« first day (2020 days earlier)      last day (2974 days later) » 

6:00 PM
Hmm. There aren't a whole lot of software visualization tools. There are tools (to a varying degree of usefulness) turn source code into UML. Not always useful. There are a few language specific toolkits and IDE plugins. Looks like GrammaTech and SciTools are the main players here.
 
6:28 PM
Hey! I actually am done with getting things 'n shit done.
For a while
@AaronHall in here asking how to do the needful again?
ah, you're blowing out your heap?
@AaronHall to answer your question - I use WinDbg the unmanaged windows debugger of course :)
This is a great opportunity to learn a thing or two about it
wow, there's a hawk gliding around just outside my window
enjoying the wind
@AaronHall WinDbg is absolutely Teh Win. If you're dealing with windows, it is the most spectacularest way to dig into a processes heap and figure out 'the hell is going on
hey there's even a Python extension for it pykd.codeplex.com
 
I don't get to install a lot of stuff, I have the Sysinternals process explorer...
 
it's an extendable debugger that digs around in a processes heap and stack via the OS references/mappings
@AaronHall yeah, no this is way cooler
with WinDbg you can attach to a running process, pause it, and start digging around in it's heap and view the stacks each thread is running etc
it may all look like gobbledygook depending on the application's language and whether or not you have extensions like that python extension installed to resolve address references to meaningful terms
 
Is it too much to ask Tableau to document their freaking SDK?
 
@AaronHall well if you just want to fix the code that's way easier than debugging it - but WinDbg is awesome! :D Everyone should learn it. That said, you can typically reason through performance issues... what's happening for you?
 
things. Things are happening.
 
6:35 PM
of course if you arent doing win32 stuff windbg isn't all that helpful
 
I have a lazy loading table. let me show the pseudocode.
 
beat my team lead in sudoku. She's not happy.
she's usually twice as fast as me. HA!
 
next time challenge her to a sudoko and a bk triple whopper
 
try:
    # Initialize Tableau Extract API
    ExtractAPI.initialize() # *their* logging
    with Extract(filename) as extract:
        table = None
        if not extract.hasTable('Extract'):
            # Table does not exist; create it
            tableDef = makeTableDefinition()
            table = extract.addTable('Extract', tableDef)
        else:
            # Open an existing table to add more rows
            table = extract.openTable('Extract')
        tableDef = table.getTableDefinition()
 
@whatsisname dude you trying to kill me?
 
6:40 PM
in college a bunch of us had a competition: run to burger king (1 mile), eat a triple whopper and solve a sudoku, run back
a challenge for the mind and body
it was awesome
 
@whatsisname what do you mean? In windows, other than assembly, everything is Win32; no?
 
and artery
 
@whatsisname that is simply unacceptable behaviour...
 
Folks, this all works better when we focus on me,
 
my team lead is very competitive. She would probably win out of nothing but sheer determination.
 
6:41 PM
That was the issue. I could have sworn I ran make clean before posting, but I guess I didn't. Thanks for the help! — John 2 mins ago
BOOM
 
@JimmyHoffa: it doesn't really do so good at .NET
 
@AaronHall haha I'm looking now; had to do other stuff
 
he shoots, he scores
 
@AaronHall I'm sure it does, for you :D
 
there is some extension but VS is where its at for .NET
 
6:41 PM
@whatsisname you're out of your mind - use the SOS extension and it's awesome! That's how I learned WinDbgh
 
obviously it is useless for JVM stuff
meh, VS is better for .net than windbg + sos
 
@whatsisname nah, I learned WinDbg debugging memory dumps from prod servers because we weren't allowed to touch them and couldn't reproduce their traffic and thus their issues outside
 
psr
44
Q: Is it possible to work more than full-time for an employer?

OverworkerI have a full-time job for my current employer and have been working here for 2 years. The problem is that I'm working a lot of hours, since I enjoy my profession a lot (I'm in software engineering/devops) and 40 hours of work is not enough for me. My productivity is no issue, people are saying I...

 
@whatsisname I still use it for memory shit like that. Finding out there's a pinned dictionary with 10k items in it (which means 10k items with free gaps between them all pinned killing compaction) because it's in a static class or for other reasons - WinDbg will let you find out that shit way easier than VS
 
psr
This after the one weekend I did a bunch of free work.
 
6:43 PM
@whatsisname is it now? even without SOS WinDbg gives you the unmanaged calls in your stack and can identify the free vs. non-free and numbers of heaps and numbers of threads et al because that's all OS level known
 
every java related problem I've had to deal with has been bugs in the java client rather than the JVM->platform area
 
every time a process in windows is blocking on some kind of critical section - it's using an OS call to do so, WinDbg will tell you about that. Everytime it creates a thread - that's an OS thread, WinDbg knows. How many heaps and how much heap fragmentation is there? WinDbg knows.
@whatsisname yeah, I'm not talking bugs, I'm talking performance issues. That's where WinDbg shines.
 
by java client I mean the jar/program in question
 
@psr He'll just have to quit his job.
 
profilers are good but WinDbg can be used to dig in and find where there's bloat and what exactly is bloating for any and all programs on Windows (shy of an assembly level driver perhaps?)
@AaronHall ok, back to you! Let's see here..
    table = extract.openTable('Extract')
tableDef = table.getTableDefinition()
printTableDefinition(tableDef)
insertData(table)
^^ You can't just bulk a load of crap into the heap like that. You need a stream.
 
psr
6:48 PM
@BarryTheHatchet He'd know that if he'd read the workplace FAQ.
 
@psr Yep
 
@AaronHall if there's no stream API, chunk the file itself independently and construct your own stream from those chunked files
 
hmm so I need to repcap for 30 days to reach 200k
wanna reach 200k on 18th April
 
@JimmyHoffa that's pseudocode. I stream my data.
 
this is going to be tough
 
6:49 PM
@AaronHall apparently not because your data is bloating the heap
a stream simply wouldn't do that.
 
but then I don't have to care about rep ever again. which will be nice.
 
So your hose is clogged. Unclog it.
 
Thanks, Jimmy.
:P
 
I'm half serious
You're not streaming it if your heap is bloating.
 
psr
@JimmyHoffa on a good day.
 
6:50 PM
@BarryTheHatchet what the hell do you get at 200k?
 
your mum's heap is bloating
@Ampt the "infinite rep privilege" immense satisfaction that will do for me for now
 
uh huh.
 
@psr and on half-day's I'm not serious at all
 
300k is far enough off after that that it's a goalpost I can convince myself to forget about entirely
 
The insertData(table) part is more like:
 
6:51 PM
so not caring about rep cap ever again is more a temporary thing
 
probably yeah
 
I was thinking that 250k is a nice round number
 
(rep, not repcap)
 
and sets you up nicely for the 500k
 
damn you, sock
 
6:51 PM
for row in stream:
    insertData(table, row)
 
now you've gone and done it
 
<3
 
@AaronHall that's not what streaming means. You need to read one row in the openTable part, then insert 1 row
 
Is this part 4 of "What Things Mean, with Jimma Hoffa"? part 2 was particularly illuminating. I mean, who knew that expressions were actually statements...
 
oh god not this again
 
6:52 PM
The openTable part creates opens THEIR preexisting table
 
Jimmy and his wacky ways of whispering sweet nothings
 
@AaronHall stream means, pick up 1 piece of data from an input, put it into an output. If you've got it all in memory, then it's all buffered - not being streamed
 
How am I supposed to use their SDK if I don't use their SDK.
This code is all theirs
 
6:53 PM
@AaronHall carefully. Pick up one piece of data, dump one piece of data; repeat.
 
@JimmyHoffa wtf
that's not what streaming means
 
Yes, I'm good with that idea.
 
every single stream I can think of has a buffering layer
 
it better or you're gonna drop packets very, very quickly
 
@BarryTheHatchet true, are you ready for how Jimmy defines words? :D
 
6:54 PM
Well, there's surely some buffering
 
@JimmyHoffa yep, got my Klingon dictionary ready
 
@AaronHall yeah, but in your code - you bufferred the whole thing
 
@JimmyHoffa Darts? Scrabble? The jump to conclusions mat?
 
@BarryTheHatchet Ok, Nikto Klatu Petronus; Ipso Facto Stream.
 
Ouiji Board?!
 
6:54 PM
kek
hmm I need a new name
shall I just go back in reverse order through my old ones?
 
dude you need to just stick with a name
 
@BarryTheHatchet why? You still have 2 weeks
@Ampt he agreed we get to name him each month
 
Can we agree to just make him Lightness and be done with it?
alternatively just pick one profile pic
I'll take either or
please
 
@JimmyHoffa now I actually am opening and closing their file, hoping that clears it up, so that's more like:
for subtable in stream:
    with Extract:
        for row in subtable:
            insertData(table, row)
So I'm kindof buffering my input too
 
@AaronHall sure, so you're streaming table chunks
row chunks is more fine grained, but chunking granularity is always a question of performance demands - Bigger chunks cost more IO and storage, but have less CPU overhead (less switching between opening/closing/diddling with the IO), smaller chunks cost less IO at a time, less overall storage, and more CPU time (take small chunk, write small chunk, clear buffer, take small chunk, write small chunk, clear buffer ...)
 
6:58 PM
And there's no benefit to doing that that I can tell.
 
@AaronHall you should probably just take row - write row - clear buffer - take row - write row - clear buffer
 
essentially open and close a file a million times?
 
@AaronHall yep! That's why I said, more overhead on CPU to do all of it, slows that piece down - but if the alternative is you run out of memory ... what choice do you have?
 
I'll make the subtables smaller, but I'm not opening and closing it a million times.\
 
shouldn't have to open and close the file though, ideally there's a flush-to-file without releasing the OS file handle
@AaronHall untrue, you will do it if the alternatives just blow up and don't work at all. :)
Now you know how to solve your problem. Yay you.
 
7:02 PM
@JimmyHoffa just planning
 
Next?
 
I'm pretty sure their API won't let me .flush()
 
maybe I should go for Supergirl for a while
ooh yeah ok
 
@AaronHall worth a look
@BarryTheHatchet no! We get to choose!
I nominate for next month, ImpendingBroom, CeeSlashCeePlus, and UncorrectException
2
 
__del__is supposed to call .close on the row object, but I'm going to explicitly do that just in case...
wow, it blows up fast, but it's making slow progress...
It's doing 800 rows at a time
and at about 19 seconds per subtable
 
7:22 PM
@AaronHall yeah see, that's the granularity of chunking difference I were talkin'boot. Bigger chunks are faster but cost vastly more storage space, smaller chunks have far more CPU and IO overhead switching and diddling with all the handles etc - slows the overall thing down but works in less memory
It's exact same concept as a file-based merge sort which is what a merge sort is for: sorting large amounts of information when you've got very little memory to hold it in
 
Maybe I could speed it up by making the swap smaller?
It was trying to use all available virtual memory before. now it's pretty steady...
hmm seems stuck
I hit ctrl-c and it jogged the logging loose somehow?
bizarre:
 
7:39 PM
@AaronHall Perhaps there was a prompt waiting on stdin that printed to some invisible stdout pipe
 
2016-03-15 15:21:18,454 1540 risk.lib.reports INFO rowcount = 813, new subtable
2016-03-15 15:21:41,161 1540 risk.lib.reports INFO rowcount = 813, new subtable
2016-03-15 15:22:03,167 1540 risk.lib.reports INFO rowcount = 813, new subtable
2016-03-15 15:22:25,632 1540 risk.lib.reports INFO rowcount = 813, new subtable
2016-03-15 15:37:55,924 1540 risk.lib.reports INFO rowcount = 813, new subtable
2016-03-15 15:38:18,788 1540 risk.lib.reports INFO rowcount = 813, new subtable
2016-03-15 15:38:45,430 1540 risk.lib.reports INFO rowcount = 813, new subtable
 
I think I'm too the point where I am an expert in docker and cannot get answers to my questions elsewhere anymore.... :(
 
@enderland hah, that point always kinda blows. Whenever I have a genuine C# question it's just irritating
 
job security? I know when I'm done with this Tableau thing I'll either be fired or unfirable.
 
1
Q: Dockerpy executing command to cat log file

enderlandThis seems simple, but I am trying to replicate the following Docker command using Docker-py: docker exec dockerName cat /var/log/foo.log Using dockerpy, it seems the following should work: e = cli.exec_create(container='dockerName', cmd='cat /var/log/foo.log') print cli.exec_start(exec_id=e[...

 
7:40 PM
"fk; I'm going to have to do the testing and or disassembly myself to figure out the answer, nobody the hell else will know..."
 
@JimmyHoffa though I did spend an hour and save some poor soul here about a week worth of work earlier today, so it's not so bad
 
@AaronHall fact: both these states are actually synonymous.
 
I know that's probably deep, but I don't get it.
 
@JimmyHoffa heh they're not bad actually
ooh docker-py that looks like something I need
my Python wrapper does half of that stuff with system calls
 
oh, fired as in I'm my own boss?
but not actually fired?
 
7:47 PM
@BarryTheHatchet lol. docker-py is a lot better :P
though it's more useful if you are doing it inside docker, since it communicates directly on the docker socket
so you don't have to actually install docker inside your other container
 
@BarryTheHatchet go dig around the edits on P.SE sometime- you'll find some thoroughly unexpected oddities..
 
FSVO unexpected
 
8:16 PM
Nope; I don't have a clue what you're talking about
But I am about to enjoy getting spacemacs and clojure and scala all tying up with eclim since I made the codes good 'n dones
 
This isn't really a discussion for this site, feels much more suitable for Programmers.SE. In my very humble opinion (as a C programmer), everything your friend told you is bullshit. — Daniel Kamil Kozar 58 secs ago
 
Fiddling with Scala feels like C#++
 
also @BarryTheHatchet a nice part about docker-py is I can run my debugger now
which I highly recommend ;-)
 
@enderland add a name or 2 to the nomination list by way of fork: gist.github.com/JimmyHoffa/78ee16ad42ba4d0c5e93
 
nomination list? for what
 
8:21 PM
Barry's April name
(forking gist just means push the button to create your own fork, and edit the text)
err edit or whatever
 
forking gist sounds like someone with speech disability is talking dirty
 
"ye barby yo forking gist"
 
psr
@BarryTheHatchet Supergirl Races In Orbit?
 
@psr those words actually make logical sense
 
8:24 PM
@psr ooh
 
Edit a fork to submit choices for Lightness' April name here
4
I guess we should just start calling your nameless now
 
Nameless in April
 
heh
how do I push my changes back
or will you just look at the forks I guess
@Jimmy should I add to or replace the choices?
 
8:49 PM
JIMMY DROP WHATEVER YOU'RE DOING AND GIVE ME ALL OF YOUR ATTENTION!
classic Jimmy
 
@JimmyHoffa Bigger chunks was the solution... :P I'm doing 100 times the data in only twice the time...
 
have you considered streaming the data
 
psr
you mean steamlining
 
no
i'm parodying Jimmy and god knows what he meant
 
hooray for logging...
just the right amount of logging is really nice.
 
8:56 PM
yes
 
@BarryTheHatchet add, replace, whatever, the revisions are all related and traversable. Preferably everything would be in latest revision
@BarryTheHatchet Jimmy installing leiningen
 
I'll write a Python program to randomly select the name I like the most.
 
@JimmyHoffa like I did then?
 
@AaronHall and now you know. Control the chunks ftw. Anytime I have to deal with that sort of crap I typically find myself making chunk sizes configurable because different input sets always end up working best with differing chunk sizes
 
Oh wow I feel dumb. For the last few days I was wondering why I couldn't access some Google services, they just timed out. I tried different browsers. I deleted cache and cookies. I tested various devices, rebooted them, to no avail. I was even starting to suspect my ISP. Turns out, I just had to reboot the crappy WiFi thingy -.- There's probably a lesson about debugging in here, but I'm done searching for stuff for today.
Then again, I encountered a problem at work that could only be solved with a clever algorithm, so that probably evens out.
 
9:00 PM
@BarryTheHatchet sure, but your fork will be promptly ignored; you don't get a say. That's the whole point. You get to pick from the list, but not create it.
 
if might have evened out if you'd actually solved it with a clever algorithm
@JimmyHoffa :(
well fork you
 
@BarryTheHatchet that would be awesome!
git clone self - AND NOW THERE ARE TWO JIMMIES! :D
 
oh gawd
so which one has to whiteboard and which one gets to work?
 
@JimmyHoffa this just in, Jimmy confirmed to be ethanol-driven chatbot. Git clone your own copy today for a free daily scotch fact.
 
9:04 PM
that's too many Jimmy Hoffae for my liking
 
Hoffases. Hoffi. Hoffens. these plurals are fun
not to forget Hoffoi, or Hoffæ with the pretentious ligature.
 
ok, I'm going to minimize/close stuff to see if that speeds up my ETL. Ciao!
 
psr
@amon Flagged as Hoffensive
2
 
hoffer goodness sake
 
user114359
9:41 PM
Tank Dating Simulator ... that is an unusual combination of words. — Vegard 6 hours ago
 
user114359
Thanks to Arqade, I now know this is a thing.
 
I agree, but am more concerned that no one prompted the asker for the information needed to make the question better (either for our site or for migration somewhere else). It is key to lead by example. programmers.stackexchange.com/q/312743 programmers.stackexchange.com/q/312761 programmers.stackexchange.com/q/312721 . I'm tired and done. Its someone else's turn now. Go through the recently closed list, check to see where there is confusion and better guidance for users. Thats what the 10k mod tools are for. — user40980 22 hours ago
^^^ MichaelT has left Stack Exchange
13
 
!?
 
What?
 
this is the type of drama I expect from SFF.SE
not over here
 
9:47 PM
Too dramatic?
 
Did he completely delete his SE account? Or just his profile on Programmers?
 
a few cursory searches imply gnat is correct and he left the whole network
 
I searched SO users, didn't see him there
 
I'm surprised he didn't even say anything in chat
 
@Ixrec two trolls and one rep whore got to be happy now
 
user114359
9:49 PM
@PeterTòmasScott He deleted his SO chat long ago
 
Well, that was... unexpected?
 
user114359
I just checked several other sites where I know he had accounts, he does not show up
 
@gnat ???
 
@gnat dafuq
 
@Ixrec check the recent meta, it's all there
 
9:50 PM
that's one hell of a ragequit
 
@gnat I don't see any Qs or As by him on meta
 
5
Q: If the scope is wrong, which of these questions belong here?

user40980Over in Is this site failing? And should it go away? I posted this list and haven't gotten a satisfactory response yet. If the site's scope is at issue, then some percentage of these questions should instead be open. If these are just poorly asked questions that need some moderate editing, then...

That was his
 
oh well
 
user114359
I know he had a lot of patience. This could not have been because of the past week of heated discussion - he had to have been annoyed for some time
 
@Ixrec I mean other's Qs and As. And comments
 
user114359
9:53 PM
@gnat like I said I checked a few SE sites and I don't see him anywhere - no user, no Qs or As that I know were his
 
user114359
room topic changed to The Whiteboard: General Discussion for programmers.stackexchange.com Pants optional when telecommuting. [coffee-day] [horror-stories] [mourning-the-departed] [my-code-is-compiling] [scotch] [the-other-monitor] [xkcd-1305]
 
RIP
 
and didn't even say goodbye
 
@Snowman Yeah, the misconceptions behind Programmers SE are definitely not new
I'm mostly just a stalker in this room, so I kinda have to ask: what triggered this? I saw quite a couple of questions on meta bringing up site scope again
 
9:56 PM
Fatigue.
 
@RobertHarvey disappointment
 
I mean the recent rash of meta discussions not MichealT leaving
 
nothing really prompted it
 
Hi Xzenon. Progammers: programmers.stackexchange.com or CS cs.stackexchange.com might be a better bet to get Algorithms ideas. — Eric S 9 secs ago
 
although certain people had been moderating more heavily than usual over the last few months, and some users were sick of it
so they ranted on meta
this prompted less-ranty meta questions to replace the ranty ones
and here we are
 
9:58 PM
:/
 
dunno, more participation in solving the problem is what people like Michael wanted
now that it happened, ragequit occurs
boggles the mind
 
9 mins ago, by gnat
@Ixrec two trolls and one rep whore got to be happy now
 
seems like some people only accept participation when the new participants agree with them...
just a theory.
 
user114359
@BarryTheHatchet I think it was more "rehashing the same crap over and over with the same resolution" that was a factor
 
@Snowman probly didn't help
more importantly, I'm on my last glass :(
 
user114359
10:00 PM
But only one person knows for sure.
 
> 'ej HumtaH 'ej DechtaH 'Iw
> 'ej Doq SoDtaH ghoSpa' Sqral bIQtIq
> 'e' pa' jaj law' mo' jaj puS
> jaj qeylIS molar mIgh HoHchu'qu'
 
aww yeah filed first bug report on an open source repo... :o
 
docker-py?
 
:|
 
@BarryTheHatchet Is this Lojban?
 
10:03 PM
@AquaTart looks like Klingon
 
Oh yeah, it's Klingon
 
It's Dutch
 
and Sticky blood, and the walls
and the river of sqral in the room to Soak Red
because the day before that a few day
evil God hohchu'qu' molar
 
that's not bad
 
10:06 PM
It's from Bing.
 
Bing, son of Google
> And the blood was ankle deep.
> And the River Skral ran crimson red.
> On the day above all days.
> When Kahless slew evil Molor dead.
 
I wish I was 1337 enough to actually know Klingon
 
user15026
I don't think I could ever wrap my mouth around the syllables enough to not sound like an idiot
 
@RobertHarvey wot
some people really have too much time on their hands
hic
 
@RobertHarvey Smalltalk in a nutshell
 
psr
10:11 PM
@RobertHarvey ifttt.com
turning the internet on itself, two sites at a time
 
Vote for if
Vote against if
Pin and cast your ballots now...
 
user114359
@BarryTheHatchet I already voted today
 
but what if you hadn't
what would you put your mark against
what would you go for
 
user15026
@gnat Huh. That's...something. Unexpected.
 
Far more entertaining than a meaningless flame war pro/con conditionals is watching @Barry spectacularly failing to ignite said flame war.
 
10:21 PM
;p
 
IMO, this conceptual question might be better suited for the Programmers section of StackExchange. — John White 1 min ago
 
bet you wouldn't be ignoring me if I asked you whether you use space indentation
 
psr
@BarryTheHatchet only if I'm doing PHP on emacs
 
@psr makes sense
 
I think this conceptual question might be better off on programmers.stackexchange.com. In here, questions with presentable code are preferred over broad conceptual questions like this one. You might want to check out: How to Ask. — John White 55 secs ago
 
10:29 PM
dammit
Please add the HTML part. — Mr Rubix 1 min ago
srsly r some ppl blind
 
sometimes it would be nice to downvote comments
 
yes
for posterity, here's the question:
0
Q: The variable is not satying for refesh, or not appearing at all

The Zip CreatorI am testing PHP and the variable from the form is not staying, and does not appear even on the same page. Here is the code: <?php echo "<form action='input.php' method='post'>Test: <input type='text' name='last' </form><input type=submit>"; echo "<h1>The last input was: </h1>", $_POST["last"]; ?>...

that's actually a valid and complete testcase so it's not even unclear or anything
how can you not see the HTML there
(FSVO "valid" .. could do with some more HTML tags, I'll grant you :P but it does reproduce)
 
11:03 PM
why am I still trying to answer questions
I'm +260
 
If an HTTP/2 upgrade occurs over TLS during the ALPN negotiation rather than over a cleartext TCP connection, is HTTP/2 identified as the protocol in the subsequent HTTP request/response lines? Or do the server and client both just speak HTTP/1.1 and then the TLS layer does all of the HTTP/2 transport stuff?
 
11:29 PM
> A man who'd just died is delivered to a local mortuary and he's wearing an expensive, expertly tailored black suit.
> The mortician asks the deceased's wife how she would like the body dressed, pointing out that the man does look good in the black suit he is already wearing.
> The widow, however, says that she always thought her husband looked his best in blue, and that she wants him in a blue suit. She gives the mortician a blank check and says, 'I don't care what it costs, but please have my husband in a blue suit for the viewing.'
 
Well I'm confused... Tried making GET requests to google in Chrome and Firefox; both using HTTP/2 over TLS. In Chrome the requests & responses all declare HTTP/1.1. In Firefox the requests declare HTTP/1.1 and the responses declare HTTP/2.0. Ah well, I'm just going to respond as normal and hope that the Apache folk know what they're doing.
 
wouldn't surprise me if HTTP/2 support were still extraordinarily patchy
can't see hidden option to turn it on in chrome://flags
 
user114359
 
According to an extension I've got running Chrome is fine with HTTP/2. In fact, if I make a secure request to programmers.stackexchange.com then it opens an HTTP/2 session (hence why it loads much faster).
Yeah, lot's of testing and reading long spec documents before I put any HTTP/2 support into the wild.
 
@Snowman hah
hmm Apache 2.4.17 supports HTTP/2
maybe I should play with adding HTTP/2 support to my CGI app
just for kicks
I'd look pretty awesome if I got to s/HTTP 1.1/\0 and HTTP 2/ on our datasheet when the rest of the office doesn't even know there's a new HTTP
 
11:44 PM
Performance improvements are pretty epic
 
especially as that means they won't know we don't need it :P
> Remember that all browsers only speak HTTP/2 over HTTPS so you may want to enable that too for your server.
oh. never mind.
almost midnight ... should put dinner on ...
 
Yeah, the browser implementers aren't currently planning on supporting plain-text TCP HTTP/2.
Firefox talked about supporting http:// urls by tunneling them through TLS or something.
But that's all I've heard about it.
 
we support customers adding SSL to instances of our app (because why wouldn't we?) but they'd need to provide the certs themselves and configure httpd themselves
we have no SSL-enabled "demo" or "dev" instances in-house so that's that
 
Ah well, such is life.
 
(sure, I could get a free cert from somewhere and use that .... but it's a PITA to set up as I recall and the browser warnings on free certs are annoying ... and then I can't impress by having it work automatically for everyone either :P)
BREAKING NEWS: champagne's finished :(
 
11:50 PM
Time to start the Whiskey I suppose
 
user114359
@BarryTheHatchet I have some sake in the fridge if you want
 

« first day (2020 days earlier)      last day (2974 days later) »