Changelog

This is the changelog for casual and all changes are listed in this document.

[1.7.0] - 2024-11-26

Added

  • xatmi: add function to reset execution_id from public api (#239)

  • queue: add –force flag to –remove-message cli (#267)

  • queue: casual queue –dequeue should take 0..* queue-ids

  • queue: introduce configurable ‘capacity’ property to groups (#232)

  • queue: runtime configuration

  • queue: local domain enable/disable enqueue/dequeue (#253)

  • queue: added –list-queue-instances -> where queues reside (#354)

  • queue: expand group concept to include forwards

  • configuration: documentation update for group.enabled

  • configuration: added CLI enable/disable groups

  • configuration: enabled on internal model server and executable

  • configuration: moved conf related cli from domain to configuration (#112)

  • service: replace --list-admin-services for --list-services -all (#426)

  • service: use random distribution for concurrent reservation (#420)

  • service: service timeout includes pending lookup time

  • service: show description for instances (remote domain name) (#353)

  • discovery: enable/disable enqueue/dequeue interdomain (#253)

  • discovery: prepared new 1.4 protocol version for discovery::Reply

  • transaction: rename cli option for scaling proxies (#229)

  • transaction: manager should wait for all spawned RM:s before on-line

  • transaction: runtime configuration (#225)

  • transaction: show description for instances (remote domain name) (#350)

  • code: conform to c++23

  • domain: cli add enabled to domain –list-servers/executables

  • domain: assassination contract lethal or non-lethal (#304)

  • gateway: honor group memberships on connections (#323)

  • gateway: runtime configuration

  • gateway: branch transaction in inbound (#328)

  • cli: improved --legend <option> to --legend --<option>

  • cli: add option to abort incomplete domain boot (#254)

  • cli: add keys for missing instances to casual domain –information

  • metric: propagate OTel parent span for service-call and conversation (#321)

  • event: log user return code in service metric event (#399)

  • documentation: added group enabled sample (#445)

  • documentation: overhaul README (#441)

  • conversation: execution span for conversation::connect::Request message (#321)

  • build: simplify install script for http-inbound (#22)

  • build: add dsl for documentation (#355)

  • build: use ubi8 isof centos

  • build: use local copy of nginx

Fixes

  • buffer: simplify the handling of the special ‘inbound’ buffer (#411)

  • buffer: zero sized buffers is not found in lookup table (#302)

  • queue: add ‘signaled’ error-code to queue api (#291)

  • queue: correct runlevel for queue-forward-group (#221)

  • service: fix routes dissapearing on config update (#431)

  • service: don’t reserve instance after fatal assassination event (#305)

  • service: refactoring service-manager - reference based -> index based (#407)

  • discovery: remove invalid sanity check (#443)

  • discovery: short-circuit fan-out discoveries -> reply directly on found (#372)

  • domain: respect ‘enabled’ in executable and server configuration (#324)

  • gateway: correct cli hints (#459)

  • event: fix casual-event-service-log segfaulting (#359)

  • http: outbound always send metric on service call (#397)

  • http: inbound handles lookup timeout

  • documentation: fix tutorial after updated cli (#453)

  • documentation: generated cli help for configuration missing

  • common: resetting system environment variables leaks memory on linux

Potentially breaking changes

CLI

  • service --list-instances column state, value remote has changed to external to be consistent with other CLI values. We use internal and external, where internal is stuff that is known locally to the “manager”, and external is stuff that is not known locally (could still be external within a domain).

  • transaction --list-instances in the 1.6 form has changed name to transaction --list-internal-instances. transaction --list-instances now lists all instances of all resources, internal and external. This makes it consistent with service –list-instances.

[1.6.22] - 2024-11-05

Fixes

  • service: discovery-known-request only reply with known discoverable services (#454)

[1.6.21] - 2024-10-09

Fixes

  • documentation: queue.forward.group.(services|queues).alias

  • transaction: error log when resource fail (#428)

[1.6.20] - 2024-09-18

Fixes

  • build: handle nginx build on rhel (#381)

  • service: fix invisible route being created when updating service config

  • queue: enqueue/dequeue to multiple queues in example server

[1.6.19] - 2024-04-25

Fixes

  • http: improved error logging in http-outbound (#367)

  • http: use route name in outbound when ack:ing service calls (#364)

  • queue: add example server with enqueue and dequeue services (#348)

  • documentation: improved configuration documentation

  • documentation: updated and removed invalid api-documentation

  • example: fix casual/example/work (#349)

[1.6.18] - 2024-03-18

Fixes

  • domain: print domain name in shutdown task (#336)

  • common: do not set environment when (re)setting execution-id (#333)

  • http: propagate execution id over http (#326)

  • gateway: fix markdown documentation for interdomain protocol (#319)

  • sf: ‘attribues’ is misspelled in describe model (#264)

  • domain: improve error messages (#315)

  • http: use a single correlation for entire service call (#318)

  • ipc: fix discard while message is only partially received (#311)

  • queue: clear available when moving message to error queue (#312)

  • domain: assassinate needs more precise logging (#309)

  • service: handle assassinate during shutdown (#299)

  • gateway: outbound sent discovery to disconnect-mode connections (#297)

[1.6.17] - 2023-11-13

Fixes

  • transaction: TM asserts when transaction is not found in prepare-handle (#288)

  • cli: header output in porcelain if –header true (#283)

  • service: perform discovery on external_discovery (#269)

  • xatmi: comply to specification signal handling (#271)

  • service: cli –list-services does not add - when contract is missing (#281)

  • cli: add precondition casual log can’t be tied to stdout (#259)

  • domain: propagate restart to configuration-get (#240)

  • domain: handle restarts attribute correct (#277)

  • transaction: cli - add legend for list-resources

  • transaction: show metric for resource-proxy pending requests in cli (#273)

  • transaction: fix resource-proxy metric (#273)

  • cli: fixed and simplified the transaction cli (#272)

  • queue: dequeue with id as selector should ignore available (#270)

[1.6.16] - 2023-10-10

Fixes

  • queue: send reply on failed dequeues (#262)

  • gateway: pending ‘calls’ could be sent after a rollback (#261)

  • buffer: buffer-field-serialize should generate with [[maybe_unused]] (#255)

[1.6.15] - 2023-09-14

Fixes

  • transaction: TM commits when it act as a resource and gets a prepare (#256)

[1.6.14] - 2023-09-03

Fixes

  • event: make sure we try to reopen the event file if it’s “broken” (#250)

  • log: do not throw if we can’t open logfile -> error to stderr (#250)

  • cli: add internal cli for generic message count metric

  • service: SM sends error-reply to no-reply request on timeout/core (#247)

  • documentation: add topology::Update to protocol documentation

  • service: improve handling of service timeout (config/cli) (#244)

[1.6.13] - 2023-08-22

Fixes

  • http: nginx could block caller for ever with payloads > ~50k (#242, #243)

  • configuration: prevent global contract setting being default overridden (#244)

[1.6.12] - 2023-07-18

Fixes

  • discovery: hops are not correctly propagated if services are known

[1.6.11] - 2023-06-22

Fixes

  • xatmi: error log unexpected error codes (#226)

  • service: handle multiple busy lookups (#217)

  • common: reduce complexity for the fan::Out abstraction

  • queue: forward stability during sqlite error (#222)

  • queue: group to handle > 1 message between persistent writes

[1.6.10] - 2023-06-09

Fixes

  • gateway: outbound does not wait for in-flight transactions

  • gateway: error-log if connection is lost unexpectedly (#216)

  • gateway: inbound does not handle prepare::Reply with read_only correct

  • service: avoid double busy lookup (#217)

[1.6.9] - 2023-06-08

Fixes

  • gateway: inbound must wait for in-flight transactions during shutdown

[1.6.8] - 2023-06-01

Fixes

  • discovery: “local” known services/queues are not added to discovery reply

  • discovery: replies should only contain a subset of the requested

  • service: include parent in service metrics

[1.6.7] - 2023-06-01

Fixes

  • discovery: aggregate external discovery requests for performance

  • domain: handle shutdown signals (#207)

[1.6.6] - 2023-05-29

Fixes

  • discovery: handle prospect services (#198)

  • domain: try to remove ipc device from exited ‘singletons’

  • gateway: handle routing problems (#197)

  • gateway: outbound order of connection for a service gets too “orderly”

  • gateway: outbound might core if connection lost during QM “restart”

[1.6.5] - 2023-05-19

Fixes

  • ipc: improve detection of removed ipc devices (files) (#194)

  • service: always send timeout duration in lookup reply

  • service: service timeout could render zombie transactions (#191)

  • gateway: always advertise discovered resources (#188) (#190)

  • gateway: shutdown doesn’t wait for inbounds before outbounds (#186)

[1.6.4] - 2023-04-25

Fixes

  • buffer: fix field doesn’t track ‘service buffer’ during auto realloc (#179)

  • queue: prevent forward causing service-manager coredump (#182) (ported from 1.5.15)

[1.6.3] - 2023-04-09

Fixes

  • sql: make sure sqlite uses synchronous=NORMAL (#177)

  • queue: fix memory leak by not (implicitly) using setenv repeatedly (#175)

[1.6.2] - 2023-03-25

Fixes

  • queue: extend c-api for get reply and get available

  • configuration: fix queue configuration default directory

  • queue: send error reply on failed enqueue

[1.6.1] - 2023-03-13

Fixes

  • queue: fix pending data structure slicing (memory leak)

  • documentation: fix sphinx-config

  • build: fix version.py to handle new minor

[1.6.0] - 2023-03-06

Added

  • cli: add force service/queue discovery to cli (#110)

  • configuration: add set operations union, intersection, difference

  • discovery: add build configuration option to set a service visibility (#86)

  • discovery: add domain configuration option to set a service visibility (#86)

  • discovery: add state and metric information to discovery (#110)

  • discovery: add directive to limit service discovery from other domains

  • domain: add start time for domain in information (#132)

  • domain: log name of exiting process (#105)

  • event: add event-logging for significant stuff that is done

  • event: add service sequential/concurrent order to event-service-log (#115)

  • general: all “managers” use multiplex send

  • http: outbound configuration to control caller fresh connect (#65)

  • http: make http payload base64 encoding optional

  • http: call::Context prepare for multiplexing and ‘http-forward’

  • http: use multiplexing in nginx inbound

  • queue: add browse-peek to queue api

  • queue: recover queue message (#74)

  • transaction: list external resources from cli

  • xatmi: extend c-api with execution service name (#140)

Fixes

  • build: package resources.yaml in /etc/casual/

  • cli: fix gateway –list-connections to indicate inbound discovery forward

  • cli: fix domain –configuration-get not outputing ‘system’ (#108)

  • cli: generalized –state option and the serialization to stdout

  • cli: fix stderr printout on error from casual call ...

  • cli: add legend for gateway –list-connections

  • cli: fix transaction cli output

  • cli: corrected legend for –list-services option to match header

  • common: add error logging on assert (#88)

  • configuration: improve configuration update semantics

  • configuration: default to one instance everywhere

  • discovery: don’t forward discovery if service/queue is known (#141)

  • domain: core servers must restart on unexpected exit

  • domain: keep service and queues sorted in discovery (#77)

  • event: fix event-service-log to also reopen ‘casual.log’ on SIGHUP

  • event: show trid in metric when trid is created by callee (#142)

  • event: make sure service-log listen to events before “connect”

  • environment: don’t create directories when constructing internal environment paths (#19)

  • forward: generate new execution id for each service call

  • forward: fix forwards dying when scaling aliases

  • gateway: reduce batch read/write in tcp handler

  • gateway: don’t let children inherit tcp file descriptors

  • gateway: fix tcp-logical-connect to detect loss of destination ipc

  • http: correct http inbound to handle keepalive

  • http: upgrade nginx to latest stable 1.22.0

  • http: remove debug log level as default (#79)

  • log: add api to configure log during runtime

  • service: fix pending “error” reply when server cores/exit

  • transaction: fix core when resources in distributed prepare (service) fails

  • queue: fix local lookup to also search external known queues

  • queue: log queuename when failing operation on queues (#106)

  • queue: fix cli to handle queue message with message attributes

  • queue: fix –list-remote

  • xatmi: tpsrvdone should never be called if tpsrvinit has not been called

[1.5.14] - 2022-09-08

Fixes

  • conversation: fix a bug where tprecv and tpsend failed to set the global tpurcode.

  • conversation: generate events in tpsend when partner calls tpreturn or tpdiscon.

  • conversation: fix handling of server process exit/core.

[1.5.13] - 2022-08-16

Fixes

  • queue: add lookup restrictions depending on caller location

  • service: add lookup restrictions depending on caller location

  • queue: fix pre/post statements for “tuning”

  • http: fix outbound to advertise services with order 0

  • discovery: fix using routes for services correctly

[1.5.12] - 2022-05-10

Fixes

  • cli: fix configuration –normalize

  • gateway: fix outbound to handle ‘gateway loop’

  • gateway: fix inbound to stop consume from a removed connection

  • service: use multiplex send as much as possible

  • service: fix lookup-forget-reply has replied for remote services

  • gateway: improve multiplex send, and consume ratio between ipc and tcp

  • discovery: reduce amount of topology related discoveries

  • discovery: use batch topology update, semantics for performance

  • tcp: guard errors in socket::address::host/peer

  • gateway: fix outbound to always remove transaction state

  • transaction: fix external resource prepare/commit request

  • gateway: handle lost connection in tcp send correctly

  • transaction: fix notification to TM for services with branch semantics

  • transaction: fix problem when to use TMJOIN in xa_start

  • transaction: be conformant with flags to xa_end

  • discovery: use multiplex send as much as possible

[1.5.11] - 2022-04-19

Fixes

  • cli: fix confusing arguments for casual service –legend

  • cli - remove extra root domain: casual configuration –normalize

  • gateway: make gateway more robust in error situations

  • xatmi: suspend involved resources during tpcall

  • gateway: absorb tcp errors and only effect the relevant tcp-connection

Changes

  • casual has moved to GitHub

[1.5.10] - 2022-03-21

Fixes

  • gateway: use multiplexing on the tcp-connection phase, no blocking if the tcp-connect takes a long time.

  • cli: fix casual domain --list-servers to show correct number of configured instances

  • transaction: fix local transaction with multiple resources

    • casual did not handle this correct (at all)

  • domain: fix shutdown if an instance of a server has died

[1.5.9] - 2022-02-18

Fixes

  • discovery: propagate topology updates upstream

    • When a domain logically connects (connector) to another domain (connectee), the connector propagates topology update to all “it’s inbounds” that has discovery:forward true configured. Hence, enable (auto) discovery for stuff arbitrary number of nodes upstream.

  • gateway: “move” listeners and connectors to failed on fatal errors.

    • Listeners: on error we move the listener directly to failed

    • Connectors: on fatal errors we move the connection to failed, on non fatal we try again after some time (as before).

  • cobol: set return status in Cobol api TPRETURN

  • gateway: keep trying to connect if the logical connection fails

[1.5.8] - 2022-01-08

Fixes

  • transaction: fix distributed transaction bug in complex topology

  • queue: forward react if queue group “dies”

  • xatmi: conversation disconnect not properly handled

  • queue: return no_queue iso no_message when writing to non existing queue

  • discovery - on discoverable::Available event: we only discover when it makes sense

Internal

  • log: made sure we only use our own stream::write and got rid of global ostream stream operator

  • common - flag abstraction degradation: equality operator for underlying enum type messed up semantics

[1.5.7] - 2021-11-26

Fixes

  • transaction: fix distributed commit when RM fails -> rollback

  • discovery: registration for discoverables more robust with request->reply

    • Primarily for unittests when booting, testing and shutdown of multiple domains within a couple of 100ms or so.

[1.5.6] - 2021-11-15

Changed

  • configuration: added system at the same level as domain

    • system holds system wide configuration, and for the time being only system.resource which replaces the resource.properties file (casual can still handle the deprecated file though).

  • cli: reverted the ‘locale’ awareness for the CLI

    • To many subtle and not to subtle problems - not worth it.

  • configuration: service restrictions for servers are now treated as regex

    • If set, only services that matches at least one regex are advertised.

[1.5.5] - 2021-11-09

Fixes

  • xatmi - conversation: major refactoring

  • xatmi: fix leak of descriptors for conversational services.

  • cli - configuration edit: made sure we sink the SIGCHILD signal from the spawned editor

  • documentation: fixed the example configuration files

  • queue: forward to remote queues did not work correctly

[1.5.4] - 2021-10-23

Fixes

  • transaction - if rm:s report error during xa_start: reply with TPESVCERR

  • configuration|cli - validate: only positive integers for instances

  • cli: version information, lowercase key names, compiler version

  • queue: support queue-peek in c-api

[1.5.3] - 2021-09-22

Fixes

  • http: outbound handle discovery (http-services are treated external as they should)

  • xatmi: support for conversations (primarily targeting COBOL)

  • gateway: fixed so outbound holds unique connections per service

  • discovery: service-manager only replies with local services

    • attention if the intention is that a domain should expose services that the domain has found in other domains (not local), the domains inbound connections should have the directive discovery.forward = true set. This also applies to http exposed services (since they’re not local)

[1.5.2] - 2021-09-17

Fixes

  • domain: fixed so order of groups (configuration) can be arbitrary, regardless of dependencies.

  • discovery: made sure we flush send to mitigate deadlocks when interacting with casual-domain-discovery

[1.5.1] - 2021-09-11

Fixes

  • gateway - respect hops: round-robin on the partition with smallest hops

  • gateway: fix interdomain protocol example tool

  • paths: made sure we do not use std::filesystem::equivalent with non existing file

  • cli: improved (subjectively) event::Notification print out (new for 1.5)

  • configuration: fix bug with empty aliases for gateway

  • configuration: fixed bug when ‘update’ an alias key ‘entity’. We moved from the one we supposed to add to…

[1.5.0] - 2021-09-01

Added

  • configuration: runtime configuration stage 1

    • cli for get, post, put, and edit which spawns an editor and when the configuration is saved casual tries to conform to the (possible) new configuration.

    • attention: under (heavy) load there might be some noise still, we aim to fix this in stage 2

  • cli: version information without a domain running

  • cli: added option to reopen casual.log. Prints information of all processes that casual are not sure can handle a SIGHUP, hence it easier for users to automate log-rotation.

  • cli: make casual cli user ‘locale’ aware, to play nicer with unix tools, such as sort

  • cli: added value auto to option --header. With auto, headers are only used if tty is bound to stdout

  • gateway: handle queue- and service information in connection state viewable in cli

  • xatmi - extended: function to browse instance services

  • cli: made sure wrong written cli command doesn’t gives error in log

  • queue - handle zombie queues: queues that are not configured anymore, but still exists in the queuebase

  • discovery: gateway inbound discovery forward

  • discovery: central ‘agent’ for handling discoveries

Changed

  • transcoding: platform-independent third-party-header-only-lib for base64-transcoding

  • paths: using std filesystem

    • attention: this might change semantics which temporary path is used.

    • https://en.cppreference.com/w/cpp/filesystem/temp_directory_path

    • On POSIX systems, the path may be the one specified in the environment variables TMPDIR, TMP, TEMP, TEMPDIR, and, if none of them are specified, the path “/tmp” is returned

  • python binding: upgraded to python3

Fixes

  • serialize - some simplifications: std::filesystem::path logs correctly

  • common: ‘helper’ for directory::create to handle corner cases with links

  • log: made the reopen on SIGHUP more robust

  • common: less semi confusing logging

  • xatmi: placed definition of functions where they should be to avoid confusion

  • cygwin compatibility

  • build: calculate version from tags -> more robust version sequences

Internal

  • configuration - better model: all managers supply their configuration

  • unittest: more deterministic unittests

  • common - algorithm clean up: moved some stuff to separate headers

  • build - improved and fixed some build related stuff: simplification

  • maintenance - simplify traits: some improvements in algorithm

  • signal: simplified signal to only have one handler for a given signal at a given time

  • build: fixed problem with linking on Ubuntu

  • build: turning archive to a shared library to avoid ODR

  • common - strong::Type: made a strong::Type that replaces “value::id” and “value::optional”

  • common: error handling and log fixes

  • build - casual-make: use new build system

[1.4.20] - 2021-06-15

Fixes

  • queue: improved performance while dequeuing messages

[1.4.18] - 2021-06-09

Fixes

  • gateway: tcp connect/accept errors might lead to inbound/outbound exit/restart

  • service: removed false error log when a timeout occur and the callee is killed

  • gateway: false error during tcp consume of logical (large) message

[1.4.16] - 2021-05-29

Fixes

  • gateway: improved tcp connection phase robustness -> spawns a process that take cares of the logical connection phase

[1.4.15] - 2021-04-14

Fixes

  • service: pending lookups did not work correctly for routes

  • gateway: made sure we terminate a detected ‘loop’ for calls

  • gateway: improved connect to be more robust

[1.4.14] - 2021-03-28

Fixes

  • gateway: outbound handle no-entry call reply correctly

  • gateway: outbound correlates multiple branched trids correctly

  • gateway: inbound respect service route name

[1.4.12] - 2021-03-05

Fixes

  • gateway - asynchronous writes: mitigate deadlock, performance improvements

  • gateway: handles bad tcp addresses in a more robust way

  • gateway: honour outbound group order for priority

  • configuration: gateway.connections -> outbound-group per connection

  • cli: call keep logical message to one correlation id

[1.4.9] - 2021-03-04

Fixes

  • gateway: outbound should prioritize replies from other domains

  • cli: call not use pending message and use the ‘flush send’ idiom.

[1.4.8] - 2021-03-03

Fixes

  • gateway: outbound/reverse-inbound might try to receive on wrong socket and get ‘stuck’

[1.4.7] - 2021-03-03

Fixes

  • domain: shutdown/restart makes ‘noise’ when provided services are under load

[1.4.6] - 2021-02-25

Fixes

  • discovery: did not work correctly on service with multiple routes

[1.4.5] - 2021-02-24

Changed

  • updated this changelog: missed it in 1.4.4

[1.4.4] - 2021-02-24

Fixes

  • discovery: did not work correctly on service routes

[1.4.3] - 2021-02-13

Changed

  • cli: updated documentation for –color with option auto

Fixed

  • cli: fixed so colors is false for Directive::plain()

[1.4.2] - 2021-02-13

Fixed

  • service: added timeout contract to cli output

  • queue: improved cli

  • service: assassination target caller instead of callee

[1.4.1] - 2021-02-07

Added

  • queue: concurrent queue-forward that is part of casual-queue

  • service: manager replies with requested queue on service discovery

  • domain runtime configuration core functionality

  • gateway: reverse connection for inbound/outbound

  • gateway: multiplexing (reverse) inbound/outbound

  • gateway: ‘soft’ disconnect from inbound

  • framework: serialize const lvalues

  • gateway: inbound/outbound groups to group connections

  • event: added event that signals when a discoverable is ‘on-line’

  • enable keepalive for tcp connections

  • added a auto to cli –color

  • making filesystem path serializable

  • transaction: resource-property-files are optional

  • possibility to terminate called servers when service timeout occurs

  • configuration: validate not empty values

Changed

  • upgrade to c++17

  • queue: reworked group and forward to be more rigid

  • service: make sure service-manager honour server-alias restrictions

  • remove 600 call trying limit in nginx

Fixed

  • gateway: did not add pending metric to request from service lookup

  • transaction: tx_begin did ‘start’ the transaction even if RM xa_start failed

  • defect in error logging

  • handle null value of queue message payload

  • queue-forward shall not handle new dequeue ‘flows’ in shutdown mode

[1.3.15] - 2020-12-24

Fixed

  • queue: missing error_condition equivalent implementation for queue::code all values for queue::code was logged as error

[1.3.13] - 2020-11-27

Fixed

  • queue: enqueue does not handle ‘shutdown’ correctly

[1.3.7] - 2020-10-14

Fixed

  • service: same instance could be added to a service multiple times

[1.3.6] - 2020-10-09

Fixed

  • gateway: rediscovery failed when there was unconnected outbounds

[1.3.5] - 2020-09-22

Fixed

  • transaction: did not remove ‘consumed’ transaction

[1.3.4] - 2020-09-17

Added

  • cli transaction support

  • handle glob pattern selecting configuration files

  • serialization: enable (de)serializing ‘unnamed’ objects

  • buffer: field-from-string more relaxed when dealing with string values

  • queue: improvement in performance

Changed

  • cli help improvements

Fixed

  • http outbound not marking transaction rollback-only on error reply

  • forward service causing error in metrics

  • handle null buffer over http

  • queue: dequeue with id

  • gateway: signal handling

  • queue: missing composite index on message.gtrid and message.state

  • queue: fix bad performance with enqueue when forwards are active

  • xatmi: did not call tpsrvdone when shutting down server

[1.2.0] - 2020-04-23

Added

  • gateway rediscover functionality

  • group restart in domain

Changed

  • improved ipc message pump semantics

  • cli improvement in aggregate information from managers

  • logging with iso-8601

  • queue: metrics and cli

  • reopen resource if failing

[1.1.0] - 2020-03-08

Added

  • http inbound and outbound handles xatmi codes in header

  • support for returning header parameters in http

  • build executables with resources

  • non blocking service lookup

  • casual queue c-api

  • support for queue retry and delay

Changed

  • logging reworked

  • improved metrics semantics

  • raw memory abstractions

  • improvement in cli

  • new implementation in http inbound

  • reworked user configuration

  • reworked marshal and serialize

  • better route semantics

  • administration api

Fixed

  • signal handling in non-blocking-send

  • multiplexing problem

  • protocol deduction error in http

  • http outbound blocking

  • base64 encoding/decoding in http

  • handling transaction branches

  • handling remote trid correct

  • cobol binding

  • service logging

[1.0.0] - 2018-08-13

Added

  • first released version 6,5 years after first commit.

  • all basic functionality supported

footnote

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.