Babel 6 and dep updates #63

rosskevin · 2016-06-16T20:23:07Z

Solves #62

Updates to babel 6
Updates all dependencies
One test assertion debugged/fixed
travis.yml added
Tests pass

What needs to be done (@jquense):

need you to look at doc generation - error on globalization during generation - I'm unfamiliar
if you want to run travis, login to https://travis-ci.org/ and turn on this repo. The config should be ready to go

jquense · 2016-06-16T20:38:18Z

.babelrc

-  ],
-  "plugins": ["object-assign"]
+  "plugins": [
+    "transform-runtime"


no need for the runtime, i'd like to not rely on it

jquense · 2016-06-16T20:41:10Z

Thanks for the PR!

the issue with the docs is not a simple one unfortunately, and why this has been mostly stuck at babel 5. I'll take a look at it though.

rosskevin · 2016-06-16T20:57:50Z

Changed two things, tests pass. Didn't change the big swath of updated dependencies - can you be more specific on which one breaks what or should I just leave as-is for the moment?

jquense · 2016-06-16T21:01:16Z

package.json

+    "topeka": "^0.3.5",
+    "uncontrollable": "^3.2.4",
+    "warning": "^3.0.0",
+    "yup": ">=0.18.3"


sorry similarly with yup, and topeka

jquense · 2016-06-16T21:02:59Z

The react-bootstrap may be fine its just used for the docs, you can leave that one, but the direct deps anything that makes a major version bump (or minor bump if < 1.0.0) is a breaking change, and i don't quite trust the tests enough to rely on that (working on that).

rosskevin · 2016-06-16T21:09:55Z

Oh I see, I don't regard minor && patch versions as significant breaking changes (especially < 1.0), these are usually desperately needed fixes. So I take the opposite approach.

Since this is going against master, and I regard master as no guarantee of stability, I think we should we go ahead with fully up to date dependencies until we find an actual problem. At that point, we can revert a single one or more preferably update the tests and code to work.

You can pick and choose as you like, but I have better luck staying up-to-date.

jquense · 2016-06-16T21:33:52Z

the thing is those deps are also owned by me. so I know they will break edge cases :P which is why they aren't updated yet. I would say tho it's a fairly common practice (in node land) to count the minor bump as breaking pre 1.0.0 you might want to watch out for that

jquense · 2016-06-16T21:35:59Z

(I ask know they don't contain major bug fixes so don't worry there)

I've got some cod elocally that will take care of those Dependencies anyway :)

rosskevin · 2016-06-16T21:49:40Z

Great! I'd like to depend on this and help maintain so getting everything up to date is a priority for me. I'll submit anything as a PR that I run into.

jquense · 2016-06-16T21:59:10Z

thanks 👍

taion · 2016-06-16T22:50:28Z

package.json

@@ -54,7 +54,6 @@
  },
  "devDependencies": {
    "@monastic.panic/component-playground": "^2.0.0",
-    "babel": "^6.5.2",


You still need babel-cli to build... that's a valid dev dep

rosskevin · 2016-06-17T20:08:23Z

Working on the merge.

…uense#60 (comment)

# Conflicts: # package.json

rosskevin · 2016-06-19T13:30:49Z

Master still has a three tests that don't pass, so it is the same in this PR.

jquense · 2016-06-19T14:02:21Z

package.json

+    "eslint": "^2.12.0",
+    "eslint-plugin-react": "^5.1.1",
+    "file-loader": "^0.8.5",
+    "globalize": "^1.1.1",


globalize was intentionally left at x.1.0 the API is vastly different then which is why those bits fail. I'd put that back

rosskevin · 2016-06-19T21:59:17Z

docs now build

rosskevin · 2016-06-20T16:33:58Z

I think this is good to merge to master. @jquense do you want to check it out now that docs are building?

taion · 2016-06-20T19:40:47Z

package.json

@@ -2,6 +2,7 @@
  "name": "react-formal",
  "version": "0.17.9",
  "description": "Classy HTML form management for React",
+  "jsnext:main": "src/index.js",


This is not correct – jsnext:main should be code transpiled to use ES modules rather than CJS modules, not the untranspiled source.

I'm not following you. The source appears to me to be ES2015, and tools that acknowledge jsnext:main can consume it as part of their transpilation pipelines. So as far as my experience with rollup goes, I need to do nothing to the src and this will work fine. Can you point to my misunderstanding?

The source is not ES2015 – it's whatever is valid Babel stage 0.

Moreover the target for jsnext:main isn't ES2015 anyway – it's ES5 with ES modules (rather than CJS modules).

the big thing here is that the source uses jsx, which isn't valid javascript at all in its uncompleted form

Let me restate then. jsnext:main isn't a standard, it's a pseudo and not one that is fully adopted. I don't think we want to bike shed it into jsnext:main:jsx or something like that in order to allow an app to use it. I see the point of wanting to constrain it to ES only, but what ES only app (have to assume it is react, though not necessarily jsx) would use this UI-only library in a transpilation pipeline that couldn't handle JSX? I realize this is a bigger and more complicated conversation, but lacking another meta indicator, why not use this until we have another way? As-is, I still have to go out of my way to include this in my webpack pipeline.

you are right we don't want bikeshedding anything, we don't need an indicator for general untranspiled code, the src folder signals enough.

JSNEXT is specifically for signally to environments and bundlers where the es module versions of your code is as an interminable solution until transpilation of modules isn't needed because they are natively supported. using it however we feel defeats the usefulness and validity of the convention ppl are trying to establish to solve a specific problem.

not sure what you mean about going out of your way to include this with webpack though... are you saying you can't just require it? you shouldn't have to configure anything to use this.

Unfortunately, webpack doesn't look to resolve jsnext:main: webpack/webpack#1979 (comment)

I'm prototyping some local resolver code that will add these to my src loader pipeline, but it is really naive and only one level down. Without some kind of indicator, I'll be leaving out react-formal. Not a big deal, I'm just looking for a way to move forward.

It's not clear why you'd even need webpack to look at that it already understands main. The code is already compiled there is nothing else to be done. Eventually you can get things like tree shaking in webpack 2 but again the point is not to have a general entry point for uncompiled code, but as a way to say "this is where code is that's in the es module format"

I'm looking at tree shaking and general benefits of navigating code while developing. My local workaround failed miserably, so we are pretty much stuck until webpack catches up and some standardization happens anyway.

This reverts commit 6b30e2f.

jquense · 2016-06-21T19:00:24Z

package.json

-    "react-dom": "^15.0.1",
+    "less": "^2.7.1",
+    "less-loader": "^2.2.3",
+    "lodash": "^4.13.1",


This breaks doc code, you need to change the imports in metadata.jsx

I reverted these.

jquense · 2016-06-21T19:04:43Z

last few comments and then gtg, please Squash the commits as well. Many thanks

rosskevin · 2016-06-21T19:13:43Z

You push merge and squash:
https://help.github.com/articles/about-pull-request-merge-squashing/

rosskevin · 2016-06-21T20:02:21Z

Are you going to push the built lib files to master or cut another release?

rosskevin added 3 commits June 16, 2016 15:05

babel and dependencies updated

3ce6d50

update failed test assertion

cb2d4d7

add travis

c695116

jquense reviewed Jun 16, 2016
View reviewed changes

remove transform-runtime, downgrade react-input-message to 0.9.2

c83a042

jquense reviewed Jun 16, 2016
View reviewed changes

'babel' removed, now it's babel-cli expected to be installed globally

438331c

taion reviewed Jun 16, 2016
View reviewed changes

rosskevin added 2 commits June 16, 2016 17:59

add babel-cli

aa77afb

unnecessary to use passPerPreset

7d673e1

rosskevin added 2 commits June 17, 2016 15:22

merged latest master - test failures on master still present here: jq…

1c8acec

…uense#60 (comment)

Merge branch 'master' into babel-6-and-dep-updates

6d6984b

# Conflicts: # package.json

rosskevin mentioned this pull request Jun 19, 2016

Customized rendering - implementing with material-ui #60

Closed

jquense reviewed Jun 19, 2016
View reviewed changes

reverted globalize version

7cc7f1f

add jsnext:main for source resolution

6b30e2f

taion reviewed Jun 20, 2016
View reviewed changes

Revert "add jsnext:main for source resolution"

02b8e34

This reverts commit 6b30e2f.

jquense reviewed Jun 21, 2016
View reviewed changes

reverted lodash/less dependencies for docs

5a0e527

jquense merged commit 17c1ce0 into jquense:master Jun 21, 2016

rosskevin deleted the babel-6-and-dep-updates branch June 21, 2016 19:21

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Babel 6 and dep updates #63

Babel 6 and dep updates #63

rosskevin commented Jun 16, 2016

jquense Jun 16, 2016

rosskevin Jun 16, 2016

jquense commented Jun 16, 2016

rosskevin commented Jun 16, 2016

jquense Jun 16, 2016

jquense commented Jun 16, 2016

rosskevin commented Jun 16, 2016 •

edited

jquense commented Jun 16, 2016

jquense commented Jun 16, 2016

rosskevin commented Jun 16, 2016

jquense commented Jun 16, 2016

taion Jun 16, 2016

rosskevin commented Jun 17, 2016

rosskevin commented Jun 19, 2016

jquense Jun 19, 2016

rosskevin commented Jun 19, 2016

rosskevin commented Jun 20, 2016

taion Jun 20, 2016

rosskevin Jun 20, 2016

taion Jun 20, 2016

jquense Jun 20, 2016

rosskevin Jun 20, 2016 •

edited

jquense Jun 20, 2016 •

edited

jquense Jun 20, 2016 •

edited

rosskevin Jun 20, 2016 •

edited

jquense Jun 20, 2016

rosskevin Jun 20, 2016 •

edited

jquense Jun 21, 2016

rosskevin Jun 21, 2016

jquense commented Jun 21, 2016

rosskevin commented Jun 21, 2016

rosskevin commented Jun 21, 2016

Babel 6 and dep updates #63

Babel 6 and dep updates #63

Conversation

rosskevin commented Jun 16, 2016

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jquense commented Jun 16, 2016

rosskevin commented Jun 16, 2016

Choose a reason for hiding this comment

jquense commented Jun 16, 2016

rosskevin commented Jun 16, 2016 • edited

jquense commented Jun 16, 2016

jquense commented Jun 16, 2016

rosskevin commented Jun 16, 2016

jquense commented Jun 16, 2016

Choose a reason for hiding this comment

rosskevin commented Jun 17, 2016

rosskevin commented Jun 19, 2016

Choose a reason for hiding this comment

rosskevin commented Jun 19, 2016

rosskevin commented Jun 20, 2016

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rosskevin Jun 20, 2016 • edited

Choose a reason for hiding this comment

jquense Jun 20, 2016 • edited

Choose a reason for hiding this comment

jquense Jun 20, 2016 • edited

Choose a reason for hiding this comment

rosskevin Jun 20, 2016 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rosskevin Jun 20, 2016 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jquense commented Jun 21, 2016

rosskevin commented Jun 21, 2016

rosskevin commented Jun 21, 2016

rosskevin commented Jun 16, 2016 •

edited

rosskevin Jun 20, 2016 •

edited

jquense Jun 20, 2016 •

edited

jquense Jun 20, 2016 •

edited

rosskevin Jun 20, 2016 •

edited

rosskevin Jun 20, 2016 •

edited