This commit is a big refactor that improves the code organization.
It focuses on how dependencies are injected into the different classes. This is now done via some factories.
Also, the code is now built around entities (data model that only exposes getters and setters), managers (used to manage an entity) and helpers (used by the managers to perform some specific tasks).
The refactoring is not fully complete, as some dependencies are still hidden...
This will improve the management of options and simplify some methods' signatures.
This commit will also help moving the code to a better state regarding Dependency Injection.
If an exception is thrown while writing data, instead of letting the developer handle this situation gracefully, Spout can attempt to delete all the temporary files that were created so far, as well as the output file as it won't be completed and therefore corrupted.
Fixes#295
If a row should be written with a custom style, the handling of empty cells should change.
Instead of being skipped entirely, empty cells will be applied the custom style, if this style has custom background color or borders.
If not, then the cell definition can still be skipped.
* Fix#276, some general refinement
* Failing test #267
* Fixed shared border definitions across different styles #267
* Fix finding the correct borderId
This fixes issues when something went wrong on reader/writer init and the developer wants to close the reader/writer.
The file handle may not be defined so we need to add a check for it, before actually using it.
Some software generate [Content_Types].xml file with sheets definition in random order.
Instead of having the first sheet (id = 1) defined first, it may be defined in 3rd position.
Therefore, to read the file in the correct order, sheets order need to be fixed.
The ZipHelper interface is now more generic and allow single files to be added.
It supports adding uncompressed files (for PHP7+), which is required to have the mime detection magic work with ODS files.
Also fixed a few issues with the created ODS file (thanks to https://odf-validator.rhcloud.com/)
Heuristics to detect proper mime type for XLSX files expect to see
certain files at the beginning of the XLSX archive. The order in which
the XML files are added therefore matters.
Specifically, "[Content_Types].xml" should be added first, followed by the
files located in the "xl" folder (at least 1 file).
Since cells are referenced by their coordinates (A2, B4...), it is not necessary to write empty cells.
This will reduce the final size of the generated XML and therefore XLSX file.
