The Whiteboard - 2015-08-27 (page 2 of 2)

« first day (1819 days earlier) ← previous day next day → last day (3185 days later) »

user55340

8:05 PM

... And there, out of delete votes too.

user55340

23

Q: I included emoji to my password and now I can't log in to my Account on Yosemite

I wanted to check if it's possible to use emoji in a password for my Account on Mac OS Yosemite. It worked but I didn't realize that the Mac OS login screen has only native keyboards, so I can't type emoji there. So the situation is that I know my password but I can't type it. The machine has re...

yosemite keyboard password filevault emoji

rofl

user55340

Emojis are evil. 👿

@MichaelT lol

user55340

@enderland and @Ixrec don't you mean 😖 or 😝?

8:16 PM

I prefer (╯°□°）╯︵ ┻━┻

user55340

But that password would work.

who in their right mind uses an emoji in their password

and also, who in their right mind allowed that as a valid character :)

user41796

It's a lot harder to record a password when the person is picking out images from a set of images. And you can make it harder by changing up the order that the images are displayed in

8:42 PM

Specific requests for a tool/library/etc are off topic here (see the close reason). I'm not sure you are going to be able to to improve it in its current form, as the core of your current question is "which tool to use?" — enderland just now

I kind of feel bad for that guy, he actually is making an effort to figure out and improve his question

weird, the only reason I didn't comment was it already had a close reason which seemed self-explanatory

what does the number next to "review" actually mean? it's been stuck at around 40 for me all day but there are far less than 40 review tasks left for me

user55340

How many outstanding review tasks exist.

for me, 9 + 7 + 4

right now the orange number is 39

user55340

39

Q: Notify users of possible reviews on toolbar

I am under the impression that increased participation in the review queue is, on the whole a good thing. It encourages people to feel an ownership in the community and partially responsible for the site as a whole. On Programmers.SE, I've seen what I believe to be an uptick in flagging to clos...

feature-request status-completed review top-bar usability

user55340

My original was yes, you should see 16. That's too computationally expensive.

8:50 PM

"It is also cached, so may or may not be very accurate depending on the amount of ongoing activity in /review at any given time."

I was expecting it to behave a bit more like the usual red and green numbers; an actionable notification which when read/handled simply goes away

user55340

shouldiblamecaching.com

it would be cool if that page actually served up no, but it appeared as yes due to caching

Is that too much to ask?

Oh. Terrible. Will drive me mad having a number there I can do nothing about at a certain point in time. Many thanks for replying! — Phil Apr 30 '14 at 2:08

user55340

Heh - looking at the headers; no-cache

user55340

Wait, that was a 307. The 200 caches for 600 seconds.

user55340

9:02 PM

@Ixrec on breaking the build - I often try for tdd which suggests checking in breaking test cases. Even if you don't have a solution to it - having a breaking build us useful.

user55340

(May I edit your with this point?)

@MichaelT are you referring to my answer on the "should CI auto-revert" question?

user55340

Yep.

user55340

You can always auto revert a broken answer.

What a gross drive-by downvote

9:05 PM

I'm not sure I understand the purpose of deliberately checking in broken test cases to the main CI branch; on a dev/feature branch sure, but on the main branch that just means Jenkins will send you "job 'tests' is still broken" on every commit and you won't notice if anyone adds a non-deliberate bug

@durron597 you suggested a pattern! Bad durron! Bad!

user55340

Bug found in production. Reproduced in unit test case. Check into trunk?

@Ampt Someone upvoted the convert to string switch case answer which is even more gross.

user55340

... Even if you don't have a fix now?

@MichaelT that sounds the same as what I just responded to

9:07 PM

Check the test case in?

@MichaelT what? broad, sweeping generalized statements can't cover every aspect of an intricate process?!?! BLASPHEMY!

If the rule doesn't fit on a chalky candy heart, I don't trust it!

do CI systems/test frameworks normally have features to designate "known broken" tests/jobs and I've just never heard of them?

because if not it sounds like a terrible idea to put known broken stuff where it can potentially hide unknown broken stuff

user55340

@durron597 I'm a proponent of check on tests as soon as they are good tests. Good tests sometimes break things. Don't let breaking things prevent communicating good tests.

user55340

Depends on the suite setup.

user55340

I've seen testng with a "known issues" suite that isn't part of "good tests preventing regression"

9:10 PM

I guess this comes down to whether "fail the build" means "it won't compile/link" or "it fails one or more tests" or "it fails a different subset of tests than it did last commit"

@MichaelT Some people advocate never check a failing test into trunk.

I think that's an opinion based question.

user55340

Once a test passes, it moves to the other suite. Thus, regression breaks the build, not known issues.

You want red to mean something, if you have a red build that's "ok" then it's not really ever a "failed build"

anyway, given the assumption that the CI setup does have a "known broken" suite/status/designation that prevents those tests from hiding new and unexpected breakages, that sounds fine to me so go ahead and edit it in (just make that assumption explicit)

what @MichaelT is saying though I think is teh way you want to do something like that for precisely that reason

user55340

9:13 PM

While I can articulate it well here, I now have doubts as if I'll be able to do it in an answer properly now.

@durron597 if I'm understanding everyone, it looks like everyone's really advocating the same thing and we're just describing it with different ambiguous words

I definitely agree that a failing test is valuable, so if it can be checked into the main branch without hiding "real" problems that seems like a good strategy

user55340

On solo projects, I check in tests first always. Broken builds are communication of information- not something bad.

depends what you mean by "broken build"

I'm used to that meaning "a failed attempt to compile and link the codebase, as well as optionally running tests that should pass"

that was sad, every single reopen vote in the queue was blatantly wrong (or only right in the sense that the wrong close reason was chosen, which imo isn't good enough)

Some bugs are "don't fix yet".

I think the worst thing is to be able to release a build that doesn't pass all it's test cases.

Create a branch "branchForFixingX", then add the test case to that branch.

Anyone who wants to fix that bug can just check out that branch and use your test case to help fix it.

well, I think the worst thing is tools that force you to do what they say, so you're automatically screwed every time they stop working for whatever reason =)

but yes I usually put my failed test cases on branches, and if I don't have time to do the fix I push the branch and link it on the ticket for someone else to use

user55340

9:23 PM

@durron597 two test suites. "Things that must work" and "things we know about"

user55340

Branches off trunk or master. Otherwise you get to merge hell trying to keep the bug bran rev up to date.

agreed, all my branches are based directly on master

user55340

Anyone else branching from master should be aware of all the known failing tests too rather than sequestering then in branches.

@MichaelT Interesting, I hadn't thought of that

Need to learn how to get maven surefire to work with a second test suite.

I suppose this depends on the nature of the project, but for me having the tests attached to the bug tracker tickets is more useful; anyone who works on that ticket will know about it, and anyone who isn't working on that ticket probably doesn't have a use for it (it's very rare for us to accidentally fix bugs we aren't actively working on, and when we do they're not the testable kind of bug)

user55340

9:28 PM

@durron597 junit? Or testng?

at least, I can't think of a use for a suite of known failing tests other than seeing if any of them suddenly start passing

@MichaelT junit

user55340

14

Q: Run Junit Suite using Maven Command

I have multiple Junit test suites (SlowTestSuite, FastTestSuite etc). I would like to run only specific suite using maven command. e.g. mvn clean install test -Dtest=FastTestSuite -DfailIfNoTests=false but its not working. Just not running any test at all. Any suggestions please.

maven maven-2 junit test-suite

user55340

7

Q: Running JUnit test suite using Maven

I have written a JUnit test suite for running multiple test cases. Now I want to run my test suite class (AllTest.java) at once so that all tests are triggered, carried and managed by one class. I know maven-failsafe-plugin is available, but is there any other easier way to invoke a JUnit test ...

user55340

Probably a better more recent way too.

user55340

9:31 PM

TestNG, however, does suites a bit more elegantly (imo)

user55340

(It has multiple groupings and groups of groups)

Wow @gnat Just saw your big argument with shog on the commented flags thing

Thanks for sticking up for me.

link?

9:48 PM

link?

@Shog9 one can argue that this particular case falls under severe content problems unlikely to be salvageable through editing, need to be removed but oh well. Sounds like you want duplicate late answers to be handled solely with custom flags to moderators - good luck with that (I think you'll need a lot of it) — gnat 15 hours ago

@enderland ^^^ there you go

@durron597 granted, Shog logically win. In the sense, I understand they can't do it now and why is that and how that makes good sense now. Thing is, this doesn't worry me because time is on your side. I saw him "winning" like that in link-only answers and in SO close queue, and in sticky hot questions and every time it went the same. After all these "victories" he eventually had to give up, not because of logic but because the issue was real and damage was big. We gonna get it, one way or another

One way or another Lyrics- Blondie

user55340

10:15 PM

@RobertHarvey ghads, that was taken poorly.

@MichaelT fcking poorly?

@gnat Things can always be status-deferred instead of declined :-/.

user55340

@gnat yep.

user55340

That's neat - I can edit a comment on a deleted post?

@durron597 doesn't matter which status it is. Request to show recent deleted posts went to status-completed straight from status-declined. You see, it's not only your pressure, I would say mostly not your pressure at all. Life pushes them. If they keep dupe-answers crap proliferating, they will eventually lose googlers, and let regular users handle these is the only way that scales. It's only a matter of time when they give up and do the thing necessary to survive

10:22 PM

@gnat Well, the dupe answers discussion is more relevant to that then vlq flag with comment.

@durron597 they are tightly coupled. Cohesive if you wish. Same guys (more or less) who do LQ review will take care of dupe answers. And the queue will be much the same, and the way to push stuff into that queue will be much the same, 'cause that's the way that works and scales. That's annotated (commented) flags. Time is on your side, you're basically riding the wave with this request, a very strong wave. Enjoy it while you're there

Rolling Stones Time Is On My Side -HQ

@durron597 yeah that's the one I initially planned to link. :) Just figured that One way or another fits a tad better, because it is really hard to tell how exactly your request will be implemented. Eg I couldn't even imagine that sticky hot questions will be solved with randomizing (not that I complain - one way or another)

user15026

@gnat This is going to be stuck in my head forever now.

@AshleyNunn if it weren't for the audio quality I'd say it's better than 1D's version

user20683

10:48 PM

@MichaelT

user20683

The "Java Life" Rap Music Video

@WorldEngineer I want those minutes back, that was terrible

user20683

@JimmyHoffa it's Java, what do you expect :P

user20683

James Bond 007 Theme Song ( Speed Programming )

user20683

here is a speedrun of composing the Bond Theme song

10:56 PM

cc3.5 and higher nvidia GPUs support threadblock preemption, but it is not exposed to the programmer (yet) - it's used by the runtime under controlled circumstances, two of which I am aware of are dynamic parallelism and debugging. That could change in the future, however. I also agree that this seems off-topic for SO. I suggest this question would be better located on Programmers SE "Q&A for professional programmers interested in conceptual questions about software development" or Theoretical Computer Science SE "Q&A for theoretical computer scientists and researchers in related fields". — Robert Crovella 20 secs ago

user20683

@JimmyHoffa would you consider unit testing to redundant on a resume if the resume also lists test driven development?

@WorldEngineer I would. But then both terms are pretty intangible as far as skills on a resume.

@WorldEngineer I consider unit testing to redundancy often; like Blake said, you never know what is enough unless you know what is more than enough. Once I've reached redundancy I know to take one step back. Enough! Or too much!

11:16 PM

I would appreciate votes to this answer: programmers.stackexchange.com/a/294759/31260

perspective of turning Programmers into debugging device scares me

@WorldEngineer HR and recruiters do not know synonyms. Listing C# does not imply .NET. ECMAscript does not imply JavaScript.

@psr but Java implies Javascript to them, right?

or it's the other way round. Javascript is bigger, it has Java inside

@gnat JavaScript is Java AND Script. Java is for people who don't know the script part.

Er, semantic jinx?

@psr testing/automated testing/unit testing/integration testing/systems testing/test-driven development/behavior-driven development/test framework design, is that thorough enough for HR?

@psr but of course!

11:25 PM

@Ixrec Not if their requirements include BDD.

"Gosh, thank you man! You just solved my problem :O It really was because of the quotes!!" => "next syntax error reported by compiler, I will get here again for help. And I will tell all my classmates how helpful are people here, and recommend them this wonderful site to come to fix their compiler errors."

the sadder part is that he didn't notice the "non-ASCII characters are not allowed outside of literals and identifiers" error message

user20683

@gnat no, he won't

@WorldEngineer well, what can I say. "Gosh, thank you man! You just solved my problem :O"

user20683

@gnat solving problems is what i do

11:41 PM

@WorldEngineer I noticed. I also noticed that you do it well. Despite that funny election promise (or maybe because of it I dunno, decisions could be easier to make when you explicitly don't take sides)

user20683

@gnat It makes it easier to play good cop

user20683

since the auto downvote can't come from me instaclosing things

user20683

they can go and look and see that I didn't

@WorldEngineer yes, that too

if I ever nominate I will seriously consider such a promise myself. I understand it may be hard to keep for a guy like me though

user15026

What promise is this?

11:48 PM

@AshleyNunn never downvote

user15026

@gnat Oh, man, I'd never be able to keep that one.

@AshleyNunn I am not sure if I could, either. But still, impressive isn't it (I mean for a moderator, for a regular user that would be maybe not really wise)

user15026

Yeah, I'd not want to promise it as a user when it's one of the fewer tools in my toolkit, but as a mod, there are other things you can do. So it's pretty impressive, because I know that I tend to downvote certain things I see (like certain question types, etc) on a frequent basis.

oh, that CR question did get "migrated"

@Ixrec IIRC discussion on it wasn't particularly welcoming :) Gotta keep an eye on it

11:59 PM

@gnat I think my response was the most positive by a wide margin

« first day (1819 days earlier) ← previous day next day → last day (3185 days later) »

Transcript for

Aug '1527

The Whiteboard

General Discussion for softwareengineering.stackexchange.com P...

coffee-day horror-stories monads my-code-is-compiling scotch the-other-monitor

join this room
about this room

00:00

06:00

12:00

18:00

all times are UTC

site design / logo © 2024 Stack Exchange Inc; legal

mobile