« first day (3351 days earlier)      last day (40 days later) » 

8:25 AM
@Wildcard yes, I really like all of Russ Cox’s articles on the topic!
On the topic of history, has anyone here read Kernighan’s recent UNIX memoirs? bookfinder.com/search/?isbn=1695978552
 
 
3 hours later…
Tim
11:49 AM
Not yet. Wondering whether that is interesting and how
 
Tim
12:06 PM
I admit I am short sighted, because I am struggling with basics of distributed systems.
cross reading several materials. figuring out what is what.
Do any of you experience using or writing internal web applications? For example, in your organization/company, in your backend teams, do you write nonfancy web applications to provide web interface to monitor the status of your backend services?
Like to know the basic architecture of such nonfancy web applications and what technologies are involved
Approximately how long will the chatroom automatically separate conversations? So that I don't have to use dash lines?
Writing several unrelated topics in one block is not nice to read
Today is a big day for UK people. Bless you all whether you are living there or not.
can anyone tell that I am writing several unrelated things?
This block never seems to end
Have to be back to study (BBS). See y.
 
1:07 PM
@StephenKitt maybe that's what I should suggest when my spouse asks what I want for a Christmas present. Gives me shivers to see "Also by Brian Kernighan: The AWK Programming Language, with Al Aho and Peter Weinberger" to know they ARE the A, W, and K
 
1:30 PM
@Tim reminds me a bit of Nagios! But semi-unrelatedly, I wrote a really simple CGI program to help me manage my work/todo items. read/wrote a CSV file. Very non-fancy.
and last I knew, Nagios was very similar in spirit -- the web page displayed the current status (based on an updated data file) and interfaced to scripts when you clicked the right "action" commands like entering downtime windows
 
 
2 hours later…
3:08 PM
So the UK votes today. Fingers crossed.
 
3:25 PM
wat up Faheem
 
3:50 PM
Hi Jesse.
 
@FaheemMitha How have you been?
 
@Jesse_b Ok. Could be better. Sleep schedule is a bit haywire.
Worried about the state of the world.
 
You find true happiness when you stop caring about politics
 
@Jesse_b Does that still work if the world is aflame?
 
wondering to know if Pretty Putty is working till now ? because I've tried and nothing changes
 
4:04 PM
@αԋɱҽԃαмєяιcαη What is Pretty Putty?
Sounds like a child's toy.
 
@FaheemMitha It especially works then because there would be no politics. However to say that is the current state of things is an extreme exaggeration.
 
@Jesse_b I don't agree with your first statement. And I didn't say it was the current state of thing. It was a hypothetical question.
 
@αԋɱҽԃαмєяιcαη Ah, yes, Putty. You realise there are other options, right?
Oh, never mind. It's a Windows question. So I don't know. It's off-topic here, really.
 
4:25 PM
This general technique, as described in the first couple of paragraphs seems useful, but still quite limited.
Because one does not have the actual match in hand. Just the start position of the match. The article doesn't appear to address it. Though I might be missing something.
Hmm, perhaps I am indeed missing something.
The article has, further down:
> After that, the last \b in the regex is guaranteed to match where the second \b inside the lookahead matched.
But I don't see why that is the case.
Trying to read the associated article on word boundaries now.
I guess I could ask on SO. Would it be on topic there? Presumably it's not on topic here.
 
4:51 PM
Turns out there is actually a regex chat room on SO - chat.stackoverflow.com/rooms/25767/regex-regular-expressions
 
@Tim Monitoring we mostly use existing stuff here (Nagios, Cacti, Smokeping, ...) but yeah, a few. Most of them are pretty straightforward, some even CGI... Depending on how plain, HTML is hardcoded or comes from a template (with Template Toolkit or HTML::Template)
Some are as trivial as cron jobs dumping HTML (or plain text files) somewhere in the web root
Mostly though custom monitoring is by writing Nagios plugins
 
@derobert Hey. Care to address my regex question?
 
Link?
@FaheemMitha you have a link to the question?
@Tim One of the ones I wrote here is more complicated (even including client/server stuff and a model of the system to analyze logs)... but that's rare.
 
5:23 PM
@derobert I haven't posted a question. See the para starting with
58 mins ago, by Faheem Mitha
I was reading https://www.regular-expressions.info/lookaround2.html
I might post a question on SO, but not if the answer is simple. Though I suppose it might help someone later. The more simple-minded and trivial my questions are, the more upvotes I get.
 
Oh... So the lookahead guarantees that at the position we're matching, there is a match for \b\w{6}\b ... a 6-letter word, i.e., a 6 "word" characters surrounded by a word break
\b\w*cat\w*\b then matches a word break (same as in the lookahead), followed by 0+ word characters, "cat", followed by 0+ word characters, followed by a word break.
Since "cat" is all word characters, the only way that both match is if the final \b line up
@FaheemMitha does that make any sense? (Did I even try to answer the right question?)
 
@derobert The question is mainly
57 mins ago, by Faheem Mitha
> After that, the last \b in the regex is guaranteed to match where the second \b inside the lookahead matched.
I don't why that is the case. I've actually been looking for a while for something to save position, but the docs for word boundaries don't say anything about this.
Though perhaps I'm looking in the wrong place.
Or perhaps I'm simply misunderstanding what the writer is saying. Wouldn't be the first time.
 
It's not about saving the position. It's just that's the only way the two patterns can both match from the same start character.
(They match from the same start as a lookahead is zero-width)
 
Ah, I think I see what the writer is saying. If the beginning word boundary is the same in both the lookahead and the following regex, then the next \b must match the end of the word, since there is only one end of the word.
 
yes, and both patterns are only allowing one word
 
5:37 PM
@derobert Do you know of a way to save the position in such a case, like a register, or something?
This is all fine, but it doesn't generalize. What about if you want to match a bit of text which isn't demarcated by a word boundary?
@derobert Is that because of the \bs?
 
@FaheemMitha Perl has a bunch of regexp extensions for doing stuff like that... at some point though, you switch to something simpler to maintain.
@FaheemMitha well, it's because they both only allow word characters. \w does not permit, e.g., space
So "hi bye" (two words) wouldn't match, because the space between the words doesn't match \w. Even though after editing the message its now 6 characters ;-)
 
@derobert But PRCE doesn't?
@derobert Ah yes, I see.
 
@FaheemMitha I'm not sure. It definitely doesn't include some of them (e.g., some of the Perl extensions are run some Perl code here) ..
 
@derobert Suppose one wanted to do something similar to the example at the beginning of that page, but just a text string with no blanks on either side. I.e. not a word.
Could one still use a similar technique? Because here one wouldn't have the helpful word boundary to tell you when to stop.
 
@FaheemMitha Basically, match 6 arbitrary characters containing the substring "cat"?
 
5:43 PM
@derobert Inside a larger string, for example. But yes.
And using the same basic lookahead technique.
Parenthetically, do you find yourself using regexes regularly?
 
Yes, I use them all the time. (I think almost everyone working in Perl does)
 
@derobert Is Perl special in that respect? :-)
 
Probably. There is a reason PCRE is Perl-Compatible...
E.g., when I was working in C++, used them nowhere near as much.
 
@derobert Does C++ even have regexes?
 
Nowadays, it's part of the standard library. Back then, it wasn't, but of course implementations were available.
(Trying to think of an easy way to match "cat" with at least 3 characters total around it... without using any crazy Perl features, or just writing all the permutations)
 
5:49 PM
@derobert I've not been keeping up with the ever changing C++ standards.
At least it's keeping people busy. Probably increasing the employment rate.
@derobert I'm looking for just a pure regex. If I'm going to us it, I want to work it.
@derobert Apparently not obvious, then? I don't see why they have a "save location", and then one could say - go till here.
 
So... maybe you can do it with both a lookahead and a lookbehind...
 
@derobert Oh? In which order?
 
Not sure. Still trying to come up with a way to force the match length to be 6
 
Anyway, I'll think about it, but I should probably move on. I learned a little bit about regexes. I should probably entitle this episode - Look! I Can Write Line Noise Too!
@derobert Don't worry about it. I just thing that the regex specs are a bit crippled.
But perhaps it's by design. If you want to do something outside this, write a program.
@derobert So, how you feeling about your job ending? Looking forward to a change?
 
@FaheemMitha It's sad seeing all the stuff I've worked on over the last decade+ vanishing.. but yeah, exciting to be doing something new.
 
6:04 PM
@derobert Why is it vanishing?
 
Company is closing
 
@derobert Yes, but can't the work be saved?
I mean some of it is code, right?
 
Sure, some of it is code... but most of it has no use outside the company. Some of it I need to ask about open sourcing. But most of it is going to die.
And of course all the servers are being decommissioned, drives wiped, etc
 
@derobert You should try to save what you can save. And you should ask about license changes sooner rather than later. Once the company is gone, presumably that would be harder.
 
It's a tiny company, so it's not like there is one little bit of it I can point to and, yes, I did that. It's everything...
So, sure, I can save a copy of some particularly neat script or whatnot, but it's not really that.
 
6:15 PM
@derobert So very collaborative? How many people did you have writing code?
No useful standalone software?
 
While I've been here, I think 4 people total wrote code. Maybe 5.
 
Not like releasing Blender then.
@derobert How many are still working for the company?
 
Of the people who wrote code? Two.
Another few people doing other things.
 
That's indeed a tiny company.
 
There might be some useful standalone bits, I need to find any and see about open sourcing them. Some of the bigger code things aren't standalone (e.g., customer portal which integrates a couple things) and/or aren't useful outside the company (e.g., survey creation tool... the first thing I wrote after being hired back in '05)
 
6:19 PM
@derobert And yes, you said 4/5 people wrote code. I'm asking how many of those 4/5 are still working there. If it's 2, then roughly 40/50%.
Blender seems quite actively maintained, but I guess it's relatively general interest.
 
A lot of what I've been doing here is sysadmin work. All that stuff is going away. A few of the servers will have a future life, but only after wiping them (which I've already started on...)
 
6:32 PM
@derobert Have you thought about going more in a development direction? Perhaps something with a bit more of a research type component? There are a lot of interesting software projects out there. Not sure how easy it is to get paid to work on them, though.
 
7:20 PM
@derobert Would LPEG work better for this?
 
BTW... seems dban is dead, anyone know of a good alternative? Before I try to hack something together...
 
7:57 PM
:( re: DBAN being dead; quite a useful tool
 
 
2 hours later…
10:04 PM
@Jesse_b this definitely isn't true
 
@MichaelHomer hello
Not sure what you're referring to
 
foo=abcdefghi; echo ${foo/c*g/XXX}
Also globs aren't used "mostly for files" they are used only for files. — Jesse_b 22 mins ago
I don't understand the question either but there are definitely non-file uses of globs even in the shell
 
Would that be considered a glob though?
it's listed under "pattern matching" in the bash manual
I guess extglob also applies to that so yeah probably
 
It's not the glob function, I guess, but that is what people call shell pattern matching in general
 

« first day (3351 days earlier)      last day (40 days later) »