"51.0.2704.37 beta
Alex.03.61.0735.0056G3.0021"

TEST CASE 

"The text navigation levels are:
    - Group
    - Object
    - Line
    - Word
    - Character

Using the up/down arrows will navigate the current level, and using the left/right arrows will navigate the next more specific level.  

Default behavior is ""groups"" for up/down and ""objects"" for left/right.  When up/down is ""characters"", left/right is still ""characters"".

For example, navigating the words of ""test phrase"" will read ""test"" first then (after navigating again) ""phrase"".  Navigating the characters of that phrase will read out ""t"" then ""e"", etc.


1. Navigate to the wikipedia page for cat in your language.  If there isn't one, pick another wikipedia page or website in your language which contains a lot of text.

2. Use search+shift+up and search+shift+down to navigate the page.
Verify: the navigation level matches expected behavior described below (""group"" by default).

3. Use search+shift+left and search+shift+right to navigate the page.
Verify: the navigation level matches expected behavior described below (""object"" by default).

4. Use search+shift++ to increase the granularity.  Repeat #2 and #3 until there is no finer granularity (i.e. ""character"").

5. Press search+shift+- until the granularity returns to the default ""group"" level.


Verify:
    - All words read should be in the correct language.
    - The division level should match whatever makes sense for this language.  
    - Use search+shift+- to decrease the granularity back to the default ""groups""


Group:
Navigation occurs by paragraphs or lists.  If the list contains a classification side panel, each line of the classification is highlighted separately (but not the individual words).  In the table of contents, list with sublists will be selected one by one, but lists that have no sublists will be grouped together, e.g. ""list with 5 items"".

Object:
Navigation separates links, bold or italic phrases, list items, and all text in between.  Unlike groups, each link or item in the classification side panel will be read individually.  Also unlike groups, a list in the table of contents with no sublists will have each item selected one by one.

Line:
Navigation reads an entire line of text, regardless of sentence structure or objects contained within the line.

Word:
An entire word.  If a word is in a link, following punctuation might be ready separately, e.g. ""link"" ""text"" ""period"" for ""<link text>.""

Character:
- A single letter."

RESULTS 

"Issue 1 (GOOGLE-3971): When in the Character level, the voice reads all the numbers in English, except 0 (zero) which is in Italian. 

 Issue 2  (GOOGLE-3973): When in the Line level, if the line starts with regular text and then has text with link that falls on a separate line, the voice reads the first line only partially, stopping before the link, and including the link text in the second line. 

 Issue 3  (GOOGLE-3976): In the Word level, the voice reads punctuation marks inconsistently. Periods are read, but commas aren't always read. Parentheses are not read. "