I have a nice little class that loads in all .pngs files named 'arbitrary_name_<n>.png' in my art directory (where <n> is a value).

It stores them in a dictionary so that I can do things like this:


All is well and good... ish.

But this dictionary lookup seems a bit... well... high level and expensive. I'm loathe to replicate data if I can avoid it, so I don't want to hand-create a bunch of enums which map one-to-one against my files.

I've been considering using my graphical processing script (in python) to spit out a nice enums.h file in C format.

Anyone else doing anything cleverer?

Dictionary lookups ought to be pretty quick. Have you profiled with Shark and had it show you that this is a hot spot in your code?

ThemsAllTook Wrote:Dictionary lookups ought to be pretty quick. Have you profiled with Shark and had it show you that this is a hot spot in your code?

Not sharked this yet; the root of all this is something I chose not to explain as I thought it might make my initial question far too confusing. It seems like getting a text file to result in code-readable information would be quite useful. Perhaps I'm getting ahead of myself and optimising too early... it's been known to happen.

Madrayken Wrote:Not sharked this yet; the root of all this is something I chose not to explain as I thought it might make my initial question far too confusing. It seems like getting a text file to result in code-readable information would be quite useful. Perhaps I'm getting ahead of myself and optimising too early... it's been known to happen.

Definitely profile it before you worry about how fast a dictionary lookup is - you may be surprised. And don't worry about optimizing at all until all major features are in place - there is no point at all in hand-tightening code that you may throw away in a month or a week. That's time wasted!