* Move importer to an inline-template, allows for translations and easier passing of data from laravel to vue.
* Pull the modal out into a dedicated partial, move importer to views/importer.
* Add document of CSV->importer mappings. Reorganize some code.
Progress.
* Add header_row and first_row to imports table, and process upon uploading a file
* Use an expandable table row instead of a modal for import processing. This should allow for field mapping interaction easier.
* Fix import processing after moving method.
* Frontend importer mapping improvements.
Invert display so we show found columns and allow users to select an
importer field to map to. Also implement sample data based on first row
of csv.
* Update select2. Maintain selected items properly.
* Backend support for importing. Only works on the web importer currently. Definitely needs testing and polish.
* We no longer use vue-modal plugin.
* Add a column to track field mappings to the imports table.
* Cleanup/rename methods+refactor
* Save field mappings and import type when attempting an import, and repopulate these values when returning to the page.
* Update debugbar to fix a bug in the debugbar code.
* Fix asset tag detection.
Also rename findMatch to be a bit clearer as to what it does.
Remove logging to file of imports for http imports because
it eats an incredible amouint of memory.
This commit also moves imports out of the hardware namespace and into
their own webcontroller and route prefix, remove dead code from
AssetController as a result.
* Dynamically limit options for select2 based on import type selected, and group them by item type.
* Add user importer.
Still need to implement emailing of passwords to new users, and probably
test a bit more.
This also bumps the memory limit for web imports up as well, I need to
profile memory usage here before too long.
* Query the db to find user matches rather than search the array. Performance is much much better.
* Speed/memory improvements in importers.
Move to querying the db rather than maintaining an array for all
importers. Also only store the id of items when we import, rather than
the full model. It saves a decent amount of memory.
* Remove grouping of items in select2
With the values being set dynamically, the grouping is redundant. It
also caused a regression with automatically guessing/matching field
names. This is starting to get close.
* Remove debug line on every create.
* Switch migration to be text field instead of json field for compatibility with older mysql/mariadb
* Fix asset import regression matching email address.
* Rearrange travis order in attempt to fix null settings.
* Use auth::id instead of fetching it off the user. Fixes a null object reference during seeding.
* Rearrange travis order in attempt to fix null settings.
* Use auth::id instead of fetching it off the user. Fixes a null object reference during seeding.
* Improvemenets to unit tests.
* Break up modelfactory into multiple files, populate many states.
* Begin testing validation at the unit test level, test relationships.
* Add tests for Asset::availableForCheckout.
* Model factories now generate all needed relationships on demand,
which allows us to unit test with a empty database.
* To faciliate the empty database, we move to using sqlite in memory as
the unit testing database.
* Fix bug with logs of checkouts to non users.
* Fix location finding for assets. Also Fix location show page to show users associated with location. Still need some work to show assets.
* More test and generator improvements
* More unit test fixes. PermissionsTest is borked still.
* More Updates
* Rewrite permissionstest. Check that we have access on the model level rather than via web requests. Also test delete permissions.
* Fix seeders.
* Make the default asset model factory generate assets that are rtd for testing.
* Save progress.
* Rebase tests, fix department unit test, update database for functional tests.
* Update functional and api tests to use new modelfactory signatures.
This is round one of the rethink of checkout-to-everything. A location
now has a manager field, and the manager (by default) be responsible for
assets checked out to the location.
We currently depend on a "vue-strap" module to provide the modal dialog.
The default npm vue-strap is not compatible with vue2, so we need to use
a fork. Update package.json to reflect this. Probably makes sense to
move away from vue-strap in all my copious amounts of free time.
