openQA-4.5.1526033850.1ad9148d-bp150.3.3.1

- Fix missing share directory when worker is local to the webui (bnc#1111325)
  PATCH-FIX-OPENSUSE Add-tmp-dir-for-share-directory.patch bnc#1111325
  https://bugzilla.opensuse.org/show_bug.cgi?id=1111325

- Update to version 4.5.1526033850.1ad9148d:
  * Add _source field to every detail generated by parsers implementations (#1647)
  * Fix links to #live not starting live view (#1648)

- Update to version 4.5.1525548058.3f6bb61a:
  * Avoid appending multiple runs with --nocleanup (#1644)
  * Limit number of tasks for limit_assets and limit_results_and_logs GRU tasks
  * Add test for GRU enqueue limit option
  * Update apparmor profile for worker (#1645)
  * GRU: Add support to limit enqueuing by task
  * Update unit tests for TAP parser
  * Mark test suite as failed if test steps have failed
  * Add information about openQA internal results
  * Update TAP format documentation
  * Remove dots from filename
  * Ensure that the details description is set
  * Use a single result to contain all the test steps
  * Update changes to TAP format
  * Temporary version of the parser for TAP still failing
  * Log a warning when step data is not found
  * Allow LTP parser to load files that contain special characters
  * Replace extra dots in the filenames (#1642)
  * Increase the required coverage to 88.8
  * Do not fail a pull request if there are unexpected changes
  * Use same width for each colored dep status span
  * Show parent and child deps on job page in nice tbl
  * Enhance unit tests for GRU TTL
  * Set 2 days of ttl for limit_assets and limit_results_and_logs when posting isos
  * Add unit test for GRU task TTL
  * GRU: Add TTL for Minion jobs
  * Use a smaller font for the 'ago time' to fix rendering (#1635)
  * Prefetch comment counts for /tests
  * Update bootstrap to 4.1.1 and fontawesome to 5.0.10
  * Do not load full application to generate assetpack data (#1628)
  * Improve the duplication tests (#1627)
  * Hide categories when filtering test details
  * Ensure needles directory exists in 18-tests-details.t
  * Require 'Selenium::Remote::WDKeys' in CPAN file
  * Add test for filtering test details
  * Allow filtering test details

- Update to version 4.5.1524586233.92db89bc:
  * Add unzip-plain to AppArmor profile (#1629)
  * Use current SUSE Logo (#1626)
  * Avoid skipping dependencies when cloning jobs. (#1623)
  * Improve some details in t/ui/18-tests-details.t
  * Randomize GET /tests/99946#step/yast2_lan/1
  * Use usual prove flags on Travis as well
  * Hide failed modules on overview table for small screens
  * Fix positioning tooltips with Bootstrap 4 on overview page
  * Add Minion Link to navbar and move under admin
  * GRU: delete gru and deps on failure
  * Fix spacing of 'Build + Bars' rows on dashboard
  * Adjust padding and media break points for dashboard
  * Use exact same colors for cards/panels as before
  * Add unit test for job custom_module
  * Display GRU failure as job test module in openQA
  * Fix dropdown menus in fullscreen mode
  * Use full width in full screen mode
  * Make the gru service to be restarted on-failure
  * OpenQA::Task::* Die on fatal errors
  * Catch all tasks output also in case of error
  * Mark the jobs associated to the gru task incomplete in case of gru task failure
  * Adapt to new Minion::Job API introduced in Minion 9.0
  * Remove unused Data::Dumper import
  * Use Minion::Command::minion::job to get jobs list
  * Enable the Minion Admin interface under /minion
  * GRU/Minion: Take account of failures
  * Use gru enqueue in controllers and keep relation with minion jobs
  * Adapt test and fix leak of memory
  * Gru: get dsn from DBI for Minion
  * Move GRU tasks to Minion

- Require Minion >= 9.02

- Update to version 4.5.1524056562.74e334c1:
  * Close candidate menu after selection
  * Fix positioning of dropdown icon under Firefox
  * Fix handling multi-tag tests
  * Adjust UI tests to handle new candidates selection
  * Implement full diff
  * Allow to show just the screenshot again
  * Add shortcuts to custom combo box
  * Make style for result preview part of openqa.scss
  * Use customized bootstrap dropdown instead of combo box
  * Workaround issue when clicking menu item in 21-admin-needles.t
  * Port card in overview correctly to Bootstrap 4
  * Use same link color as before
  * Make badges look like before
  * Adjust colors
  * Fix paging of tables for running and scheduled jobs
  * Remove symlinks for Glyphicon fonts
  * Fix UI and fullstack test after Bootstrap 4 migration
  * Adjust media breakpoints
  * Fix grammar in tour
  * Upgrade tour to Bootstrap 4
  * Restructure stylesheets
  * Fix live-view and other small issues under Bootstrap 4
  * Make it look more like before under Bootstrap 4
  * Fix permanentely moved links on dashboard
  * Fix overall appearance with Boostrap 4
  * Fix filter panel on the main page
  * Patch bootstrap theme to be a little more like it was before
  * Remove hidden class
  * Port the navbar to bootstrap 4
  * Start to migrate to bootstrap 4

- Update to version 4.5.1523608608.fca0aa46:
  * Enforce sorting the results returned by the sql query to use in is_deeply (#1621)
  * Remove dead code in Worker/Jobs.pm
  * Update current test module details when cancelling running job
  * Prevent failures in worker reset_state() (#1618)
  * Group candidate needles by tag (#1606)
  * Test image selection of needle editor (#1617)
  * Fix 'Take image from' selection in needle editor (#1616)
  * Refactor needle editor code
  * Fix showing needle editor when no tags present
  * isotovideo: wrap log_ calls with eval
  * Require latest RWP
  * Start isotovideo inside a container using cgroups
  * Checks if failure status file exists before trying to slurp it
  * tests: cover case when one of the cluster jobs already running
  * Consider running jobs in filter_jobs
  * Fix fullstack test in case /usr/bin/isotovideo is present
  * Limit showing new needles to the 5 most recent ones
  * Use 'new' prefix when showing new needles
  * Add documentation about showing new needles in editor
  * Improve tests for needle editor
  * Show errors/warnings/info in needle editor with consistent margin
  * Show added/updated needles in editor
  * Add database entry for needles when saving via editor
  * Simplify save_needle_ajax code
  * Extend needle table with tags and timestamps
  * Capture serial patterns documentation
  * Don't show configure button for untracked assets
  * Fix mojo not handling max_size_limit when is a transaction
  * Allow optional overwrite of the productlog request depth with a query parameter (#1601)
  * Enable JSON representation of the test overview (#1600)
  * Add asset size limit option
  * Return reference for test result file lists
  * Make OpenQA::Client::Archive property of OpenQA::Client
  * Remove chdir that caused errors when downloading test files
  * Add file _download_file_at method for logs and uploaded logs
  * Move progress monitor code to a separate fuction
  * Ensure that uploaded test logs can be downloaded
  * Add first version of OpenQA::Client::Archive
  * Move test_uploadlog_list and test_resultfile_list to the model

- Use Minion for GRU tasks, this requires Minion and Mojo::Pg as dependency

- Fix handling UTF-8 when executing make install

- Fix handling UTF-8 in update-cache.sh

- Update to version 4.5.1520515419.c3df6f91:
  * Update the chromedriver url from 2.33 to 2.36
  * Allow saving needles for OS with dots in the version
  * docs: Clarify some serial terminal terminology and other elaboration (#1596)
  * Do not reset the api version just because the worker is offline
  * Fix scheduler comment - now default MAX_JOB_ALLOCATION is 80 (#1595)
  * Always retry until chunk is uploaded and retrials exhausted (#1594)
  * Refactor and add error test
  * PARALLEL_CLUSTER is no more
  * Adapt scheduler_full test
  * Avoid to re-consider the jobs that are being allocating in a round
  * Do not take account of spaces in PARALLEL_WITH
  * Exclude allocating jobs from prefer_parallel cut
  * Consider allocating only when we have them
  * Bump MAX_JOB_ALLOCATION default to 80
  * Exclude allocated when blocking dependencies
  * Consider the jobs allocated during the round as running in prefer_parallel
  * Prevent cluster tests to be allocated if their group is not allocated

- Update to version 4.5.1520100590.2279151e:
  * Fix sort order of the review comments (#1593)
  * Expand the serial terminal console documentation (#1586)
  * Emit openqa_job_create events when an ISO is posted (#1584)
  * Replace die for 500 code and move global constants to OpenQA::Constants
  * Avoid circular dependencies when running the tests.
  * All the available free workers need to have the same websocket interface version
  * Move the interface verification to the worker ack

- Update to version 4.5.1519929327.f726adcc:
  * Enable wrapping cells in overview table (#1591)
  * Perltidy
  * Protect during collected unsubscription
  * Prevent to determine job status from collected event when timout expires
  * Remove check backend timer
  * Fix check_backend detection
  * cpanfile: require at least Mojo::IOLoop::ReadWriteProcess 0.19
  * Adapt jobs tests - we are not dying anymore
  * Don't fear the (sub)reaper
  * Add AMQP doc
  * Lock tidy version
  * Make prove verbose

- Update to version 4.5.1519128406.810fb68b:
  * Show correct port for created job (#1578)
  * Use rwp also to call rsync (#1585)
  * Force a restart when the children or grand-children processes die
  * OpenQA::Parser::Format::JUnit mark testsuite failed if one of testcases fails
  * Replace the exec call with ReadWriteProcess module
  * Add use Mojo::DOM in JUnit and XUnit Parser (#1580)
  * Use the generated Mojo::URL as Client base_url
  * Add get_piece() in OpenQA::File to ease writing of downloader class
  * Add test for OpenQA::Files's prepare()
  * Make integration tests of uploader with fixed chunk size
  * Fire events inside subprocess before and after chunk processing.
  * Move uri checks inside common function in OpenQA::Client::Handler and guard from not configured scheme in host
  * Add test to check if retrials numbers are matching
  * Use subprocess for potential loop blocks
  * Adapt workers to error handling
  * Add tests for OpenQA::Client::Upload retrials on failures, and events propagation
  * Emit errors on OpenQA::Client::Upload
  * Swap multi-chunk uploader in worker with OpenQA::Client
  * Make OpenQA::Handlers EventEmitter(s)
  * Split OpenQA::Client
  * Wipe out content of written chunk, try only weak check and back to Storable
  * Do not always perform total cksum
  * Drop Mojo force to File asset, display stats of uploaded chunks
  * Write as we receive in a temp file
  * Avoid to relay too much on the serializer
  * Use C backend to read whole file when computing total cksum
  * Enhance error detection
  * Allow to chunks to dynamically grow and shrink
  * Make total checksum verification run faster
  * Let specify chunk size in config file, and default to 1000000 bytes
  * Also add API endpoint to inform of upload status, adapt tests
  * Do not make split() to load the file
  * Pass asset option in multichunk upload
  * Create empty file before writing it if doesn't exist
  * Try to swap current uploader with multi-chunked one for assets
  * Add spurt to OpenQA::Files
  * Add write_verify_chunks() and write_chunks to OpenQA::Files
  * Add methods to verify against chunks, also add test when joining/verifying in-place
  * Add serialize/deserialize also to single result of parser
  * Add OpenQA::File, wraps Mojo::File and adds seek capabilities
  * Add worker version check on the server. (#1552)
  * Add basic test for OpenQA::WebAPI::Description
  * Add POD documentation to Controller/API/V1 packages
  * Test 27-errorpages.t with a predefined API desc
  * Replace Pod::Tree for Pod::POM in WebAPI::Description
  * Add POD documentation to Controller/API/V1 packages
  * Move API description methods to its own package
  * Add POD documentation to Controller/API/V1 packages
  * WebAPI: Add API description from Controller's POD

- Update to version 4.5.1518003872.3977d2f5:
  * Run perltidy again (#1577)
  * Fix perltidy complaints (#1576)
  * Give 404 rather than 500 on invalid asset route (#1575)

- Update to version 4.5.1517920216.2250682a:
  * Take latest version of Selenium::Remote::Driver (#1570)
  * Big rework of the GRU limit_asset task (#1569)
  * Fix broken links in openQA documents
  * Update Installing.asciidoc
  * Add pager for comments on group overview (#1561)
  * Add help box for group filter
  * Fix testsuite descriptions if default config not present (#1554)
  * Test subgroup filter and form
  * Allow to expand nested groups by default
  * Add filter for subgroups
  * Check qemu pid before starting job (#1565)
  * Require < 1.21 of Selenium::Remote::Driver
  * Downgrading selenium drivers during test phase

- add fonts for chromium test suite - and remove currently broken
  tests (https://progress.opensuse.org/issues/19652)

- Update to version 4.5.1516716144.98db9c99:
  * Improve error handling in group overview (#1562)
  * Do not remove forwarded-for=https headers (#1558)

- Update to version 4.5.1515767133.56cd6ee5:
  * Support nested OpenQA::Parser objects
  * If no parser format is supplied, return a OpenQA::Parser::Format::Base object
  * Add POD documentation for OpenQA::Parser::Result::OpenQA
  * Add documentation for OpenQA::Parser::Results and move search() in proper class
  * Move search_in_details() in a new class that inherits from OpenQA::Parser::Results as it is openQA specific
  * Add documentation of OpenQA::Parser::Result
  * Move methods and attributes from Parser to Base format
  * Simplify spec format definition and move common displayed fields to openqa specific result classes
  * Drop support for creating collections with package global variable (useless now)
  * Make possible to upload in all the supported format and add tests for it
  * Add other tests, bring back coverage and fix wrong Node val() assigment
  * Return a OpenQA::Parser::Result::Node object when getting unstructured data from result
  * Make parser now really recurive, add also support for nested data structure with unit test
  * Make LTP parser tests output compliant to openQA tests results
  * Run parser test in parallel
  * Add possibility to store / serialize file content too within the parsed result
  * Add test to check DB consistency with parser extracted results
  * Move format specifications to separated namespace
  * Add XUnit tests
  * Add basic XUnit parser and schema
  * Clean LTP and include schema inside parser file
  * Add support for nested Results collections (required for XUnit)
  * Use functional interface in exposed api
  * Make collections serializable for unstructured data
  * Add accessor for unstructured data
  * Ask to include result only if the specific parser can
  * Move include_result option to specific parser that needs it
  * Provide also functional interface
  * Serialize single Results object as seen by to_hash() and add tests
  * Add tests when parser errors
  * Support hashifications of unknown objects
  * Add unstructured data test
  * Add reset method to parser, add and cleanup tests
  * Make serializing possible for collections with different OpenQA::Parser::Results objects
  * Add possibility to export/import the Parser to JSON, add also tests
  * Add serialization tests for junit spec
  * Tidy OpenQA::Parser::Result
  * Hide name field in JSON encoding of OpenQA result - but keep serializing it
  * Add example of new format specification for LTP
  * Simplify LTP spec definition even more
  * Get hash representation automatically of attributes if they can be rendered as hash
  * Add backword support for v2 format and add unit tests
  * Rename subtest to avoid Perl 5.18 complaints
  * Refactor write of parser result data
  * Add optional extra list that can be filled by type spec parser
  * LTP specification: Environment is optional inside result in new format
  * Add support to serialize/deserialized nested data structure
  * Add serialization capabilities to parser, and add tests
  * As classes are representing merely what's the json content, include them in the LTP module file
  * Add test data example for LTP results format
  * Add support for LTP specification in OpenQA::Parser
  * Generalize parser results, and subclass to specific implementations
  * Adapt parser tests, since now default value for script is 'unk'
  * Generate test module result from parsable extra_tests
  * Parser: return unk for test script if does not exist
  * Add search between parsed tests, link results and tests
  * Draft for JUnit parser
  * Prevent worker status selection to sort (#1555)
  * Unify code for scheduled/running table and fix sorting (#1556)
  * Remove leftovers in t/api/02-assets.t (#1557)

- Update to version 4.5.1515509269.fa07a084:
  * Remove assets from disk when deleting database entry (#1553)

- Update to version 4.5.1513750232.be133588:
  * Add basic setup of multi-machine worker (#1544)
  * Allow deleting assets via web UI (#1551)
  * Deal properly with environment variables that can change logging (#1537)
  * Fix deletion of jobs referenced as last use of asset (#1549)
  * Show assets by job group
  * selenium test: Allow configuring mojo port
  * Split asset tables in used and untracked assets
  * Refactor data size rendering in JS and reduce precision
  * Set App mode in WebSocket server (#1548)
  * Fix repos gone too early: assign REPO_10 to jobs as well (#1547)
  * Retidy - Perl::Tidy 20171214 has different formatting (#1546)
  * Update installation steps for postgresql (#1538)
  * branding: openSUSE: Add Kubic (#1539)
  * Change the logs timestamp to ISO8601 (#1542)
  * Add documentation for postgresql migration (#1530)
  * Fully fix asset caching without test caching (POO #29208)
  * Fix job group update issue (#1533)
  * Hide worker properties if no login (#1534)
  * Switch to fontawesome 5 (#1531)
  * `limit_assets`: remove asset that breaks limit, not next one (#1536)
  * Remove all assets from DB that are gone (#1535)
  * Allow to query live data from api, adapt tests
  * Display additional information in the worker's table
  * Add panel and popover to show job detail
  * Make searchable field and fix DataTable loading
  * Add openqa-setup-db service that will create postgresql database (#1528)
  * Retrieve worker stats data
  * Do not query by default websocket server to display worker informations
  * Fix tests to expect new columns
  * Show size of assets exclusively kept by a group
  * Accumulate size of assets exclusively kept by a group
  * Document fast initialization of local tests (#1527)
  * Add column to groups with size of exclusively used assets
  * Also store if the asset is fixed in the DB
  * limit_assets: Use DBD::Pg directly for querying job asset
  * Check all columns in test for asset table
  * Show last use of asset in assets table
  * Ignore chromedriver logs
  * Show asset size
  * Write info about assets kept during cleanup to db
  * Improve comments in limit_assets
  * Fix PRJDIR error
  * Add support for colons on worker class
  * Add WORKER_CLASS to job_list filter
  * Fix caching of assets without caching of tests (#1521)
  * Accept a lower coverage - we removed code
  * Move worker log test to 24-worker.t
  * Schema.pm: Remove unused functions (increase coverage)
  * Remove SQL files for SQlite and for 2016
  * Have our services require postgresql server to be up
  * Adopt generate-packaged-asset to use postgresql
  * Import log_error in OpenQA::IPC
  * Test: increment unstable_worker ticks and wait for each one to be ready
  * Remove more traces of Sqlite support
  * Provide t/test_postgresql to setup a temporary postgresql
  * Fix more test cases
  * Prepare a proper DB for the scheduler full test
  * Fix t/ui/22-job_group_order.t by providing sort_order in fixtures
  * t/ui/13-admin.t: Postgresql sorts RAID at the end
  * Sort table settings by key to have reliable tests
  * t/api/02-iso.t: Do not rely on hardcoded job number
  * Give WebSockets server its assets
  * Sort the previous jobs in worker table by ids
  * Skip all tests requiring a database
  * Test fixtures: restart the postgresql sequence
  * Fix t/17-build_tagging.t not to depend on Sqlite specifics
  * ui/04-api_keys.t: Don't rely on ids in fixtures
  * 14-grutasks: don't rely on DB sort order
  * Drop sqlite3 support
  * branding: Update SLE staging links (#1523)
  * Use log_info for worker activities (#1519)
  * branding: openSUSE: Add caasp (different bugzilla naming variant) (#1522)
  * Tidy the sources
  * Do not cast to int worker_status timer
  * Propagate error on IPC dispatch call
  * Use WORKERS_CHECKER_THRESHOLD to detect dead workers as well
  * Remove printing to debug the result of safe_call
  * Reduce OpenQA tests verbosity
  * limit_assets: don't wipe assets for PENDING jobs (#1518)
  * Allow same group name within different parent groups (#1507)
  * Apply log level correctly (#1516)
  * Check if job is valid before starting working on it.
  * Upload workers log to the webUI (#1502)
  * Only test basic tests on perl 5.26 (#1512)
  * Remove IDs from fixtures if they are in sequence (#1513)
  * Move the $shared_dir to setup_share_dir
  * Print project directory upon worker's init failure
  * Fix documentation generation script
  * Avoid warnings during test runs (#1511)