jgonera / openopj Goto Github PK

It seems that we understand the format wrong. Here are some ideas (unproven and untested, just some ideas for now).

There is a design pattern when designing a file format. Called TLV, tag-length-value.

1. A file may be it is just a sequence of blocks. But special blocks. tag \n value \n tag is a string from predefined list of strings. if it is not, if it is u4 (uint32_t), then it is size.
2. if it is size, and if we are not in a section, then tag is in the first 2 bytes of header, but it is binary one. I feel like 0xfffc stands for a file header, and 0x0001 stands for column header. then header means entering a section.
3. when in a section subsequent blocks types are determined from the algorithm of parsing of that section. I mean that algo consumes blocks and probably section header contains count of blocks to consume. Or maybe consume untill null block. IDK for now.

Again, these all are untested, just sharing some thoughts because I have not so much time to dive into it now.