#general
- r
rough-minister-58256
03/28/2016, 5:48 PM@happy-kitchen-89482: no sign of
in our pluginsstore_false - w
witty-crayon-22786
03/28/2016, 5:52 PM@limited-country-20626: awesome. thanks mateo... cc @limited-terabyte-34878 - h
happy-kitchen-89482
03/28/2016, 5:52 PMAny objection to not supporting it, and not supporting explicit defaults for boolean options? One can always invert the sense of the flag. It would make the code simpler. Thoughts? - l
limited-country-20626
03/28/2016, 5:52 PMI say drop it - b
bored-art-40741
03/28/2016, 5:53 PM++ - l
limited-country-20626
03/28/2016, 5:53 PMHaving the False as a default value is much more readable imo - w
witty-crayon-22786
03/28/2016, 6:56 PMhey gang: rfc on what we're thinking for integrating the engine: - w
witty-crayon-22786
03/28/2016, 6:57 PMthe initial plan was that "integrating the engine" was going to mean 2.0.0. - w
witty-crayon-22786
03/28/2016, 6:57 PMas we've been able to trim the scope, we have... the first time was trimming to integrating only bootstrap/codegen into the engine before shipping it, and trying to target that for 1.2/1.3 without too many deprecations - w
witty-crayon-22786
03/28/2016, 6:58 PMnow we're considering trimming scope a bit further - w
witty-crayon-22786
03/28/2016, 6:58 PMin particular, if we allow for synthetic target creation in https://github.com/pantsbuild/pants/blob/master/src/python/pants/engine/exp/legacy/graph.py#L17 , we can allow it to be used before bootstrap/codegen is ported to the new engine - w
witty-crayon-22786
03/28/2016, 6:59 PMand since that wouldn't involve any backwards incompatibility, we could probably target it for 1.0.x or something - b
bored-art-40741
03/28/2016, 6:59 PMI think it's going to merit a minor bump at the very least - w
witty-crayon-22786
03/28/2016, 7:00 PMit would be behind an option most likely - b
bored-art-40741
03/28/2016, 7:00 PMI find it very unlikely that such a large change won't break anything, even if it was theoretically backwards compatible - b
bored-art-40741
03/28/2016, 7:00 PMI'd prefer to move "fast" to a new minor version with these kinds of iterations - b
bored-art-40741
03/28/2016, 7:00 PMAnd not fork into options - w
witty-crayon-22786
03/28/2016, 7:00 PMwhat would actually be shipping would be likely a single option to "1) use the new engine to parse the build graph and 2) enable the daemon" - b
bored-art-40741
03/28/2016, 7:01 PMRemember we'll have branches to work off of to test this stuff. It would work well for minor bumps - b
bored-art-40741
03/28/2016, 7:01 PMSpeaking of, we should probably get a 1.0.0-dev branch running soon so we can start cutting RCs - b
bored-art-40741
03/28/2016, 7:02 PMIf we're launching in a month we'll need to be testing RCs sooner! - w
witty-crayon-22786
03/28/2016, 7:02 PMyes, matt is working on the strategy. - w
witty-crayon-22786
03/28/2016, 7:02 PM - w
witty-crayon-22786
03/28/2016, 7:03 PM@bored-art-40741: the truth is: we've already implemented the graph such that it is backwards compatible, and so the majority of the work to allow for an option to use it is already done - w
witty-crayon-22786
03/28/2016, 7:04 PMwe wouldn't be saving an appreciable amount by not putting it behind an option - w
witty-crayon-22786
03/28/2016, 7:14 PMin any case, one of our next proposed steps is to clean up the
package, to make sure nothing that is only relevant toengine.exp
is in thereengine.exp.examples - w
witty-crayon-22786
03/28/2016, 7:14 PMand then, to explode
out into pants coreengine.exp - w
witty-crayon-22786
03/28/2016, 7:15 PM(https://github.com/pantsbuild/pants/issues/3106 for the first bit) - b
bored-art-40741
03/28/2016, 7:24 PMYeah, I'm fine with all of that--just suggesting that it's a natural pipeline to do the integration as minor releases - b
bored-art-40741
03/28/2016, 7:24 PMThe point of minor releases is that we'll have branches and RCs to let other orgs do realistic testing before we cut the final release