• (nodebb)

    If there are no errors or warnings, but the build is still failing, perhaps prayer or visiting a psychic can help?

  • A-nony-mouse... (unregistered)

    Exit Status

  • anon (unregistered)

    Failed successfully.

  • RLB (unregistered)

    Someone forgot to buy that build tool flowers.

  • Tim Ward (unregistered)

    Bog standard for TeamCity. Doesn't matter what you're trying to build.

  • Gargravarr (unregistered)

    "Nothing went wrong, I just don't like you."

  • Anon E. Mouse (unregistered)

    If a build fails on the server with nobody there, does anyone hear it scream?

  • 🤷 (unregistered)

    That's the equivalent of someone saying "We need to talk" and then simply walking away and blocking any attempt to communicate with a hateful "You know damn well what's wrong!"

  • Hal (unregistered)

    You code is find but I don't feel much like compiling today

  • Anonymous') OR 1=1; DROP TABLE wtf; -- (unregistered)

    While the lack of useful output is certainly annoying, TRWTF is not reading the fine manual. The default verbosity of the dotnot publish command is minimal. Crank up the verbosity by passing the -v <LEVEL> option with one of the normal, diagnostic, or detailed levels. If you still don't get any useful output there, then TRWTF is dotnet publish.

  • Sole Purpose Of Visit (unregistered) in reply to Anonymous') OR 1=1; DROP TABLE wtf; --

    "Net or DotNet, Young Jedi. There is no DotNot."

  • ooOOooGa (unregistered)

    Even with the verbosity of the process at minimum, if the process fails there should at least be one (probably unhelpful) error message.

    The only error-inducing problem that I could forgive for giving no errors is a 'media full' problem. Running out of disk space could prevent logging the error messages. Or running out of RAM, but that is even more rare.

  • (nodebb) in reply to Anonymous') OR 1=1; DROP TABLE wtf; --

    Nope, TRWTF is not telling the user ANYTHING about what failed or why by default regardless of verbosity level. It's one thing to suppress expected messages, but as a rule failures (in this context) are unexpected and therefore should get SOME sort of useful diagnostic. Print a fucking message. Why would you ever want to suppress fatal error messages on a build tool, LET ALONE by default?

  • (nodebb) in reply to Anonymous') OR 1=1; DROP TABLE wtf; --

    You seem to be confusing printing message content with flat out reporting that there weren't any messages.

  • Paul T (unregistered)

    Now I wait to see if any of my former co-workers read the Daily WTF. At a previous job, I got chided for doing "code review by DailyWTF".

  • CoyneTheDup (unregistered)

    Was the input source file empty?

  • (nodebb)

    Working as an external dev. When the automated build fails, I need to write an Email for someone to send me the logs.

  • löchleindeluxe (unregistered) in reply to Anonymous') OR 1=1; DROP TABLE wtf; --

    I'm only halfway through my first coffee and read "diagnostic" is "disgusted". Still would be a fine log level, though.

  • 🤷 (unregistered) in reply to löchleindeluxe

    DEBUG, INFO, WARN, ERROR, DISGUSTED. Seems like a perfectly reasonable log level to me.

  • MiserableOldGit (unregistered) in reply to Paul T
    Now I wait to see if any of my former co-workers read the Daily WTF. At a previous job, I got chided for doing "code review by DailyWTF".

    I've had a few co-workers who appeared to code by copypasta from TDWTF.

  • Gargravarr (unregistered) in reply to Anon E. Mouse

    Depends if someone installed volkswagen-ci on the build server.

  • CdrJameson (unregistered)

    Post build step failed

  • Gerry (unregistered)

    As implied above by "Anonymous') OR 1=1; DROP TABLE wtf; -- ", the error handling failure is from dotNet Publish. it is outputting:

    [15:50:34][publish] Build FAILED. [15:50:34][publish] 0 Warning(s) [15:50:34][publish] 0 Error(s)

    and exiting with error code 1

    TeamCity is simply repeating what it's told:

    WTFin/WTFout

  • MIke Swain (unregistered)

    Azure ARM Deploy Pipelines have a habit of doing this too. Sometimes if you dig around in the Resource Deployment Logs you can find a bit more info. One thing you can guarantee is that any error message you do find will never have the correct line number against it - what sort of a coder would want to godirectly to line in the json garbage that caused the error! Sometimes it gives you a 'character number' but it doesn't match the offset in the file as they've probably expanded LFs to CR/LFs or something. Horrible

  • WTFGuy (unregistered)

    Ref

    DEBUG, INFO, WARN, ERROR, DISGUSTED. Seems like a perfectly reasonable log level to me.

    If someday we have semi-sentient hardware you can rest assured "disgusted" is the nicest thing they'll be thinking while processing most human-written software. Can you imagine how processors "feel" dealing with all this WTFery all day? Not good.

  • ichbinkeinroboter (unregistered) in reply to Tim Ward

    OMG. MS Build was just the same. Drove me out of a (very well paid) contract gig because I just could not solve something. I cant think of anything else that I have ever found simply unsolvable in a 30 year career.

Leave a comment on “Passive Regressions”

Log In or post as a guest

Replying to comment #:

Log Out

« Return to Article