* Fix error conditions for setParent #4989
* Error now thrown when calling `a.setParent(b)` on non-null b if the output/previous connection of a is not connected to an input/next connection of b without removing block from old parent's child list.
* Commented out error for when calling `a.setParent(null)` if a is connected to superior block.
* Adjusted comment to reflect that blocks were no longer being disconnected in this method.
* Also changed == to === per #4924 (`newParent` and `this.parentBlock_` must both be instances of `Blockly.Block`).
* Fix error conditions for setParent #4989
* Error now thrown when calling `a.setParent(b)` on non-null b if the output/previous connection of a is not connected to an input/next connection of b without removing block from old parent's child list.
* Error now thrown when calling `a.setParent(null)` if a is connected to a superior block.
* Adjusted comment to reflect that blocks were no longer being disconnected in this method.
* Also changed == to === per #4924 (`newParent` and `this.parentBlock_` must both be instances of `Blockly.Block`).
* Fix error conditions for setParent #4989
* Error now thrown when calling `a.setParent(b)` on non-null b if the output/previous connection of a is not connected to an input/next connection of b without removing block from old parent's child list.
* Commented out error for when calling `a.setParent(null)` if a is connected to superior block.
* Adjusted comment to reflect that blocks were no longer being disconnected in this method.
* Also changed == to === per google#4924 (`newParent` and `this.parentBlock_` must both be instances of `Blockly.Block`).
* Fixed lint errors.
* Fix error conditions for setParent google#4989
* Error now thrown when calling `a.setParent(b)` on non-null b if the output/previous connection of a is not connected to an input/next connection of b without removing block from old parent's child list.
* Commented out error for when calling `a.setParent(null)` if a is connected to superior block.
* Adjusted comment to reflect that blocks were no longer being disconnected in this method.
* Also changed == to === per google#4924 (`newParent` and `this.parentBlock_` must both be instances of `Blockly.Block`).
* Fixed lint errors.
* Adjusted comment.
* Removed unnecessary set to null/added tests
* One is failing (commented out), will investigate later
* Lint fix
* Removed failing test that correctly fails
* Update comments to conform to style guide
Capitalize first letter, period at end
An intermediate version of this code read:
fs.mkdir(demoStaticTmpDir, {recursive: true}), done);
but apparently `fs.mkdir` doesn't honour the recursive option, so I
tried to revert the change but munged it instead.
This commit cleans up the mess I made.
npm test has been failing if previous build output was present,
because the generated build/msg/??.msg files are missing a trailing
newline and the line to exclude build/* from being linted was
inadvertently omitted from commit 9e72378.
(Much of the rest of the build output would also fail lint checks but
was already excluded by the *_compressed*.js exclusion.)
Unlike goog.provide packages, goog.module modules are loaded using
XMLHttpRequest in uncompiled mode. Because browsers treat file: URLs
as each being a separate, unique origin, this means that CORS rules
prevent goog.module modules being loaded from a file URL.
To work around this, use the http-server npm to serve the repository
root directory on localhost:8080.
Provide a script so that you can type `npm start` to start the local
http server and automatically open a browser window pointing at
tests/playground.html.
Modify the setBackgroundColour function in the playgrounds to provide
the usual lilac background on locahost URLs and a sickly green one
if using a file: URL, all of which will soon stop working.
(The background doesn't actually appear to be visible in the advanced
playground.)
I had forgotten that I needed to change the value of BUILD_DIR
in several different places.
Added comments warning future editors about this as well as filing
issue #5007 to track fixing this properly.
Despite being misconfigured and therefore failing, the typescript
and metadata test scripts were exiting with status 0, indicating
successful completion. These have been fixed so they should fail
on any error, including misconfiguration.
- Reconfigure package_tasks.js to use BUILD_DIR from config.js
(i.e., build/) rather than repository root as source of files to
package.
- Add check to packageTasks.package to ensure that certain required
files exist in BUILD_DIR, to verify that buildTasks.build and
typings.typings have been run.
- Fix packageUMDBundle to use generated, rather than checked-in
version of en.js.
- Fix packageDTS to use the generated (rather than checked-in)
versions of blockly.d.ts and msg/*.d.ts.
- Modify run_all_tests.sh to run packageTasks.package before running
node tests, since they depend on it. Previously this was only
working because 'npm install' runs the 'prepare' script, which would
run the 'package' script - so the code being tested by the node tests
was not the current source but whatever precomipled code had
previously been checked in.
- Remove the 'prepare' script from package.json, since it is no longer
needed (and is now broken, since it requires that build and typings
have been done first.) Note that no scripts at all are included in
the version of package.json that is created in dist/ and subsequently
included in the published npms, so this deletion does not affect what
happens when the Blockly npm in installed - it only affects what
happens when 'npm install' is run after the blockly repo is cloned.
- Factor out the actual recompilation steps from releaseTasks.recompile.
- Rename releaseTasks.recomple to recompileDevelop, since it deals
specifically with the develop branch. (The npm script name is
unchanged.)
- Ensure that a full recompile and repackage is done before publishing
(beta and non-beta) npms.
