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

    rough-minister-58256

    09/09/2015, 6:47 AM
    invoking pex with -c implies the caller understands they’re basically computing a product that will act for all intents and purposes like the -c executable
  • e

    enough-analyst-54434

    09/09/2015, 6:47 AM
    Making the weird to me and him behavior moot
  • r

    rough-minister-58256

    09/09/2015, 6:48 AM
    yeah
  • r

    rough-minister-58256

    09/09/2015, 6:48 AM
    exactly
  • r

    rough-minister-58256

    09/09/2015, 6:48 AM
    without that first thing in the docs, this would have never come up
  • r

    rough-minister-58256

    09/09/2015, 6:48 AM
    unless you’ve wrapped your head around it and done it yourself
  • e

    enough-analyst-54434

    09/09/2015, 6:48 AM
    Agreed
  • r

    rough-minister-58256

    09/09/2015, 6:49 AM
    https://github.com/pantsbuild/pex/issues/119 <— task for publishing pex as a pex
  • r

    rough-minister-58256

    09/09/2015, 6:51 AM
    if we kill the docs and then just have pex distributions for the tool itself (which afaik contain no native libs), we shortcut the awkward acts a person would have to go through for this w/ all the gains
  • r

    rough-minister-58256

    09/09/2015, 6:52 AM
    and in fact then its even cleaner/faster to get started with pex
  • r

    rough-minister-58256

    09/09/2015, 6:52 AM
    dl the tool, drop it in, go
  • r

    rough-minister-58256

    09/09/2015, 6:53 AM
    no pip/easy_install necessary
  • r

    rough-minister-58256

    09/09/2015, 6:55 AM
    poor dan will probably never touch pex again after this bug tho
  • r

    rough-minister-58256

    09/09/2015, 6:55 AM
    mindbending, for sure
  • l

    limited-country-20626

    09/09/2015, 3:27 PM
    Hey twitter folks - what was your experience with twitter-commons? Is putting a commons-style repo something you would do again?
  • w

    witty-crayon-22786

    09/09/2015, 3:50 PM
    yes, it's something we plan to do again
  • w

    witty-crayon-22786

    09/09/2015, 3:51 PM
    the primary reason it went out of sync was that we merged our monorepos
  • w

    witty-crayon-22786

    09/09/2015, 3:51 PM
    once the dust settles, i absolutely want to resume syncing out our open source that way. but it will be much, much larger
  • l

    limited-country-20626

    09/09/2015, 3:55 PM
    What were the advantages do you see from it?
  • w

    witty-crayon-22786

    09/09/2015, 4:02 PM
    realistic OSS testbed for pants
  • w

    witty-crayon-22786

    09/09/2015, 4:02 PM
    easier cross-project contribution to twitter oss projects
  • w

    witty-crayon-22786

    09/09/2015, 4:02 PM
    (in that everything can be integration tested together)
  • w

    witty-crayon-22786

    09/09/2015, 4:02 PM
    not needing two build systems for simple oss projects that don't want to go through the trouble of creating their own
  • w

    witty-crayon-22786

    09/09/2015, 4:03 PM
    consumption of "snapshots" of our entire stack, rather than dealing with semver
  • w

    witty-crayon-22786

    09/09/2015, 4:05 PM
    (sirois and i worked on an internal doc for this, but since then i've realized why something like https://github.com/jsirois/sapling is important)
  • w

    witty-crayon-22786

    09/09/2015, 4:07 PM
    for most of our open source projects, i don't expect that the oss monorepo would be their primary home
  • w

    witty-crayon-22786

    09/09/2015, 4:07 PM
    but for some it would
  • w

    witty-crayon-22786

    09/09/2015, 4:08 PM
    and they would get a guaranteed working build, and automatic syncs for free
  • l

    limited-country-20626

    09/09/2015, 4:26 PM
    thanks stu. 4sq is considering something similar, I wanted to hear your thoughts. We are moving some open source stuff into our monorepo (for an easier upgrade/consumption cycle) and rethinking how we push changes back to OSS.
  • w

    witty-crayon-22786

    09/09/2015, 4:27 PM
    yea
1...333334335...2420Latest