I'm currently trying to workaround a problem of ST2/3 with huuuuuge text files. And I'm not talking about a 10000 lines of code file, I'm talking about hundreds of megabyte of ascii, without any linebreaks. Such files a pretty usual on mainframes and I currently need to work with them.
Some are structured (EDIFACT for example) some are record based (fixed intervals of fixed length text).
Currently I preprocess these files for viewing externally (i.e. split EDIFACT by specific patterns and split the record based files into separate lines according to the record length).
Now it would be awesome if I could do that as a ST plugin so it happens magically. I imagine something like when opening a file beyond a certain size (>= 10 MB maybe) a popup asks me how to proceed and if split is intended, how to do it (by length or by pattern). Popup and splitting aren't the problem, however I can't find a way to intercept the file loading _before_ ST loads it completely - because that takes ages. So I have to "fix"/preprocess the buffer before ST tries to do anything with it (parsing, or whatever else it does while loading a file).
BTW. just as a side note: vim, scite, notepad++, gedit, kate, ... all struggle with these kind of files. The only editor I have so far that loads them instantly is UltraEdit. So I don't blame ST for not handling them well, I just want to fix it