[WIP] 1691 dg calculus on database (#1752) * Add database dependencies. * first experimental version of DB interface - just store DG names * also write dg nodes to db * hints for further programming * Connect to PostgreSQL databases using a Rails database.yml. * Move the schema to an own module. * mapping of Hets datatypes to DB * more explanation on index maps * Save the LogicGraph and Development Graph in the database. * Add config option for DB file. * Move dbConfig into HetcatsOpts, move migrations into onDatabase. * Save errors to the database. * Associate the data with a given file version. * Fix saving independent documents, non-signature symbols and remove debug prints. * Remove theoryMap from DBCache. * Fix names of Loc/Ids. * Drop unneeded database columns. * Save SignatureMorphisms and SymbolMappings. Major refactoring in Persistence.DevGraph. * Handle Library and NativeDocument properly. * Rename Range to FileRange because of constraints in Ontohub. * Use PostgreSQL Enums. * Rename Warning to Warn because of constraints in Ontohub. * Use proper location as Loc/Id. * Include persistent-fork into Hets repository as a git submodule. * Find the FileVersion of an import and save the Document accordingly. * Add advisoryLocked function. * Ignore the _dependencies directory. * Only save a Document if it does not yet exist. * Lock language migration and document saving. * Use mutex, write logic graph in ins own command Write logic graph: Since both commands, DG analysis and LG output need to write Logics to the database, we need the same mutex kex for these two tasks. This splits the LG output into its own command to allow for multiple processes to write analysis data at once. * Get the submodule via https. * Add readme for the persistence layer. * Fix findFileVersion function. * Fix export warning. * Find a FileVersion at the current commit instead of the previous version. * Make the symbolKind part of a symbol's Loc/Id. * Make concatMapM more generic. * Correctly associate symbols along heterogeneous links. * Set evaluationState of FileVersion. * Allow to omit the --database-fileversion-id option. * Simplify the travis configuration. * Add database job to travis. * Find FileVersion of the last change. * Fix dependency relation order. * Add MySQL support. * Add restack target to Makefile. * Add option to disallow migrations That means, PostgreSQL can be used without Ontohub. * Remove unused import. * Update Stack release and use persistent-template of the fork. * Make the existing columns in the schema MySQL compatible. * Work-around migration errors caused by class table inheritance. * Use PostgreSQL and MySQL in travis. * Use better job labels in Travis. * Possibly fix travis. * Add SQLite to Travis. * Add installation instructions. * Add DGVerificationGeneric constructor. * Rename ConsStatus to ConservativityStatus. * Add error handling. * Fix symbol's fullName. * Set Logic name to Language Name when there is no Logic name. * Do not print migrations. * Fix key length for MySQL. * Add namespace to Loc/Ids. * Add option for re-analysis. * Add a testfile. * Improve error messages. * Create the freeness parameter OMS. * Add failing testfiles. * Remove code duplication. * Temporarily remove failing test files - until #1655 and #1597 have been solved. * Try longer locId in MySQL. * Fix mapping type. * Add default implementation of symbolKind. * Add Indexes. * Quiet down PostgreSQL on CREATE INDEX IF NOT EXISTS. * Add PostgreSQL 9.6 to addons. * Allow Hets-database to run without ontohub. * Reduce column size for MySQL. * Move the database-specific dependencies into their own module. * Upgrade persistent (rebase on master). * Minimise the module exports on Database. * Upgrade stack.yml to new stack version. * Fix SQLite module. * WIP: Try System.Plugins. * Remove System.Plugins again. * Deactivate MySQL. * Remove System.Plugins from the dependencies. * Finalise db_mapping.txt * Update readme. * Don't use PostgreSQL for hets-desktop. * Use hets_server in Travis/PostgreSQL job. * Add PostgreSQL as a dependency to the server package. * Update some documentation. * Fix runtime dependencies. * Describe main db options in the UserGuide. * Add libpq5 to debian dependencies. * Clean the database.yml files and link them from the UserGuide.

Fix _JAVA_OPTIONS environment variable on Travis.

Split travis build into 'Hets-lib' and 'make check' jobs. (#1741)

1722 integrate more tests (#1730) * Integrate Hets-lib test script. * Put the script and file list into the Hets repository. * Pull Hets-lib if it already exists. * Install hets-desktop.

Create a loose stack and clean_stack target to setup stack.

Travis: Use addons key to install packages.

Fix some of the issues stated in the review.

Add another cache directory to travis.

Use stack in travis-ci.

travis has only jdk7 (is always trusty)

travis: install hard dependencies, only (also saves ~1 min)

add missing texlive-fonts-recommended (ecrm1000), travis: use native pkgs

make travis CI happy; add a "test" note to the README

removed parsec1

removed parsec1

gtk2hs needs newer alex

try a standard installation

try cabal-install-1.22

remove verbosity

try in one go (and use parsec1)

install glade separately

output dependencies

newer versions

forget programatica

add /home/maeder/.cabal/bin to PATH

added happy for gtk2hs-buildtools

added alex for gtk2hs-buildtools

try to install a newer ghc

try to set HETS_MAGIC properly

set HETS_MAGIC

added 'make check'

* Compile haskell sources in batches of 200 to (hopefully) avoid #1276 * Speed up compilation by keeping compiled *.o files This should provide a signifcant speedup as most pull requests change few files

download programatica sources

need to build derived sources

install gtk2hs-buildtools, updated programatica link, hide programatica

added travis configuration file