later) » 
00:00 - 17:0017:00 - 23:00

5:00 PM
@FaheemMitha yup. Thanks for the compliment about the bowtie.
 
@MaxVernon You're welcome.
I think of it as a British thing.
 
@FaheemMitha, it would be a interesting analysis on predicting how the upvotes on a question work.
 
@Ramesh We should lay bets.
 
@FaheemMitha I was born in Britain, but I call myself Canadian.
 
@FaheemMitha, I would be losing much more than I lose in Vegas.
 
5:02 PM
@MaxVernon Ah, well, that explains the tie then. :-)
@Ramesh :-)
 
50 bucks this ends in the hot list
 
@Braiam I'm not betting against that.
 
I have posted my question, I hope I have everything needed: unix.stackexchange.com/questions/149049/…
 
This is a dumb question, but what does non-wiki answers mean for stack exchange?
 
5:16 PM
@Braiam if someone is good at porting you can use an OSX program called soundflower
the source is on github too in ma++ github.com/Cycling74/Soundflower
 
5:31 PM
It seems google pretty much forces one to use gmail for email accounts when using google accounts. is that accurate?
 
looks like it
you can change the address in your google account but they want you to have a gmail as well
@FaheemMitha I am wrong, check this out accounts.google.com/SignUpWithoutGmail
so you can't use interactive with force for rm?
 
I'm trying to submit an issue for chromium, but nothing is happening at the last step when I click the submit button. Frustrating.
 
@FaheemMitha So you tried that form I linked right?
 
Maybe this web page doesn't like Chromium. Maybe I should try Firefox instead.
@NoTime What form?
 
the create account without gmail form
 
5:46 PM
@NoTime I'm not sure what to do if I already have an account linked with a gmail address.
 
What are you trying to do exactly?
 
Any way to debug a form that won't let me submit?
@NoTime Submit an issue/bug for the chromium browser
 
yeah same one
I gotta go I forgot I have to do screening
ty for help with stuff earlier
 
5:55 PM
Oh, weird. I opened the page again, and got a different looking bug report form, but with similar content. The strangest thing is that it was partly filled, including my title.
 
6:20 PM
sorry for asking such a noob question, but in AIX what does "There are running jobs." mean when I type exit? Probably that there are running jobs; however I did not knowingly create any jobs at all...?
 
@MaxVernon Normally these are jobs that you are created/started.
 
Could using an ampersand between too commands cause that? like tail -f 20140103.txt & tail -f 20140117.txt - I pressed CTRL-C to exit tail.
 
@MaxVernon Hmm. List running processes
 
ps ?
 
In linux one does ps aux.
@MaxVernon With suitable flags. Check the man page
 
6:26 PM
ok
 
For starters, you want to restrict to processes owned by you
So here, no a. But since this is AIX, the flags are presumably different anyway
 
it seems ps by itself shows a list of my processes
 
@MaxVernon ok
 
@MaxVernon yes, that's exactly what it would do.
 
wow thats a big man page
 
6:28 PM
The message means you have a job running in the background.
Run jobs to see them.
 
cool.
 
@terdon Interesting. Did not know that.
 
thanks. jobs works perfectly. Now I'm confused about why tail -f xxx & tail -f yyy [Enter] followed by [CTRL-C] would result in a background job running.
 
@terdon Shell builtin?
 
and presumably exiting the session would end the job?
 
6:32 PM
@FaheemMitha yup. It goes with bg and fg.
 
@terdon ah
 
@MaxVernon Yes, unless you've done something tricky like use disown or nohup
 
@MaxVernon yes. the shell or whatever. unless you are doing screen or whatnot
or that tmux thing that is in vogue these days. same diff
 
thanks guys. clearly I have a lot of learning to do
is there another way to end the job other than exit - perhaps ej???
 
@MaxVernon Just don't start by reading the ps man page. It's scary.
 
6:35 PM
@terdon no kidding.
 
@MaxVernon just run fg and then kill it with Ctrl+C
 
@terdon that worked. So why didn't it kill when I hit CTRL-C right after running the tail -f xxx & tail -f yyy command?
thanks for your help, again, by the way!
 
@MaxVernon Because that's what the & does. Your command launched the 1st tail in the background and immediately launched the second.
Since it was running in the background, fg brought it to the foreground where it can now catch signals like Ctrl+C
See help fg and help bg
 
OOOHHHHH
so if I want to tail too files in the same tty, can I launch one in the background and still see its output?
multitail is not installed on this box.
by the way I'm looking at:
2
Q: How to tail multiple files using tail -0f in Linux/AIX

Web NashI tried tailing two files using the option: tail -0f file1.log -0f file2.log In Linux I see an error "tail : can process only one file at a time". In AIX I see the error as "Invalid options". This works fine when I use: tail -f file1 -f file 2 in Linux but not in AIX. I want to be able t...

had an issue this weekend where I would have loved to see the output of too logs at the same time
 
@derobert it's made up (I looked it up when this came out, if you google you should be able to find the author's blog)
 
6:42 PM
@MaxVernon, I am not sure if you tried my suggestion but if you have tmux installed that may work for you in a sense
 
@ryekayo ksh: tmux: not found. :-(
 
@MaxVernon No. Well, kinda. You can launch it in the background but to see its output you will need to bring it to the foreground. To launch one tail and then another, you would use ; to separate the commands, not &.
 
Ah yeah, you would have to install its package to get that to work, what about running a screen session?
 
hmmm... I do have the ability to login using multiple Putty sessions, and that is likely what I'll do in future. I'd just like to not have to switch windows to watch output.
 
hmmm.. yeah screens would essentially work. its a bummer you dont have tmux installed, thatd be perfect for what you are trying to do. But fiddling with screen's can also work as well
 
6:47 PM
interesting screen is installed. I'll look at it.
 
Good luck. I am not a 100% with screen commands but screen --help will give you the basics
 
thanks!
 
anytime
 
@Ramesh ! isn't, # is. The character # goes by many names (see Wikipedia); in music it's called “sharp”
in music isn't exactly the same character as # as in number. Unicode has ♯ U+266F MUSIC SHARP SIGN, distinct from # U+0023 NUMBER SIGN. But it's a good enough approximation for ASCII.
 
@ryekayo screen is sweet.
 
6:56 PM
I use it all the time :)
 
CTRL-A 0 and CTRL-A 1 to switch between screens
nice
 
really nice!!
I updated my answer to that question
 
@Gilles And handwriting.
 
@FaheemMitha indeed, the dv/uv ration on U&L is the 26th smallest among the 123 sites on the network (as of about a month ago)
site-name           dv/uv
tex                 0.010
tridion             0.016
craftcms            0.020
gardening           0.020
expressionengine    0.025
genealogy           0.025
scicomp             0.027
matheducators       0.028
blender             0.029
aviation            0.030
bricks              0.030
homebrew            0.030
boardgames          0.035
mathematica         0.035
opendata            0.035
outdoors            0.035
pets                0.036
productivity        0.037
sound               0.038
 
@Gilles where did you see that?
or is that mod-only?
 
7:11 PM
@MaxVernon from the data explorer
plus a bit of local postprocessing
the site as a whole has a dv/uv ratio of ~0.042. This doesn't include deleted posts (which tend to have a lot more downvotes than posts that don't get deleted).
My personal dv/uv ratio is 0.042 as well, including deleted posts. I don't downvote enough.
 
@Gilles the top member on our site has a ratio of 0.001 - don't feel bad.
(dba.se)
 
@Gilles tex is at the bottom. unsurprising. good for them.
Interesting statistics.
@Gilles I don't like downvoting and downvotes. Except for spam, trolls and lunatics. And we don't get many of those, fortunately.
 
@FaheemMitha spam needs flags, not downvotes
 
@Braiam Flagging with a spam flag gives an automatic downvote, so, two birds.
 
Spam → flag as spam. Trolls → flag as offensive. Lunatics → flag as VLQ and possibly downvote. Downvotes are for unhelpful posts, not the crap of the crap.
 
7:21 PM
@FaheemMitha which actually accomplishes nothing
 
@Gilles VLQ?
Very Low Quality?
 
@FaheemMitha There is absolutely nothing to gain by downvoting spam, and it makes it stay longer because the post becomes less visible to users who might add their own flag. Don't downvote spam.
@FaheemMitha yes
 
@Gilles I agree with tex.sx's take on it - where one of the top rep people wrote a post in meta saying they didn't believe in downvotes, and that seemed to be the prevailing opinion about the top rep users. Of course, the TeX world seems to be a very civilized place, for reasons that remain mysterious, so perhaps they can get away with such a policy.
Hmm, maybe they should teach TeX in schools. It might be good for the little ruffians.
 
@FaheemMitha downvotes is the least rude you can do to a post...
 
@Braiam It is?
 
7:25 PM
@FaheemMitha I disagree. If something is wrong, it's wrong. You don't get a star for participating. If you don't downvote wrong answers, this tricks readers into believing that it's a good answer.
 
I wonder if the downvote/upvote ratio rises over time as a site gets more and more like SO. I'm thinking so.
 
tex.se is justified in having few downvotes because it has few wrong answers, but refusing to downvote is anti-social.
 
@Gilles Well, you could just not upvote them. Or explain to them that they are wrong, and give them a chance to fix it. When wrong posts get upvoted, then that is a problem, I agree.
 
@Gilles I wish Facebook had downvotes.
 
@MaxVernon make a data explorer query to find out how they've evolved over time
given your home site you should be a lot better at this than me
 
7:27 PM
@Gilles LOL. I'm not certain there are historical records in data explorer.
I'll look
 
@MaxVernon before going too deep, check this out meta.stackexchange.com/a/2678/213575
 
@Gilles I don't say never. I just try to avoid doing so. On occasion, when people have asked me to, I have downvoted. But such occasions are rare.
 
@MaxVernon there's a table of vote with a date for each (rounded to the nearest day)
 
Maybe if this site gets worse, everyone will start downvoting more in self-defense. I hope that doesn't happen.
 
@FaheemMitha that's what we've been starting to do over at dba.se unfortunately. Trying to keep the question quality reasonably above moron.
 
7:30 PM
@MaxVernon Discovering that you need to downvote in self-defense?
 
Personally I like to upvote anything that is half reasonable. I VtC questions that are off-topic before downvoting them. I downvote the egregiously bad ones.
 
The question quality here is not that far above moron right now.
 
like this question:
0
Q: Update script using CASE WHEN Clause , non-complaint results

user44587I'm tryin to update a column table* DMT_TEE_OTC_Sales.TEE_StandardCostAdj_GrpCur Intermediate calculation 'Rate Obsolete Depreciation' : If BWH_OTC_Sales.PlantMaterialStatusCode = '1' then 0,3 else 0 Intermediate calculation 'Rat...

 
@MaxVernon That sounds like a reasonable policy.
 
@MaxVernon turns out there's one already: data.stackexchange.com/unix/query/201630/…
 
7:32 PM
@FaheemMitha since I've been asking questions? lol
@Gilles sweet
 
@MaxVernon :-)
 
the data is very noisy, it would be nicer to have it grouped by week or by month
 
I mean, in recent months. Things don't seem to be getting better, either.
Fortunately, we don't have the problems SO does.
 
Change SE format?
 
@NoTime Hmm?
 
7:35 PM
@FaheemMitha Maybe process of working site needs to change as size becomes larger/has more users
Seems like similar complaints in other metas
 
@NoTime Maybe. What seems to happen is that the regulars become much more agressive. Judging by SO at least, which is the biggest, of course.
 
either relating to things like review process, basically question quality seems to become lower
 
@Gilles working on that
 
criticism welcome, all I know about SQL is what I learnt by fumbling in the dark on SEDE
 
it looks like it is plateauing but at a lower rate
from graph
 
7:41 PM
the graph is not good because the long horizontal lines are because of my crappy date conversion
I convert the date yyyy-mm-… to the number yyyy*100+mm
which looks nice in the table, but horrible in the graph
@MaxVernon How do I properly convert back to a date?
 
@Gilles thanks for the info. Is there some field that you are unaware of? :)
 
isn't there a date format?
like one that goes seconds from 'X' year
I can't remember the small amount of SQL I learned 3 years ago
 
@Ramesh bovine dentistry
(That's a quote, btw. There are actually a lot more fields that I know nothing about.)
oh, and technically “Is there some field that you are unaware of?” is unanswerable except by “yes” (or “know”). I am aware of bovine dentistry, but I know nothing about it.
 
does this help? w3schools.com/sql/sql_dates.asp or is it too low level
 
@Gilles no worries. If bovine dentistry is the next big thing, there would come a meta.bovinedentistry.stackexchange.com
 
7:45 PM
is anyone here geeky enough to know where the quote is from?
 
no
:)
 
@Gilles you can CAST(xxx AS DATETIME) assuming the xxx is a valid date-formatted string or number
 
@MaxVernon thank you. Man, any kind of datum manipulation in SQL is verbose.
 
its actually pretty tricky. dba.se has reams of questions on how to convert from a VARCHAR field to a DATETIME field
@Gilles yup, and you're welcome
 
7:47 PM
isn't it like you would do in access?
 
@Gilles @MaxVernon example query data.stackexchange.com/askubuntu/query/198615/…
 
I mean functionally?
 
@NoTime not so much
 
Isn't date/time determined as a variable as seconds past 'X' years (I can't remember something like X=Jan1 1960)
 
hello, hello hot question list unix.stackexchange.com/q/148956/41104
 
7:50 PM
@Braiam hmm, that counts a week/month/year from now, which isn't so nice
@NoTime in unix, yes. The epoch (origin date) is 1970-00-00 00:00 UTC
 
@NoTime actually a DATETIME column in SQL Server has 3-millisecond precision, and can represent dates in the range of January 1, 1753, through December 31, 9999
 
You can see it with date -d @0 on Linux
 
@MaxVernon yeah I couldn't remember the exact date if it was like Unix inception or some 18th/19th century time
I thought 1893 and 1960 I am bad with dates
 
@NoTime who knows why they chose 1753 - I mean 1753???
 
is it end of a war?
 
7:53 PM
177
Q: What is the significance of 1/1/1753 in SQL Server?

DanielWhy 1753? What do they have against 1752? My great great great great great great great grandfather would be very offended.

 
@Ramesh ohh julian dates
 
StackOverflow has a somewhat interesting graph for the ratio: data.stackexchange.com/stackoverflow/query/215871/…
 
stack exchange has an answer for everything. Wow, this is amazing.
@NoTime may be. I am yet to read that answer. I am not sure.
 
ahhh Kalen Delaney. She is a goddess in the SQL Server world. And by goddess I mean she knows a lot!
 
Ha ha.
Your great great great great great great great grandfather should upgrade to SQL Server 2008 and use the [DateTime2][1] data type, which supports dates in the range: 0001-01-01 through 9999-12-31.
 
7:57 PM
Yeah they are saying to make a national calendar the date started was 1750 when finished with Parliment the British calendar was off so the date became 1753
So really everything prior to that date is really subject to scrutiny honestly...
I say to my history teacher, who flunks me anyway
 
@NoTime and of course all the other countries in the world who didn't give a fig about the British date change didn't notice the three missing years
 
Must have been hard for people born in 1752
 
lol, my brothers at dba.se just scolded me for doing such poor datetime conversion.
data.stackexchange.com/stackoverflow/query/edit/215871#graph has a much better way of doing the conversion now
 
that is simpler
can you make the column show yyyy-mm without disabling the graph?
 
no doubt!
 
unfortunately no. Data Explorer only accepts numbers or dates as the first column
 
apparently not
 
@Gilles precisely
 
what is the rise rate?
or the function?
 
hmmm, I prefer this way of looking at it. The graph now shows a rise in the ratio of downvoted questions. data.stackexchange.com/stackoverflow/query/215871/…
 
8:06 PM
thats 18% of all questions being down voted (at maximum)
is that good or bad?
 
@NoTime depends on your definition of bad I suppose.
 
Better than stack overflow I would say.
 
What is it in relation to the amount of users?
 
you can run it for SO as well. Its seen a pretty sharp rise in the rate of downvotes recently
 
@Gilles What, bovine dentistry?
 
8:10 PM
@FaheemMitha answering “is there anything you don't know” with “bovine dentistry”
 
@Gilles right
@Gilles MASH?
 
@FaheemMitha No. A lot less well-known.
 
@Gilles ok
 
and only on page 2 of Google for me
 
I'm getting a hit on something called "When Worlds Collide"
 
8:16 PM
@FaheemMitha wow, it got fanfic
 
@Gilles So, some marvel thing, then?
 
no
read the footnote
 
Casey and Andy? Never heard of it.
 
@FaheemMitha Then go read it
 
@Gilles Yessir.
I hope this doesn't sound retarded, but amd64 is the same as x86_64, right?
@FaheemMitha x86_64. — kyr 56 mins ago
 
8:19 PM
@FaheemMitha yes
 
@Gilles Ok, then I don't understand what is going on here.
 
11
A: Why are 64-bit distros often called 'amd64'?

EarlzAMD intially implemented x86-64, before Intel. For this reason, most distros (and other OSs such as OpenBSD) did a first release when it was still AMD only, or renaming to x86-64 wasn't considered worth the effort. I think one of the another contributing factors to this naming confusion was a co...

it could be worse, at some point intel had the working name IA-32e for their 64-bit support
 
@Gilles Yikes.
@MaxVernon at any rate, i have visited dba.sx occasionally, and I am happy to say that it is a friendly site.
 
> This account is temporarily suspended because of low-quality contributions. The suspension period ends on Aug 7 '15 at 19:47.
user have +200 answers
 
How do I improve this ?
Probably wrong chat
 
8:38 PM
@FaheemMitha thanks!
@NoTime looking now
 
I don't know SQL so that was cutting from the post before and trying to figure it out
 
yes my graph was sideways but I didn't know how to move things
the Max(Id) couldn't switch
on unix /linuxsite
 
data.stackexchange.com/unix/query/215895/… now shows change from month-to-month as well as the total number of users.
 
@MaxVernon yeah was just trying to ask how you do 2 separate things :)
 
8:46 PM
@NoTime glad to help
 
no thanks really. I wanted to combine that graph from before (with the downvotes) to see difference between total users and % of downvotes
I really need to learn SQL or some other query languages.
 
@NoTime You'd need to download the csv of both queries, and combine them in Excel probably
 
I don't have windows
I use LibreCalc though
Can you exportt csv to MySQL?
or i could just do data to calc
 
@NoTime I'm working on a potential solution in the data explorer
@NoTime This query data.stackexchange.com/dba/query/215895/… has the data, the graph on the other hand is useless since it needs two separate X scales. Pretty much any charting package should accept the .csv from that query
 
@Gilles This reminded me of:
Willow: Is there anything you don't know everything about?
Giles: Synchronized swimming. Complete mystery to me.
 
8:57 PM
@MaxVernon that looks pretty nice. If you don't work with databases often it seems you forget a lot.
 
@NoTime I think I've forgotten more than I know.
 
I am noticing that I can remember some of the things in your production of script that I forgot, but knew at one time
 
yup - I know how you feel!
 
@MaxVernon same here. I used to know a lot more about stuff you had to know in mid 90's for computers, that would be helpful now..

I have to make dinner though I'll be on and off
 
does anyone know of a command line tool that can enter stuff into a web form? can curl handle that?
 
9:04 PM
@FaheemMitha curl can do it if it's simple
for more complex stuff, there's perl with WWW:Mechanize
 
@Gilles Ok. How about indiapost.gov.in/Tracking.aspx for example?
 
@FaheemMitha write the code and see
 
@Gilles ok
 
9:17 PM
pays 5 bucks to @FaheemMitha
 
9:42 PM
@Braiam Huh?
We now apparently have imaginary money, in addition to Imaginary Internet Points...
 
5 hours ago, by Braiam
50 bucks this ends in the hot list
 
Gentlefolk, is LC_PAPER customarily set or unset?
 
5 hours ago, by Faheem Mitha
@Braiam I'm not betting against that.
@FaheemMitha set
 
On my system (Archlinux): LC_PAPER="en_US.utf8" so it defaults to US Letter. But if I start from terminal with some euro locale (as I said above) it defaults to A4. — don_crissti 35 mins ago
@Braiam Not here, it isn't.
echo $LC_PAPER gives bugger all.
Does anyone know what the 411 on this is?
/me tries to go with the hip American slang.
 
@FaheemMitha mm... is a gnu extension
 
9:50 PM
@Braiam What is?
 
@FaheemMitha man 7 locale says that LC_PAPER is only defined after glibc 2.2 and it's a gnu extension
 
@Braiam Ok, but it's standard on recent linux distributions though, right? Why is it unset here? Or should that be another question?
glibc 2.2 was ages ago, wasn't it?
 
@FaheemMitha isn't libpaper a debian thing?
 
@Gilles Dunno. Is it?
 
I don't know
on debian, it's a native package, and the copyright file doesn't mention any upstream
but I see that it also exists on other distros
 
9:56 PM
@Gilles huh. so, is LC_PAPER normally set or unset? That seems like something you would know.
 
@FaheemMitha I don't think there's a “normally” about this
it's set if the admin or the user configured it
 
libpaper (1.0.4) unstable; urgency=low

* New maintainer.
* libpaper is now a native Debian package.
 
on Ubuntu it seems to be set by the installer
 
Looks like upstream died. Maybe Debian is the new upstream. This would explain why (a) it is a Debian native package and (b) why other distributions have it. They use the Debian one.
@Gilles Oh, ok. I wonder how would would set it in Debian. You saw the question I asked about Chromium?
libpaper (1.0.3-14) unstable; urgency=low

* Added debconf support, thaks to the patch by
John Goerzen <jgoerzen@progenylinux.com> (Closes: Bug#70959).
This was the previous version. -14 indicates it was not a native Debian package then.
 
My minimally-configured Ubuntu 14.04 VM has it set in /etc/default/locale
 
10:00 PM
@Gilles Interesting. I thought you were a Debian user, though.
@Gilles Yes, that seems like a good place to add it. Thanks.
 
@FaheemMitha my home Debian machine only has LC_CTYPE there, but that's a heavily customized setup
 
@Gilles The home machine is customized?
Let me check and see if Debian has anything to say about it.
 
My Ubuntu VM has weird mix of fr_FR.UTF-8 and en_US.UTF-8. Strange, I don't think I ever told it anything about French. Except the France timezone.
 
root@orwell:/home/faheem# locale | grep PAPER
LC_PAPER="en_US.UTF-8"
Huh, I guess I do use the US locale, usually because I can't think of anything better.
dumb question, but does this mean it is set or unset? I thought it was an env variable, but $LC_PAPER doesn't give anything.
 
@FaheemMitha LC_PAPER defaults to $LANG
 
10:04 PM
@Gilles Ok, I guess I could override it then.
 
You must have LANG=en_US.UTF-8. That's a bad idea: it sets LC_COLLATE and that's almost always a bad thing
 
root@orwell:/home/faheem# echo $LANG
en_US.UTF-8
@Gilles What's the alternative? I don't know what LC_COLLATE is.
 
LC_COLLATE doesn't describe correct collation, it's too restrictive (it goes character by character)
remove LANG and instead set LC_CTYPE and LC_PAPER
 
@Gilles Ok, what is your recommendation?
 
plus LC_MESSAGES if you want messages in a language other than English
 
10:05 PM
@Gilles where do I do that?
in locale?
 
@FaheemMitha wherever you're setting LANG, which could be many places
 
@Gilles Ok.
I think I am probably setting it in locales. dpkg-reconfigure locales gives a possible list of locales to generate. What do you recommend here, if anything?
Actually, this might make a reasonable question. I was completely unaware of these issues.
faheem@orwell:~/personal/business/smyt/download$ LC_PAPER="fr_FR.UTF-8" chromium
[30863:30863:0808/034123:ERROR:browser_main_loop.cc(240)] Gtk: Locale not supported by C library.
Using the fallback 'C' locale.
Created new window in existing browser session.
Locale not supported by C library.?
Do I need to generate it first?
 
yes
dpkg-reconfigure locales
 
@Gilles Right. Got it. Doesn't seem to be working. I guess I'll close all chromium windows to make sure.
 
@FaheemMitha debian or ubuntu?
on ubuntu, I think it's sudo locale-gen fr_FR.UTF-8
 
10:17 PM
Huh, to my surprise, that actually does seem to work. I closed everything and run chromium from a shell with the french locale set.
@Gilles I meant the chromium printer dialog thing, not the dpkg-reconfigure locales thing.
The dpkg-reconfigure locales isn't a problem.
@Gilles you should write a post about locales. should I ask a question?
 
@FaheemMitha yes
 
@Gilles ok
If I use fr_FR.utf8 for example, would it know what I meant?
 
yes
I think the encoding (the part after the .) is case-insensitive and ignores non-alphanumeric characters
 
@Gilles Great
@Gilles Done.
@Gilles I use Debian. Is Ubuntu different?
@Gilles Having read your earlier post,
13
Q: Does (should) LC_COLLATE affect character ranges?

GillesCollation order through LC_COLLATE defines not only the sort order of individual characters, but also the meaning of character ranges. Or does it? Consider the following snippet: unset LANGUAGE LC_ALL echo B | LC_COLLATE=en_US grep '[a-z]' Intuitively, B isn't in [a-z], so this shouldn't outpu...

 
10:36 PM
@FaheemMitha yeah, they changed locale-gen for whatever reason
 
I'm still fuzzy what character ranges are.
@Gilles oh
Is this character ranges strictly a shell programming issue?
 
@FaheemMitha no
I even give an example with grep, not in a shell
this could happen in any program that uses locale settings and sorts strings
 
@Gilles ok
@Gilles isn't '[a-z]' bash shell syntax, though?
 
@FaheemMitha it's shell pattern syntax
nothing to do with any hash
it's also regex syntax, with the same meaning
 
@Gilles sorry, that should have been bash
 
00:00 - 17:0017:00 - 23:00

« first day (1398 days earlier)      last day (3539 days later) »