1.5.6

• jdbc: DB_READONLY=true env var can be used to make the whole app read-only (e.g. when migrating the DB, instead of Heroku maintenance mode)
• jdbc: CrudRepository.get() now has optional forUpdate parameter
• HttpExchange.fileName() added for setting of Content-Disposition
• json bugfix: json within json (and '') are now properly escaped
• json bugfix: render Any type as 'any' instead of 'string' in TSGenerator
• csv: new module for CSV parsing/generation

1.5.5

• jdbc: support reading of collections of Decimal from DB array columns
• jdbc: db.upsert() now has an optional "where" parameter
• server: AssetsHandler.headerModifier now gets file parameter to make decisions based on it
• server: allow having annotations on AssetsHandler for e.g. access checks

1.5.4

• json: ValueConverter can now be used to transform strings into types
• json: JsonNode.getList() signature fixed
• json: improve error message when trying to parse an empty stream
• json: JsonParser.readArray can now be used to stream json arrays without loading them into memory
• jdbc: close connection (return to pool) even if autoCommit/commit/rollback fails
• server: support for boolean query parameters without values, e.g. ?flag
• server: support for anonymous handler annotations (fixed in Kotlin 1.8)

1.5.3

• server: TypedHttpClient/JsonHttpClient uses a better logger name (from nearest user class)
• json/jackson: TypedHttpClient/JsonHttpClient now have overridable trimToLog property in case you need to process how requests/responses are logged
• jobs: JobRunner.schedule() now takes kotlin.time.Duration values
• jdbc: PooledDataSource - a simple and easy to configure connection pool (used in DBModule by default)
• jdbc: use HikariModule instead of DBModule if you still want to use Hikari (also add dependency on com.zaxxer:HikariCP)
• liquibase/serialization - dependencies updated

1.5.2

• core: experimental Decimal class to be used for monetary values, with numerical equality (unlike BigDecimal)
• core/json: use default values for explicitly passed nulls if property is not nullable
• core/json: unwrap InvocationTargetException, so that any validation exceptions thrown from data class constructors is propagated properly
• core/json: workaround for a bug in kotlin-reflect, which boxes null values even if nullable inline type is used: https://youtrack.jetbrains.com/issue/KT-57590
• json: TSGenerator can now receive additional library types to generate from command-line, e.g. klite.TSID
• i18n: do not trim translations by default (keep all whitespace)
• jdbc: support binding of Int values to data classes (DB usually returns integers as Long)
• jdbc: possibility to use table aliases when getting of columns from ResultSet with joins/using create (Postgres only)
• jdbc: allow whitespace between -- and keywords in changeset sql files
• server: default RequestLogFormatter will not log StatusCodeExceptions anymore
• server: useHashCodeAsETag() introduced to avoid sending of same responses
• server: run onStop handlers in reverse order of registration, so that e.g. connection pool is stopped after jobs are stopped
• jobs: do not start new jobs on shutdown while waiting for running jobs to finish

1.5.1

• json: ValueConverter.from() can now have access to the expected KType
• json: TSGenerator to generate TypeScript types for data classes/enums
• server: AppScope.async now is the standard async function that returns a Deferred. Use AppScope.launch if you want exceptions to be logged

1.5.0

• core: toValues() functions moved here from klite-jdbc
• core: TSID introduced as an alternative to UUID
• json: new lightweight json parser
• i18n: now uses the lightweight klite-json, not jackson
• jackson: package changed to klite.jackson to avoid conflicts with klite-json
• jackson: .parse extension function now passes type parameters to Jackson, not only the main class
• jdbc: fromValues() was renamed to create()
• jdbc: switched db.select() <-> db.query(), taking "where" as a list or varargs, to allow for duplicated columns

The release is not fully backwards-compatible, however most old functions are provided as @deprecated.

This will migrate the most important parts:
find -name '*.kt' -exec sed -ri 's/klite.json./klite.jackson./; s/mapOfNotNull/notNullValues/; /db\.(query|select)/{N; s/db\.query/db.xxxselect/g; s/db\.select/db.query/g; s/mapOf/listOf/g; s/emptyMap/emptyList/g};' {} \; -exec sed -ri 's/db\.xxxselect/db\.select/; s/(db.update\(.*, )mapOf\((.*?\)), (mapOf\(.*?\))\)/\1\3, \2/; s/(db.delete\(.*, )mapOf\((.*?)\)/\1\2/' {} \;

Beware: if you use the replacement above, then make sure that no deprecated query/select usages are left, i.e. check that all "where" maps are replaced with lists/varargs, which is done automatically only if on the same or next line.

You may also use IDEA migration of deprecations, but it will most likely break code formatting, etc.
Also, you may need to add some imports manually.