#general
- w
witty-crayon-22786
10/09/2015, 6:30 PM(to be clear: by "workspace", you mean the one in .pants.d) - e
enough-analyst-54434
10/09/2015, 6:30 PMyes - w
witty-crayon-22786
10/09/2015, 6:30 PM@enough-analyst-54434: but also cacheable - e
enough-analyst-54434
10/09/2015, 6:31 PMyes - but we'll by 10x slower than go even after that - e
enough-analyst-54434
10/09/2015, 6:31 PMwe'll always be appriciably slower than go - b
bright-dinner-81701
10/09/2015, 6:32 PMso it won’t be so bad for node if i make a couple changes - b
bright-dinner-81701
10/09/2015, 6:33 PM1) i dont want to resolve npm_remote_module targets anymore - in fact, they dont really serve a purpose since they arent individually cacheable - b
bright-dinner-81701
10/09/2015, 6:33 PM2) we shouldnt fully resolve node_module targets without a prepublish - e
enough-analyst-54434
10/09/2015, 6:34 PM1 makes sense to me - I originally had a little extra logic to only resolve remotes if in the target roots - ie: I want to do
to play with typ./pants repl contrib/node/3rdparty:type - e
enough-analyst-54434
10/09/2015, 6:35 PMvs I want to repl a local module that transitively needs typ - e
enough-analyst-54434
10/09/2015, 6:35 PMSounds like you want to go further and kill the
target altogethercontrib/node/3rdparty:typ - e
enough-analyst-54434
10/09/2015, 6:36 PMWhich is sounding like it may make sense - e
enough-analyst-54434
10/09/2015, 6:36 PMREPLing a 3rdparty dep would still be nice, but there could just be a flag for that, `./pants repl.node --extra-deps=typ - e
enough-analyst-54434
10/09/2015, 6:37 PMor something - b
bright-dinner-81701
10/09/2015, 7:29 PMexactly - b
bright-dinner-81701
10/09/2015, 7:29 PMi was thinking about killing it altogether like you were saying - b
bright-dinner-81701
10/09/2015, 7:30 PMit’s only advantage is that it makes it easier to standardize on particular versions of 3rd party modules across your codebase - b
bright-dinner-81701
10/09/2015, 7:30 PMwhich seems like it has some value - however, every JS developer ive talked to says “whats the point?" - b
bright-dinner-81701
10/09/2015, 7:31 PMand if thats not useful, then why bother with all the boilerplate of the extra targets? - w
witty-crayon-22786
10/09/2015, 7:33 PM"however, every JS developer ive talked to says 'whats the point?'" - that's not necessarily because they know any better =P - w
witty-crayon-22786
10/09/2015, 7:33 PMbecause obviously shrinkwrap is useful - e
enough-analyst-54434
10/09/2015, 7:33 PM=Stu, but agreed that controlling via 3rdparty doesn't work here - e
enough-analyst-54434
10/09/2015, 7:33 PMwe need another means - w
witty-crayon-22786
10/09/2015, 7:33 PMand it is a form of shrinkwrap - b
bright-dinner-81701
10/09/2015, 7:34 PMits kind of a form of shrinkwrap, except it doesnt control the 3rd party module’s transitive deps - w
witty-crayon-22786
10/09/2015, 7:34 PMso do that - w
witty-crayon-22786
10/09/2015, 7:34 PMlike with go. - e
enough-analyst-54434
10/09/2015, 7:34 PMyeah - I'm not a fan of the half way - e
enough-analyst-54434
10/09/2015, 7:34 PMI'd rather fully shrinkwrap - w
witty-crayon-22786
10/09/2015, 7:34 PM+1