https://pantsbuild.org/ logo
Powered by
#general
  • e

    enough-analyst-54434

    08/16/2015, 5:00 PM
    Task -> global1 -> global2, Task -> global2
  • e

    enough-analyst-54434

    08/16/2015, 5:00 PM
    That is real and can happen
  • h

    happy-kitchen-89482

    08/16/2015, 5:01 PM
    with globals yes
  • h

    happy-kitchen-89482

    08/16/2015, 5:01 PM
    I don't care about those for my purpose... In fact we could hypothetically get rid of them entirely
  • h

    happy-kitchen-89482

    08/16/2015, 5:02 PM
    assuming that instantiating a subsystem instance is cheap, I suppose that might not always be true
  • h

    happy-kitchen-89482

    08/16/2015, 5:02 PM
    and assuming that a subsystem is just config, which also might not always be true
  • e

    enough-analyst-54434

    08/16/2015, 5:02 PM
    Well, and cutting off scoped setting
  • e

    enough-analyst-54434

    08/16/2015, 5:02 PM
    It seems the only 1 set f settings for a global is a desirable trait in some cases
  • e

    enough-analyst-54434

    08/16/2015, 5:03 PM
    if you were to take away the global you'd need to emulate othe write side to only allow one set of settings
  • e

    enough-analyst-54434

    08/16/2015, 5:03 PM
    even if they're always read from a subscope
  • h

    happy-kitchen-89482

    08/16/2015, 5:03 PM
    well, if you instantiated the task-specific one and it had no task-specific values it would inherit all its values from the global one
  • e

    enough-analyst-54434

    08/16/2015, 5:03 PM
    yes - read side good
  • e

    enough-analyst-54434

    08/16/2015, 5:03 PM
    write side could be bad unless restricted
  • h

    happy-kitchen-89482

    08/16/2015, 5:03 PM
    but yeah, I suppose it's useful to be able to say with confidence that the instance you have is identical to the instance everyone else has
  • e

    enough-analyst-54434

    08/16/2015, 5:04 PM
    I think so
  • h

    happy-kitchen-89482

    08/16/2015, 5:07 PM
    I guess global subsystems are implicitly depended on by the GlobalOptionsRegistrar, in that they have the same scoping relationship with the global scope that per-task subsystems do with task scope.
  • e

    enough-analyst-54434

    08/16/2015, 5:09 PM
    Its the opposite I think, global subsytems implicitly depend on the GlobalOptionsRegistrar
  • e

    enough-analyst-54434

    08/16/2015, 5:09 PM
    The can use its values, the GOR cannot use theirs
  • e

    enough-analyst-54434

    08/16/2015, 5:09 PM
    Everything implicitly depends on the GOR
  • h

    happy-kitchen-89482

    08/16/2015, 5:37 PM
    Then that's inconsistent with how subscoping works. E.g., 
    jvm.test.junit
    is "the options scope for the JVM subsystem instance used by test.junit", so by analogy, 
    jvm
    should be "the options scope for the JVM subsystem instance used by '' " - i.e., the global instance.
  • h

    happy-kitchen-89482

    08/16/2015, 5:38 PM
    I think the inconsistency is that the conceptual "global scope" != the GOR
  • h

    happy-kitchen-89482

    08/16/2015, 5:38 PM
    "global scope" is, conceptually, all of pants.
  • e

    enough-analyst-54434

    08/17/2015, 4:19 PM
    @witty-crayon-22786 & @important-piano-50289 - does Twitter actually have go src trees in use yet? I've been applying the plugin to a monorepo of existing go projects and it naturally raises issues and such - I'll be getting Medium folks opinions on these issues, but then I'll be looking for other real-users of the plugin's thoughts too
  • s

    salmon-helmet-84594

    08/17/2015, 4:20 PM
    There doesn't seem to be any documentation for resource_targets in python, can someone point me to documentation, or a quick rundown of how it works?
  • w

    witty-crayon-22786

    08/17/2015, 4:20 PM
    @enough-analyst-54434: no, not yet. we haven't released a version containing the plugin
  • e

    enough-analyst-54434

    08/17/2015, 4:20 PM
    K
  • w

    witty-crayon-22786

    08/17/2015, 4:21 PM
    but cody has another two weeks to iterate to make it more usable
  • e

    enough-analyst-54434

    08/17/2015, 4:21 PM
    I have a shorter time frame so I hit things pretty hard this weekend on the remote side, biggish RB coming, but on the local side I'm guessing there could be controversy
  • e

    enough-analyst-54434

    08/17/2015, 4:23 PM
    Its a bit odd to have to define targets at all for the local go stuff - as a go coder - odder than in other languages. That's as opposed to having to write down go_remote_libs which has a value-add of version pinning
  • e

    enough-analyst-54434

    08/17/2015, 4:23 PM
    I'll see what the real users think in a few hours, but I'm guessing this reaction
1...178179180...2420Latest