Updated config to match a conversation knut and i had about the relationship between global, site, and integrator configuration
(Will update docs)
Renamed wrapEnabled to wrap
Poor man's caching for calculateTextDimensions, wrapLabel, and breakString (actually makes a huge difference)
addition: bounds.init will clear models
addition: added loop model widths id instead of using title as the id
removed parseMessage debug message
addition: added configuration labelBoxWidth and labelBoxHeight for loop label box width/height
change: adjusted svgDraw drawText to support anchor and valign and whether to output a tspan
change: drawText returns an array regardless of array size
change: hardcoded label width/height uses conf.labelBoxWidth and conf.labelBoxHeight
change: Math.round() on many of the calculations to clean up bounds values
change: getTextObj anchor, width, height start as undefined
Fixed default config clobbering issues
Updated/corrected sequenceDiagram.spec to set the config using mermaidAPI
Enabled freeze on mermaidAPI to protect defaultConfig
Added inline config and init(ialization) grammar
Added reinitialize functionality to mermaidAPI (not to be confused with initialize)
Added actorFontWeight, noteFontWeight, messageFontWeight, wrapEnabled, wrapPadding
Added wrapLabel and breakWord functions to intelligently wrap text based on a pixel-based width instead of column-based
- The implementation is largely from Carys Mills: https://medium.com/@CarysMills/wrapping-svg-text-without-svg-2-ecbfb58f7ba4
- Made slight modifications for mermaid-js
Fixed dark theme color inconsistencies for sequence diagrams
Removed !important from sequence scss as this prevents any client overrides
Fixed various invalid css values in sequence scss which prevented proper rendering of various elements
Added detectInit to utils for initialization json detection
Updated detectType to support the existence or absence of the intialization configuration
Updated calculateTextWidth to include fontWeight
Previous expression with noteWidth does not make sense as noteWidth is
not used in case of note over two participants - forceWidth is used
instead. It tried to be symmetrical over cases A,B and B,A but failed.
Can be seen with messages over conf.width size (cannot reproduce in the
live editor as it seems to use not the last version where the code
does not use noteWidth).
The changed code explicitly calculates which of the actors is the
"left" one and starts rendering half margin to the left of its cetral
line.
This commit fixes some bugs, and I believe, improves upon the current
implementation.
In no particular order, it adds:
1. Control over note font size, family and alignment (now defaults to
center)
2. Dynamic actor resizing - actor's width will now scale if its
description is bigger than the static configured width
3. Dynamic actor margins - the margin between actors will now be
dynamically calculated by taking into account the width of connecting
messages or notes
4. Fixed a small visual annoyance where a loop arrow would intersect
with the text it loops on
5. Fixed a bug where if global config -> fontFamily wasn't defined, it
would override the actorFontFamily with an undefined
6. Removed some stale / commented out code
7. Added missing config variables to the global config object in mermaidAPI.js
8. Added messageFontSize, messageFontFamily to control message (non-note)
font settings
9. Memoized the actor widths in a pre-calculation that takes notes and
signals lengths into account
10. Removed redundant console.log lines
11. Extracted out actor width & margin calculation to getMaxMessageWidthPerActor, and
calculateActorMargins
This commit allow user to write multiline messages in sequence diagram.
For instance:
```
sequenceDiagram
Alice->>+John: Hello John<br>How are you?<br>I mean...<br>Are you OK today?
Alice->>+John: John, can you hear me?
John-->>-Alice: Hi Alice, I can hear you!<br>Keep calm!<br>I'm there.
John-->>-Alice: And I feel great!
```
This extends the `br` tag regex to accommodate an arbitrary amount of
any type of whitespace instead of a single space character, which holds
more closely to the HTML spec[1]
This extends the fix to #702 added in #1153
[1]: http://www.w3.org/TR/REC-xml/#sec-starttags
