Changelog for df1-0.4
Version 0.4
-
COMPILER ASSISTED BREAKING CHANGE: Change functions names
Df1.render
toDf1.Render.log
,Df1.renderColor
toDf1.Render.logColorANSI
, andDf1.parse
toDf1.Parse.log
. -
Exported
key
,message
,iso8601
,segment
andvalue
from moduleDf1.Render
.
Version 0.3.2
-
Added
ToValue
instances for common types likeInt
,Bool
, etc. -
Changed colours in
Log
rendering.
Version 0.3.1
- Added missing language pragma.
Version 0.3
- The
segment
,key
,value
andmessage
functions are now methods of the newToSegment
,ToKey
,ToValue
andToMessage
classes. Identity instances, as well as instances forString
, lazyText
and strictText
, have been introduced for these classes.
Version 0.2
-
BREAKING CHANGE:
Segment
,Key
,Value
andMessage
don't strip surrounding whitespace anymore. When renderingSegment
,Key
andValue
, the whitespace will be percent-encoded. When renderingMessage
, the whitespace will be kept as is. -
BREAKING CHANGE:
Segment
andKey
now wrap lazyText
, rather than strictText
. This is to align their APIs withValue
andMessage
, which already wrapped lazyText
so as to prevent loggedValue
s andMessage
s from to use much memory. It's unlikely thatSegment
s andKey
s are affected by this, since in practice they are almost always created statically. So, this change is mostly to make the API less surprising to users: LazyText
is used throughout. -
Added draft BNF specification.
Version 0.1.2
-
Fixed escaping of control characters in
Message
. -
Percent-escape less punctuation characters when rendering
Key
,Segment
andValue
.TODO: write spec.
Version 0.1.1
- Fixed compilation.
Version 0.1
- Consider this a preview release: The API is likely to stay stable, but extensive testing, formalization and tooling is due.