It seems that when ObserveSequence observed a Cursor, it used to stop() the observe when the main autorun was invalidated, creating a “gap” during which no callbacks would be received (or fired). This was before we used Deps.nonreactive in the main autorun (shielding the cursor.observe from being stopped). Now the observe is only stopped upon re-run or when the ObserveSequence is stopped, and there is no gap.
Removed references to this gap from comments.
I believe the current code is correct, and in addition, we could now optimize the cursor-to-same-cursor case (and basically not do anything if seq===lastSeq and is a Cursor, i.e. not stop the old handle, create a new handle, or diff).
This was originally introduced with
f1b77fec96, but it looks
like all of our tests now pass. (Maybe eliminating `rewind` in
b5a0613f85 made this no
longer necessary?)
If we find that this commit did break something, let's make sure to
add a failing test before reverting.
It seems like a good practice for tests to always be grouped under their package name, so that it’s easy to find the tests when they fail. It’s confusing to have the tests in packages like “spacebars-tests”, “spacebars-compiler”, and “templating” be in various groups like “spacebars - templates” and “spacebars - templating”.
I didn’t go through and enforce this convention in all packages, just a few, but I think it would be good and not too hard to do the rest sometime.
We now call 'removedAt' and 'changedAt' with indices (instead
of 'removed' and 'changed'). While at it, fix a bug where
the 'movedTo' callback was firing with incorrect
indices.
commit f44d03d2c8bbfca32ae08fdabef35472c62042eb
Author: Avital Oliver <avital@thewe.net>
Date: Tue Apr 15 14:54:29 2014 -0700
A more complicated movedTo test that should catch a bug
commit 24f9ba3ffc708cf1af34e434ddf033c54d246614
Author: Avital Oliver <avital@thewe.net>
Date: Tue Apr 15 14:05:44 2014 -0700
Update tests and {{#each}} for new observe-sequence
This breaks a strong dependency between the webapp package and the
bundler. Now we can change the standard page format, add more hooks,
etc without changing the tools (and hot code push works properly).
This is an incompatible change to the definition of the
browser-program-pre1 format: it no longer contains a "page" field
pointing to a boilerplate defined using ad hoc ##FOO##
substitution. Instead, the manifest can contain "head" and "body"
entries for data added with compileStep.appendDocument(). (Note that in
Blaze, <body> in a template file no longer calls appendDocument.)
WebApp.addHtmlAttributeHook callbacks now return attribute objects (like
those that Spacebars supports) rather than strings.
- If the items in an array are strings or numbers, use those as
the id.
- If the items in an array are objects with no '_id' field, use
the index in the array.
- In any case, if the id to be used is already present in this array,
generate a random replacement value and print a warning.
