« first day (1979 days earlier)      last day (3025 days later) » 

12:00 AM
alright gtg
 
@JimmyHoffa coffee came!
 
Still completely dissatisfied with Git workflow in Visual Studio. Am I to understand that you actually need to buy a 500 page book on command line Git, drink all of the Git philosophy Kool Aid, and then map that knowledge to however things are supposed to work in a Visual Studio Team environment?
 
If you aren't using TFS; Visual Studio is an IDE- not a source control tool. Just use it as such and deal with source control separately.
 
I know nothing about Visual Studio, but yes, to use git properly in any setting you do need to actively study the command line for a while until you let the DAG into your heart and learn to love the history rewrites and all that jazz
it's very powerful once you get it, kinda like C++, but getting it takes a while
 
and after you know the command line you have to devine how to invoke the commands you need from it
 
12:12 AM
@ratchetfreak I didn't know somebody vined them
 
git insert-magic-here insert-something-else-here
 
give me a console to type the commands in rather than dissecting a gui made by people who can't understand UX
 
personally I'm used to version control being totally outside the IDE and I think I like it that way
git rebase -i master
git stash -p
fun times
git diff coworker/master upstream/master~3 -w
 
@Ixrec honestly, fully integrated TFS with visual studio is...really nice.. I've worked in and out of source control in visual studio and it's definitely pretty nice that checking in is no different than saving or compiling and is done in the same tool and UIs
 
@Ixrec git weep
 
12:14 AM
git push origin --force because-github-cant-auto-squash-PRs-sadface
@JimmyHoffa to me saving files to disk and making a commit are very different things, so that level of integration has annoyed me more often than not; I end up never saving at all until I have a complete piece of work or making a zillion terrible commits with no messages that aren't remotely coherent
 
yes that
 
@JimmyHoffa hows that coolaid?
Thirst quenching???
 
also, I would be very surprised if any IDE managed to make interacting with multiple git remotes any easier than the command line
 
user55340
@DeliriousSyntax The Society of Mind is a very accessible book. "Ingenious, amusing, and easy to ready" - there are things in there like "what is a joke?" because that does play an important role in that model of the mind.
 
12:22 AM
@Ixrec yeah i cant imagine not being able to dofferentiate saves from commits
I hit save at the weirdest times
 
user55340
@Ixrec I've got to look at Tower again.
 
Annnd i changed a letter. Better hit ctrl s 3 times
 
user55340
 
user55340
 
ctrl+s is not "I've finished a meaningful piece of work", it's "I'm typing this stuff for a reason and it would majorly suck if I lost all my progress"
 
12:25 AM
or I'm going to test this piece and I can't compile if it's not saved to disk
 
@MichaelT yeah but that costs money
Intellij has great git integration
 
user55340
@Ampt You're saying that to an IntelliJ advocate... I don't mind paying for good tools.
 
How great can one make a vcs tool
I mean its not cheap right?
40 bucks?
 
user55340
If it saves an hour over a year... its worth it.
 
12:33 AM
How does it do so???
Better merges?? Better git history or comparisons??
 
git is one of those tools where once I learn one interface, I have a hard time seeing the point of learning another interface
I already know how to do everything I need to do
 
user55340
The thing is, I don't use it professionally - only as a hobby (Employer^ was git, Employer^^ and current employer are svn)... so I don't do some of the more complex things on a regular basis.
 
and all the resources about how to use git online are about the command line for obvious reasons, so that one kinda wins by default despite all its inconsistincies
 
not having to tab over to another window can be a win over the course of a year
 
user55340
Staging individual lines on the command line... I don't do that often enough. Or bisect... eh...
 
12:35 AM
staging/unstaging specific files I do fairly often, never individual lines
rebase -i is probably the messiest thing I do regularly, and that's only because the github PR workflow requires you to manually squash all the time
 
user55340
I've done it with methods often enough.
 
I've started using stash -p surprisingly often
 
user55340
The "ok, made a bunch of changes... this section is one commit that I want as one thought, and the rest as another commit"
 
I can't actually remember what the hunk-based version of git add is
 
user55340
Anyways... I played with Tower back two years ago. There was a key feature missing for me to buy it then.
 
user55340
12:39 AM
But there's also a part of me that is "I will buy well made tools"
 
user55340
If nothing more than to say "this is worth it" and cast my economic vote for "that part of the market needs to thrive"
 
user55340
Now... all that, @Ixrec feel your ire rise... go take it out on
 
user55340
 
@MichaelT all 8 delete votes spent
 
user55340
@Ixrec You need more.
 
12:43 AM
I do
I think that was still a big chunk of the way to finishing off
 
user55340
@Ixrec You need to rant about break.
 
user55340
100
Q: How can I teach a bright person, with no programming experience, how to program?

Richard FantozziI have been asked to take a person in our IT department who has no programming experience but is a smart and capable person and help him move into programming at lets say an entry level developer supporting existing .Net applications. I definitely believe this person can do it but I am looking fo...

 
user55340
^^ That one is locked locked and would need to poke a mod like @Yannis for them to look at it.
 
user55340
Given
 
user55340
26
Q: Is learning C# as a first language a mistake?

JuniorDeveloper1208I know there are similar questions on here, which I've read, but I recently read this post by Joel Spolsky: How can I teach a bright person, with no programming experience, how to program? And it got me thinking about my way of learning and whether it might actually be harmful in the long run....

 
user55340
12:46 AM
Is tagged with C# only... maybe a retag to just .net there. They're both... meh.
 
I've been tempted to ask on meta if we should seriously consider nuking some of our historical locks, even the ones with a bajillion views
 
user55340
That's the place for it... I've gotten a few re-examined.
 
user55340
9
Q: Reconsider historical lock on hardware recommendations

MichaelTSome of our older questions are about hardware recommendations. Well, Hardware Recommendations.SE is now live - lets clear up some confusion. No, I'm not suggesting migrating these questions there, but properly and finally removing them from here. In particular, I am looking at What are good k...

 
user55340
13
Q: Reconsider historical lock on "How many monitors do you use? Why? How they are used?"

MichaelTThe post How many monitors do you use? Why? How they are used? has the question body of: Why more or less monitors is worse than your configuration? Why is important use it in this way (position)? Productivity is obvious, but what specific advantage? A picture would be nice. ...

 
user55340
This one is still open:
 
user55340
12:48 AM
1
Q: Two locked 'favorites'

MichaelTThe favorite tag is all cleared out except for two questions, which are locked. What are your favorite version control systems? This is more a discussion question than an actual attempt to determine the "best", since that clearly varies by the needs of the organization. I'm more curious abou...

 
Woo! Mobile from my tablet!
 
1:10 AM
I decided that I am a lot more sick than I actually feel.
 
user15026
@KitZ.Fox Oh no :(
 
user15026
I am headachy and tired, I am hoping it doesn't mean I am slowly getting sick
 
Oh no!
you should go rest. With tea.
I am listening to tingle videos and getting sleepy.
 
user15026
@KitZ.Fox Gotta make it home first
 
user15026
1:13 AM
Then yes there will be tea most likely
 
user15026
@KitZ.Fox Tingle videos?
 
Wow. I think I've had that conversation before.
 
user15026
@MichaelT The next one is awesome too
 
@AshleyNunn oh, no tea at work? That's a bummer.
 
user15026
@KitZ.Fox There is tea, but I am leaving pretty much now to catch my bus
 
user15026
1:14 AM
so it will have to wait
 
@AshleyNunn also called asmr. Some people have it, some don't.
 
user55340
@AshleyNunn Yep. It is indeed.
 
@AshleyNunn oh! Kbai!
 
user15026
@KitZ.Fox ASMR is awesome :D
 
Right. I forgot that's a thing we have in common.
Omg and I forgot I can edit in mobile now! Yip yip yoooo!
 
1:21 AM
Boss says we need to keep our own master branches, fork a branch for each new feature, merge the branch into our master branch, and then do a pull request on the production master branch. Someone responsible for integration testing accepts the pull request.
 
That sounds complicated.
 
blargh, it's much easier to have one local repo and two remotes
just feature branch off master, push the feature branch to the remote that is your fork of upstream and then make the PR to upstream
 
have you been breaking the build much?
 
Not so much breaking the build as being unable to push changes due to merge conflicts.
That, and build problems. Getting a project to clone and build from the master branch is damn near impossible.
 
That would make it difficult. So this change is supposed to help reduce conflicts?
 
1:26 AM
Well, if everyone is working from their own clone branch instead of directly off the master branch, and then folding their features in with pull requests... Maybe?
 
Assuming you can get your clones to build from the master, of course.
 
Yes, assuming that. Of course, once you get it to build once, it will continue to build, provided you can keep it stable.
 
1:56 AM
@Ixrec are you here?
 
yes
 
why does this not work:
 
I am cooling off after election season by reading up on why exactly STV is the best thing ever
 
var payload = {}; payload.message = "wat"; $("div.hello").text(payload.message);
<div class="hello">ABC</div>
 
1:58 AM
unfortunately I do not use jQuery, and I'm guessing the bug is in the selector somehow
 
@durron597 why is your identifier "div.hello" ? that seems wonky to me
 
@enderland .hello doesn't work either.
 
yeah, I thought it was normally $("className") or something
not div.className
 
also hello neither.
jsfiddle is blocked on the work network, can someone test it for me, i'm using an internal fiddle tool
wrapping the div in <html><body></> didn't work
 
154
A: How do I select item with class within a DIV in JQuery

David ThomasTry: $('#mydiv').find('.myclass'); JS Fiddle demo. Or: $('.myclass','#mydiv'); JS Fiddle demo. Or: $('#mydiv .myclass'); JS Fiddle demo. References: find(). Selector context.

does it have more than just htat? like an ID?
 
2:02 AM
aha, i figured it out
it's not the selector, it's the amazon environment
 
hah
 
i had to inject jquery using some wonky amazon proprietary syntax
/me is not a javascript developer
 
2:14 AM
jimmy will convince you someday @durron597 :)
 
 
2 hours later…
user55340
4:12 AM
0
Q: Three way light switches in a tower

MichaelTTalking with an electrician some time back when the house was being remodeled we were discussing the virtues of the remote control of lights. He had boasted that while the technology is neat, there are some things that just don't work well unless you wire them traditionally. The example that he...

 
6:08 AM
Better to ask this on Programmers.SE. — Mars 24 secs ago
@Pascal, I think I see your point--maybe Johan's question is not specific enough even for Programmers.SE. But perhaps with more detail it would count as an instance of "conceptual questions about software development"? e.g. with more detail about matrix sizes and kinds of operations, amount of RAM he has, etc. (I can tell you really quickly that certain kinds of processes with a certain amount of data are a bad idea in R on a low-RAM machine. Learned the hard way.) — Mars 20 secs ago
 
 
1 hour later…
7:47 AM
@Vizllx if you already have a working solution and want to optimize it, Code Review is the site for you. If you are only interested in the algorithm and don't want to actually use it implemented Programmers is more suited. Any way you turn it, your question is off topic for this site. — Gerald Schneider 34 secs ago
 
 
2 hours later…
10:07 AM
@JimmyHoffa that being said I no longer remember anything about it so take that how you will
oh yeah I do now
This is a pretty good example of why we shouldn't confuse the STL and the C++ standard library, by the way. — Lightness Races in Orbit 7 secs ago
 
 
2 hours later…
12:10 PM
6
A: How is C++ function's default parameter passed?

Lightness Races in Orbit My question is, in the second example, should the function parameter be passed by value, so that change of arg will have no effect on defaultvalue? No. It is impossible to pass an array by value (thanks a lot, C!) so, as a "compromise" (read: design failure), int[] in a function parameter l...

well that was easy
finally is executed even if the catch block rethrows, right? that's the whole point, right?
yes. good.
even if it returns, which is nice
 
1:10 PM
0MQ doesn't have a broker? WTF is it, then?
Oh. It looks more like a pipelining framework than a message queue.
This is weird
 
Interesting. I wonder if I'm famous.
 
I looked at the Workplace one, the formatting is really bad
 
@Ampt I said it works better than a lot of people realize, I didn't say it's the end-all-be-all. I prefer Git, and agree that Git should be done by command line. Just saying, integrated tools are nicer than you'd think if you never worked in a wholey integrated stack.
 
@JimmyHoffa I'm going to have to (re)learn git again and this time I'm going to really learn the command line
 
1:14 PM
So is this a Programmers book or a Stack Overflow book?
 
@enderland the dev's are out for blood this morning. Watch your 6.
 
@Ampt your "Day X" stuff has inspired me to write an internal blog about my experiences as a new hire, I think
 
Haha. This guy just took a bunch of SE site questions and turned them into ebooks.
I wonder how much money he's making from this.
 
Hopefully not very much, the information is free online and... better
 
Anyway, 0MQ...I don't understand the advantage of 0MQ versus something like ProtoBuf over a raw socket.
The whole point of a message queue is to have a broker that your components connect to. 0MQ doesn't have a broker, so your components need to know about each other instead of every component knowing about one broker.
I mean, 0MQ does look cool. But I guess I'm not understanding what problem it's solving.
They do get bonus points for including my favorite Sun Tzu quote in their docs.
> If you know the enemy and know yourself, you need not fear the result of a hundred battles. — Sun Tzu
 
2:01 PM
Happy Coffee Day.
2
 
Happy Coffee Day
@ThomasOwens I really don't understand the hype around 0MQ either, I think it's something that makes sense to some particular dev culture around a particular technology; I suspect some subset of C++ folk who work in particular types of systems. The fact that it's neither server nor client, but more of a template for creating your own just makes it seem totally useless when there are high quality options that don't require you to build them from scratch yourself.
I think the place 0MQ makes the most sense is for in-process rather than out-of-processs communication. As a means to transmit information between different modules and or threads within the same application it may be useful (for some languages where doing that isn't absolutely straight forward already). For out-of-proc communication though just use something already built to do it; a server and client platform that exists
 
I think for 0MQ to be useful, it needs to be components that you all control.
Like if I made a box that had some some components made by you and some components made by me, 0MQ isn't a good idea. I'd want a third party broker. But if I had a box that had all kinds of components made by me in different pub/sub or pipelined configurations, then 0MQ may make sense.
But I'm still not sure why I would need it over sockets, though.
 
@ThomasOwens this is why I say for in-process stuff. Where you have one whole homogeneous application, but as soon as you want to communicate with another application, it might be different. Why would you use 0MQ?
I don't know. I just really don't get it, but again I suspect it makes sense for some particular technical sub-space I just have absolutely no experience with
 
@JimmyHoffa I can see it being useful across applications. For example, we have a box that sits on an airplane. It's a physical box with 4 single-board computers. On those single board computers, we run a bunch of different, loosely coupled applications. One sits and reads data from hardware and writes it to disk, another receives tasking instructions from a ground control station, and another turns the raw data coming off of the hardware into a usable file format for later processing. Others, too
0MQ looks to be a good way for us to communicate across the pieces of this box, because we own everything.
We can construct our own publishers and subscribers or pipelines for handling the data as appropriate.
But if we were going to host someone else's thing on this box, 0MQ stops making sense.
This physical box, as a whole, communicates with the outside world on an ActiveMQ message queue, with other physical boxes made by other people. But everything else internal to this box is 0MQ.
Or, would be 0MQ.
 
2:20 PM
@ThomasOwens but again, why? If you went 0MQ you'd have a shit time every interoperating something different in the future. You'd be locked into whatever you built with it where one of the biggest benefits of loose coupling is you can improve/replace disparate pieces over time as newer functionality is required, as newer technology is available, etc
 
@JimmyHoffa Why would we be locked in?
 
If you can use ActiveMQ, what reason would you use 0MQ?
@ThomasOwens because you have to build 0MQ stuff from scratch yourself - it's not a standalone component
 
@JimmyHoffa If that was the architecture, what's the problem with that? We would be tied to 0MQ, but there is 0MQ support in most languages - C, C++, Java, Python, Perl...
 
ActiveMQ is standalone, with independent client libraries you can just use like a SQL connection. 0MQ you have to build out the client and server yourself
@ThomasOwens but why would you use something that locks you in for the long haul when there are other high quality options that don't?
That's the thing I never understand about 0MQ, there are other choices that do a bunch of stuff it doesn't. It's like using a 486 instead of a Pentium because "It too can do computery stuff"
sure...a 486 can... but why would you? I just don't grok that. There's better choices, 486's are good, great even for many metrics of the term, but so many better choices..
 
@JimmyHoffa There's nothing that ActiveMQ has that we need. Although with the latency and performance requirements, 0MQ may actually be better.
 
2:24 PM
486? I played Doom on my 486 twenty years ago.
 
But my preference would be for a broker-based thing.
 
@ThomasOwens Nobody needs to upgrade later right now. But later on, they'll fkn wish they made it a possibility.
 
An internal ActiveMQ instance, or something comparable.
 
@ThomasOwens yeah, there are comparable choices; RabbitMQ is much lower overhead than ActiveMQ
 
And I think RabbitMQ has many more language bindings than ActiveMQ.
 
2:27 PM
Eh; I'm not certain I buy that. ActiveMQ really has a robust set of transports, I can't see RabbitMQ for ActiveMq being different regarding client availability
I've used ActiveMQ with STOMP over WebSocket from JavaScript. It supports AQMP which is the "standardized" protocol (RabitMQ's original big selling point: First implementor of a standard MQ protocol making it available to any client that implements the standard)
 
Either way, it boils down to why aren't you using sockets if you are using 0MQ?
 
I have 0 million questions about that
 
Anything that you can do with 0MQ, you can do just as easily with sockets.
 
@ThomasOwens because it's so barebones, it doesn't really do anything for you; which is my whole point
 
I spent some time looking into 0MQ as well as it was the preferred "messaging technology" at my work place. Unfortunately, what I wanted was a message Broker, which 0MQ is not. And while I could build my own broker (and my colleagues asked me why I didn't) I think that's a waste of effort. It's even almost arrogant to believe that I could build a better broker than all the ones out there with hundreds of dev, test, and prod hours behind them.
I eventually settled on RabbitMQ.
 
2:33 PM
you either want to do your own thing from scratch, or you want to do a standards based broker - 0MQ is so close to the former of those choices as to lend no actual value from what I've seen.
 
So it is safe to say that 0MQ doesn't add much over raw sockets?
 
has anyone here actually used 0MQ for anything? (I don't know what the point of it is either)
 
Because I can't find anything.
 
I thought the point of 0MQ was that it made implementing basic messaging patterns easier.
 
@ThomasOwens it adds a ton! Look at all the documentation! You don't have that much documentation for raw sockets! It adds lots of blog articles, fan boys, probably heaps of code and unit tests and uh ... stuff. Oh you mean useful stuff? No, none at all to my visibility.
 
2:36 PM
documentation is nice I guess
 
@MetaFight but... those are easy.. Why do we need a library with a bunch of design/architecture conventions being encouraged when you can just create all the messaging patterns you want with standard data structures and or protocols (protobuf, msgpack, etc)
 
I find sockets to be well documented.
 
If I remember correctly, 0MQ makes pub/sub trivial because it takes care of some of the nitty gritty for you. eg, Subscribers will automatically scan the network for a matching publisher.
 
@ThomasOwens I was kiding
@MetaFight to my recollection, 0MQ required you to implement portions of the network facility code because it acted more as an engine template where you had to create all the IO connectors to it
 
@MetaFight I didn't remember seeing that in the docs.
I suppose some things could be useful. Like "If you're using TCP and a subscriber is slow, messages will queue up on the publisher. We'll look at how to protect publishers against this using the "high-water mark" later."
 
2:38 PM
I may be confusing it with another technology... but I'm pretty sure I used 0MQ to create a pub/sub prototype at work. I ended up ditching it because my publisher was transient. It came to life, published, and died too quickly for the subscribers to discover it. Adding a Sleep(2000) helped.
 
this kinda reminds me of how git isn't really a usable VCS out of the box, but it's great after you read all the e-books and tutorials and figure out what set of aliases and branching patterns you want to use
 
meh git
only gits use git
it's all about SVN
 
I use SVN too
 
@Ixrec that's true about 0MQ, but the difference is, git is a useful standalone application. 0MQ is like a paint-by-numbers template, you have to build it yourself more than anything
 
then you are, like me, a god amongst men (and women (and whatever else))
 
2:40 PM
I use zip files on dropbox... and a naming convention that is failproof. Codez-Yesterday1.22.zip
 
@JimmyHoffa good point
 
real programmers save their code as phone-camera photos of an IDE, embedded within Word documents stored on floppy disks
 
SVN? You're trolling me, right?
 
and they use butterflies and air currents and all that
 
wrong
SVN is the win
 
2:40 PM
@KitZ.Fox me or Lightness?
 
it's easy, and it works. what more could you possibly want
though admittedly I can see the draw of DVCS in some realms
 
if you want simple idiot-proof VCS that anyone (even non-technical) can learn in 10 minutes, SVN is the win
 
Coffee. I could want coffee.
 
I use SVN for non-coding projects
 
I got tea and Red Bull :D
 
2:41 PM
never seen anyone who couldn't understand "update before changes, commit after changes"
 
Might as well use renamed folders for that.
 
@Ixrec heh I have :D
we tried getting product managers to store tech docs in SVN. it did not work.
 
ok, no one who contacted me asking to help with my fan translations
 
TortoiseSVN shell extension, even!
 
@Ixrec SVN feels like visual source safe did to me
 
2:42 PM
never used that one
 
it makes all the - what should be super basic standard VCS actions - fall into the "advanced" category
 
So favorite dev told me that I'm taking it too personally, that he can't work with an emotional wreck, and that I need to get my shit together and get in the game because the team needs me.
 
@JimmyHoffa that's git
 
@KitZ.Fox No trolling: from what you've been saying lately, he may be right!
With sympathies
 
2:43 PM
I used to use TortoiseHg.
@LightnessRacesinOrbit He's my favorite dev for a reason. He's always going to tell me straight.
 
maybe it's time to take a hip flask to work and bring the cool
@KitZ.Fox Keep those people close :)
 
@JimmyHoffa gotta agree with Lightness on this one, SVN is the one where the basics are dead simple, git's the one where you have to learn about branches and remotes and fast-forward merges and the index and staging area before you can even do one "real" commit properly
 
I guess 0MQ could be advantageous. I like working in messages instead of streams, and we have messages that are too big for a single UDP packet.
 
@Ixrec yeah like wtf's "rebasing" jesuis
 
@LightnessRacesinOrbit no, git has all the exact opposite features hard/easy from SVN. SVN checking in is pie, but any VCS actions (checking in doesn't count, that's basically saving) are confusing and weird to operate. Git, all the vcs stuff (branch, merge, distribute, etc) are unbelievably simple, the hard part is just adding files/removing files/committing changes heh
 
2:44 PM
That right there could be a huge win.
 
@JimmyHoffa sorry which SVN actions are confusing and weird?
 
(yes, you do need to know about fast-forward merges right away, or you end up with a fake master branch pretending to be the real master for SIX MONTHS before anyone notices the problem)
--ff-only
 
@Ixrec branches and shit are the basics of VCS. SVN makes all those VCS things a pain.
 
always --ff-only
@JimmyHoffa oh.
 
committing isn't even a VCS thing; that's like uploading a file to yahoo
 
2:45 PM
I meant basics as in "update" and "commit"
that's more basic than branching
 
what's confusing about svn copy svn://host/project/trunk svn://host/project/branches/branch-name
then svn co svn://host/project/branches/branch-name my-wd
then svn co svn://host/project/trunk my-trunk; cd my-trunk; svn merge svn://host/project/branches/branch-name; svn commit
 
the magic of DVCSs is that they did branching well, so that becomes basic...after you get through the learning curve
 
@Ixrec I don't even associate those with VCS to be sure, VCS is all about a bunch of vastly broader things than just "upload my changes" or "download latest changes"
 
I see that as incredibly straightforward, but maybe that's just familiarity talking
besides, I do it in the GUI shell so it's even better...
 
@LightnessRacesinOrbit Can you use your GUI to track the killer's IP?
 
2:47 PM
conflict resolution can be quite amusing .. I wouldn't do that at terminal
@ThomasOwens I haven't used VB in a while, but I'm a PyGTK fanboi at the moment so who knows!
 
rebase conflicts are fun
 
@LightnessRacesinOrbit shrug, the other thing is SVN constantly spits errors out at me when I commit my changes complaining about not being able to do something or another - even though when I double check the commit was completely succesful, and then I have to delete my whole local repo and reget it (changes come down fine) to fix the thing
 
that's never happened to me
 
GtkTrackKillersIpWidget
@JimmyHoffa that's like never happened to me ever - you did summat wrong
 
I've only used SVN as may main VCS for 6 months now; but I find it irritating as hell compared to git or TFS
 
2:48 PM
inb4 it shouldn't be easy to do summat wrong
inb4b4 it's not!
are you sure you're not trying to use SVN as if it were Git and tripping up that way?
cos that ain't SVN's fault
 
@JimmyHoffa the coffee is good - I'm not sure that it's worlds better than the stuff I get from berres brothers
 
@LightnessRacesinOrbit I really don't do anything special with SVN. I check out my files, I do my codey work, I check them in. Sometimes I look at the commit log. That's literally it.
 
but then again, the stuff at berres brothers is probably super fresh too, so I'm just spoiled that way
 
then your installation is completely borked because that experience happens to literally no-one else ever...
 
@Ampt berres brothers? and which did you try?
 
2:49 PM
I have seen SVN get a bit weird about conflict resolution, but for the kinds of projects that tend to actually have merge conflicts I'd either learn how to do svn commandline or switch to git
 
I'm not trying to be difficult, but honestly...
@Ixrec yeah that's prob the worst part of it IMO
although sometimes you can have some fun with mixed-revision WDs on commit .. again, though, pretty easy to resolve .... just svn update then try the commit again
 
or I might try one of the other DVCSs and see how much of the git weirdness is really git's fault
 
one thing I don't understand (never bothered to research it) is recursive vs non-recursive commits
 
@LightnessRacesinOrbit I don't doubt you really. The whole thing just has me creeped to the point I don't want to fuck with it for fear of trashing the entire source control.
 
TortoiseSVN complains at me about that occasionally but I always ignore it and it's fine
 
2:50 PM
wtf is a recursive commit?
 
@JimmyHoffa eh you're not a real programmer if you don't trash the entire source control once in a while :D
@Ixrec well, quite
 
@Ixrec precisely this is something I feel I need to do when I get the time
 
I had the dark last night, and now I'm having the donut
 
@Ampt very poetic. consider writing a song?
 
2:51 PM
I've heard Mercurial is the easy to use DVCS
 
"From the Dark to the Do[ugh]nut, by Ampt Amptsson"
 
Don't get me wrong, it's good, and I'll probably order more, but it's not earth shattering
 
git is the only tool mentioned so far that has ever trashed my entire source control, but w/e
 
oh dear I just realised the missing letters in American donut spell out the word Americans use when thinking about so-called "British English" :(
 
:/ if I thought we had a snowball's chance in hell of recovering things after I trashed them, I would be a bit more fearless with SVN - but it's totally arbitrary erroriness (and frankly poorly organized nature before I got to this job) makes me not particularly interested in doing much more with it
 
2:53 PM
@Ixrec wait, are you implying that SVN is better than git?
 
SVN is orders of magnitude better than git. It just can't easily be made distributed which AFAICT is its only flaw ever
 
SVN is "better" than git in the same way that Javascript is better than C++: the latter is easier to drive yourself insane with
 
orders... of magnitude?
 
@JimmyHoffa um nightly backups of your server?
 
@Ampt ah yeah that probably is just as fresh then. You can think of Hi-Line as automatically-arriving Gerres Brothers then :)
@LightnessRacesinOrbit not my place or within my authority
 
2:54 PM
Orders of magnitude are written in powers of 10. For example, the order of magnitude of 1500 is 3, since 1500 may be written as 1.5 × 103. Differences in order of magnitude can be measured on the logarithmic scale in "decades" (i.e., factors of ten). Examples of numbers of different magnitudes can be found at Orders of magnitude (numbers). == UsesEdit == Orders of magnitude are used to make approximate comparisons. If numbers differ by one order of magnitude, x is about ten times different in quantity than y. If values differ by two orders of magnitude, they differ by a factor of about 100. Two...
 
I'm not IT
 
@LightnessRacesinOrbit yeah, I thought that maybe this coffee was spiked with something
 
@JimmyHoffa you guys don't have nightly backups of your server and nobody has questioned this with the people in authority? sorry but that's completely your place!
 
no chance in hell svn is better than git
 
@Ampt it's spiked with the blood of Git's many victims
 
2:54 PM
@Ampt more serious answer: it depends on the project
 
@LightnessRacesinOrbit whoa whoa, I didn't say that. I've asked and been told we do, but I don't have a spectacular faith in our ability to restore such and don't want to ever have to find out
 
@JimmyHoffa you are forced to use a server that has zero backups - if you're not willing to talk to IT about that then god help your business :(
@JimmyHoffa aha!
 
for a team of developers who care enough about VCS tools that they're willing to learn git properly, yes git is unquestionably better
 
@Ixrec Oh yeah, almost forgot - if you're arbitrarily committing binary blobs then SVN is the tool for you!
 
welllll that's different :)
 
2:55 PM
if you're doing software development, stick to git!
 
if it's non-devs or the devs don't want to "waste time" on that, then probably SVN
@Ampt true, I haven't had much of that particular issue yet
 
I must say I sometimes quite enjoy basically being the development team's IT guy - that is, nobody else wants to maintain our build/dev/VCS server, and proper IT doesn't know how to use Linux, so I end up doing it. which means I get to set up the backups and I know they can be restored because I do dummy replication to VMs once in a while to PoC out new servers
 
b-...but the bus factor!
 
of course the downside is it saps my time away sometimes
 
@Ampt which do you like more so far? The Donut House or the Stock Exchange?
 
2:56 PM
@JimmyHoffa hard to say - only had one of each and they weren't back to back
 
anyone here done an octopus merge?
 
@Ixrec is that where you crush an entire order of calamari into a solid lump of fried awesome?
 
that's what I imagine it feels like
 
you mean merging multiple branches right after eachother?
 
no, that's easy
 
2:59 PM
yeah - and it's waaaay better in git than SVN because git's merge algorithms weren't written by monkeys
 
I mean creating a merge commit with more than two parent commits
 
the difference being....?
 

« first day (1979 days earlier)      last day (3025 days later) »