-
Notifications
You must be signed in to change notification settings - Fork 19
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
XML Marker Parser and New Internal Format #56
base: master
Are you sure you want to change the base?
Conversation
what binary format btw? custom or a standard one like protobuf? |
I currently plan to use protobuf, specifically proto3, as it is a strictly typed and future proof standard that, in theory, could easily be used by other applications if desired. I do want to add the functionality to export these binary files back into XML for distribution to existing XML parsing applications. Several considerations have yet to be made about the schema. Some of note are: if images should be stored inside the tree or outside, if trails should be stored inside the tree or outside, how this might handle future yet-ephemeral features such as 2D areas or 3D meshes. |
Remaining tasks can be found in #64 |
I personally grouped up all the bools into a single byte.
then again, i had not saying you gotta do this too, as the code already seems complex enough with macros and generators. but might still be relevant to this because using bitflags makes a lot of sense for binary formats. |
@coderedart Using bitflags makes a lot of sense but I am going to leave these as bools both in memory and in the serialized format. Protobuf already has other aspects that make it less-than optimal for space saving. If I wanted a perfect binary format on disk I would serialize a struct with bitflags. However the convenience of automatically having the variables defined and usable by name inside an instantiated class when parsing a protobuf, and the utility of forwards and backwards compatibility, which would be lost if using bitflags, outweighed the space savings for me. For the filters/flags, like profession, protobuf wont save For the in-memory data structure, saving an extra few dozen bytes per icon is still only a few kB of memory all in all, and once I get over to GDScript it is going to be stored in memory in an even less optimal format. However if the need arises to restrict memory usage in this c++ module then I will absolutely take a look at switching to bitflags here. And just theory crafting as to how they would work for fun. Each class here, such as |
Also just because I feel like ranting about macros. The crazy complex macro system I had created originally, before deciding to swap to code generators that work off the documentation, was a bad idea. It seemed really cool, but in the end due to the things I had to do to get everything to work properly without adding extra toil to modifying a class caused the code to end up being quite slow for a dumb reason. Instantiation of each |
Reorganizing Tests
Adding a presubmit for integration tests and fixing the generators one
Renaming intigration_tests to integration_tests
Changed render_ingame to is_hidden_ingame
Changed render_on_map to is_hidden_on_map
Changed render_on_minimap to is_hidden_on_minimap
Xml converter
Fixed a bug that caused trails to not inherit attributes correctly
Changed the name waypoint to guildpoint
…ng if there are any duplicates between packs. Added an argument to bypass this if nessecary. Addded tests for both cases
Merging master changes into XML_Converter
…mbining set, error handling for argument parsing
Changed workflow to also work on the new guildpoint extention
Fix automatic builds
… conflicts are communicated
…file by suffix function to return relative paths
Changed all instances of .bin to .guildpoint
Added tracking for the top level categories in each pack.
Changed file extenstion for duplicates tests
A new XML parser using the RapidXML library which converts to a new internal binary marker format that will be read by Burrito. Before this PR is merged the new format will be implemented and usable for all of the attributes that are currently supported. The new internal format will support marker categories, and will be divided by map and marker pack. Having this done will allow for Burrito to eventually support enabling and disabling specific marker categories on a given map instead of entire marker files at once. The completion of this work will also indicate the depreciation of the existing python and rust XML parsers.
The XML parser will attempt to parse all valid attributes across BlishHUD and TacO even if those attributes are not used within Burrito at this time. The result is that the parser will double as a linter to alert users of any marker errors. Depending on time a simple documentation generator may be able to take the various parser classes and create documentation similar to BlishHUD's docs automatically.
Right now all that is completed is a design for writing the XML Parser / Linter