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

    mysterious-action-44812

    05/13/2016, 7:36 PM
    gistfile1.txt
    gistfile1.txt
  • r

    rough-minister-58256

    05/13/2016, 7:37 PM
    the way it works is to fetch the base url and scan that for 
    <a>
    tags to infer available filenames for resolution
  • r

    rough-minister-58256

    05/13/2016, 7:37 PM
    (or 
    ls
    for file paths)
  • r

    rough-minister-58256

    05/13/2016, 7:38 PM
    so you might have more luck pulling that artifact down locally - along with all your others - and then pointing pants to that dir to resolve against
  • r

    rough-minister-58256

    05/13/2016, 7:38 PM
    if you’re in a position to commit those into the repo too (e.g. modern github repo style), that’d be a clean way to manage deps for a smaller set
  • r

    rough-minister-58256

    05/13/2016, 7:39 PM
    then pants is always building against a set of committed dists - ideally you’d precompile those to bdists too
  • r

    rough-minister-58256

    05/13/2016, 7:39 PM
    iirc, this is how square’s repo works
  • m

    mysterious-action-44812

    05/13/2016, 7:39 PM
    ic so u would actually pull it down and use git-lfs to store it?
  • m

    mysterious-action-44812

    05/13/2016, 7:39 PM
    @enough-analyst-54434: is the url the way I have it supposed to work? cause if we can get this to work this would be simplest for me
  • m

    mysterious-action-44812

    05/13/2016, 7:40 PM
    if not I’ll look into your approach @rough-minister-58256
  • e

    enough-analyst-54434

    05/13/2016, 7:40 PM
    nono
  • e

    enough-analyst-54434

    05/13/2016, 7:40 PM
    @rough-minister-58256 and I speak of the same thing
  • m

    mysterious-action-44812

    05/13/2016, 7:40 PM
    oh maybe i misunderstood?
  • e

    enough-analyst-54434

    05/13/2016, 7:40 PM
    Did not realize they had no directory listing
  • e

    enough-analyst-54434

    05/13/2016, 7:41 PM
    So my suggestion won't work
  • e

    enough-analyst-54434

    05/13/2016, 7:41 PM
    You'll need a script or custom plugin to download the artifact out of band
  • e

    enough-analyst-54434

    05/13/2016, 7:41 PM
    Or you need your own serving infra (bintray is convenient)
  • r

    rough-minister-58256

    05/13/2016, 7:41 PM
    yeah - we use something akin to bintray internally at twitter. a big apache index of sdists and bdists etc
  • r

    rough-minister-58256

    05/13/2016, 7:42 PM
    but we also manage ~thousands of deps
  • m

    mysterious-action-44812

    05/13/2016, 7:42 PM
    ic, yeah we just have 1-2 trying to avoid all that
  • r

    rough-minister-58256

    05/13/2016, 7:42 PM
    for ~4 deps and living on github, you could probably just commit the deps directly to the repo and use findlinks against the binary deps dir
  • e

    enough-analyst-54434

    05/13/2016, 7:42 PM
    yeah - 256k
  • e

    enough-analyst-54434

    05/13/2016, 7:42 PM
    github will be fine
  • e

    enough-analyst-54434

    05/13/2016, 7:43 PM
    but - rate limits
  • e

    enough-analyst-54434

    05/13/2016, 7:43 PM
    so you really do want bintray
  • m

    mysterious-action-44812

    05/13/2016, 7:43 PM
    for example we have one side product producing jars for the main codebase and I wrote https://github.com/ActionIQ/archive-ivy-s3-resolver so that we could just push those to S3 and use that as a maven repo rather than managing a whole mvn repo or having to get bintray
  • e

    enough-analyst-54434

    05/13/2016, 7:43 PM
    bintray is free
  • m

    mysterious-action-44812

    05/13/2016, 7:43 PM
    oh ye forgot its free for open sourjce
  • m

    mysterious-action-44812

    05/13/2016, 7:43 PM
    sg lemme try that
  • e

    enough-analyst-54434

    05/13/2016, 7:43 PM
    for OSS licensed artifacts
1...129612971298...2420Latest