Commit graph

24 commits

Author SHA1 Message Date
Catrope 5e47bdc57f Maintain a transaction-wide offset adjustment
This fixes a bug where the second replace operation in a transaction
would cause the rebuild of the wrong range, or the adjustment of the
wrong text node.

Change-Id: I9b1c68d84999d538fe10bb193f4dfdd694121d2a
2012-06-06 15:34:55 -07:00
Trevor Parscal f6864b0c04 Changed use of "factory" to "nodeFactory" to make way for other kinds of factories
Change-Id: I96db0f28bb220f1c3b23990824e9523278cb8f9b
2012-05-31 15:20:58 -07:00
Catrope 2f18605a1a Fix buggy replace behavior when inserting content that contains nodes
Copy-pasting things like "text<IMAGE>moretext" failed spectacularly,
this commit fixes that.

* Check for content rather than structure in the inserted/removed data
* In the content case
** Run selectNodes() over the removal range, rather than just the cursor
*** i.e. no longer assume that content replacements only affect one node
** If there is structure involved, rebuild all affected nodes

Change-Id: I80e40b5b7c514a3fb105d57e4a17770d0fefaaea
2012-05-31 07:27:36 -07:00
Trevor Parscal 08765feab3 Merged ve.dm.DocumentFragment and ve.dm.Document
* Also fixed lots of whitespace issues and some global variable introductions

Change-Id: Ia6c82ef0c3b5c4938cf72afa86da6d79e4cb385a
2012-05-17 11:23:56 -07:00
Catrope 80db7a593e Rethink of structural replacement code
Splits and merges now work, or at least the tests for it pass

The strategy I used is to gather the affected ranges for each of the
following:
* removed stuff
* the entirety of each node touched by a non-zero removal
* if the inserted data busts out of its parent, the entirety of that
  parent node (the 'scope')
then get the covering range of all those ranges, and rebuild that.

Change-Id: I7c3b421abc0ba134157ac8b59042675bb1b5073c
2012-05-16 16:48:02 -07:00
Catrope 6f4b68274d Rename removeLevel to insertLevel
Change-Id: I339c830547d19af0b04927e1ebccd0468c9e2a20
2012-05-16 16:48:02 -07:00
Catrope d6cf4fff7f Add comment about bug that makes tests accidentally pass
Change-Id: I842ffe00bdad025797f9c71f550efd24d2ce39da
2012-05-14 23:16:28 -07:00
Catrope 7e1aa2336e Support splitting nodes in replace tree sync
This means inserting things like </p><p> are now synced correctly and
split the paragraph in the model tree. Merges (removing e.g. </p><p>)
aren't supported yet.

Also, this needs tests, Trevor tells me he's working on porting replace
tests from the old ve/ directory

Change-Id: Ic5050849d7d007a1696dc36548654979aedb53a8
2012-05-14 21:46:39 -07:00
Catrope 59f74de3b8 Fix bug causing replacement tree sync to adjust the wrong node
The tree sync for content replacements was adjusting the parent of the
text node affected, rather than the text node itself. This was because
it called getNodeFromOffset(), which returns branch nodes. Switched it
to use selectNodes() in leaves mode

Change-Id: I50a9be18151a1b75815ab19b787b16b6be385bf9
2012-05-14 21:42:53 -07:00
Trevor Parscal e28a1a188a Fixes typo, need to check if object is empty, not if it's an object
Related to: 14a4c7b780

Change-Id: Iede66d1e01ed010705362611e90fb4b13e7e4a1c
2012-05-14 17:22:37 -07:00
Trevor Parscal 14a4c7b780 Fixed annotation processing to support leaf nodes and detect more errors
Also added lots of tests

Change-Id: I45fe4fb569d9050d7d79f71f777d82f5b34fdd11
2012-05-14 17:16:30 -07:00
Catrope 5b398f95f8 Fix botched rename: opReplacement -> opInsert
Change-Id: I34091dbbc9a46466823ee7407f151fd67112ee03
2012-05-14 15:12:44 -07:00
Trevor Parscal e0de881a8a Removed trailing whitespace in multi-line comment blocks
JSHint and various git tools complain about this, so let's just do away with them altogether

Change-Id: I731866bd21f1ee0088fb0a71df3abf92692aca23
2012-05-14 15:05:09 -07:00
Catrope 43c2393c03 Remove outdated comment
Change-Id: I3a5db50aafa01d07bfb34359dfbe73812b3a7ab5
2012-05-11 16:03:46 -07:00
Trevor Parscal db0e050b39 Renamed uses of "replacement" to "insert"
Change-Id: I69ad0f84e8fefe1715339068b7685958f8f8546f
2012-05-11 15:25:01 -07:00
Catrope ca17ea0943 Port structural replace code from the old VE code
This makes TransactionProcessor work for regular replacements, as well
as insertions and deletions of self-contained pieces of data. This does
NOT yet work for inserting and deleting unbalanced data
(splitting/merging nodes).

I've tested this from the console for insertions and deletions and
simple replacements, but I haven't tested wrappings. We should write a
bunch of unit tests for this some time :)

Change-Id: Ic2fd75d1cf2e127bc9ae58debce67576be2c912f
2012-05-10 20:20:42 -07:00
Catrope 09631f1466 Make TransactionProcessor actually use DocumentSynchronizer
Change-Id: If0ac69026fa34430d811de5f7ca0748d9b1e01cc
2012-05-08 18:31:51 -07:00
Trevor Parscal 1d74945dfc Fixes for Roan's crazy half-baked commit
Change-Id: Ic104debc3a941e4cf8e4dca19633dcad04a0fe0a
2012-05-04 16:13:54 -07:00
Catrope 2baeb5532d Initial stab at DocumentSynchronizer
Change-Id: I7176fd7235a82a6314bf431d00870d9ddb40f370
2012-05-04 15:18:07 -07:00
Catrope 11a3b6886b Implement basic replace processor
* Implement basic TransactionProcessor.replace(), only does content
  replacements
* Add ve.dm.Document.containsElementData()
* Fix bug in attribute()
* Write basic tests for attribute() and replace()

Change-Id: Ie9c22aec3f2631be5b0bd66790408ad283565491
2012-05-04 11:09:10 -07:00
Trevor Parscal fee375cf4e Removed global variable introduction
Change-Id: I331b19cafeb20610f7b42bacaff20757a342dffc
2012-05-02 12:37:30 -07:00
Catrope d9cd1392fc Add TransactionProcessor tests, and fix TP to make them pass
* Add .process() and its helpers from ve1
* Fix applyAnnotations()
** Use data[i] rather than data[j]
** Don't add empty annotation objects for no reason
** Remove empty annotation objects
*** I thought this wasn't needed, but it is needed for clean rollbacks
* Remove unused second parameter in applyAnnotations() call

Change-Id: Ia338f62d2eaf2a76f8ef653eead05bc44757a122
2012-04-30 19:06:20 -07:00
Trevor Parscal c2d4a2d928 Added basic ve.ce nodes
* Also removed beforeSplice and afterSplice in favor of just plain splice which is the same as afterSplice used to be - beforeSplice was never used and it was making things more complex looking than needed

Change-Id: Icbbc57eac73a2a206ba35409ab57b3d1a49ab1a5
2012-04-30 16:58:41 -07:00
Catrope 80decda935 Mostly-finished TransactionProcessor
Change-Id: I7d2cd684fafa16c85d0d9c9ad2b6d42d7e6f1448
2012-04-30 11:26:22 -07:00