This is actually a great question. The order depends based on the specific application.
[*]
Subtractive EQ before Comp:[newline ]
Subtractive EQ can be thought of as a way to soften your signal, then you would compress your tamed signal. Here, it's ok for the compressor to come after as it's only working on the frequencies that you're most concerned with.
[*]
Subtractive EQ after Comp:[newline ]
This method would make the compressor work harder as it's addressing all frequencies, both tame and excessive and nasty. This may be the sound that you're after, you have to experiment to see how you like it, but then you would tame your frequencies after the whole signal has been feed into your comp.
[*]
Boost EQ before Comp:[newline ]
Similar to the last example, but even more exaggerated as you're feeding even MORE to your comp. This might be desirable, as you can manipulate certain reactions from your comp by feeding more of a boosted frequency range. (again, one of the easiest ways to notice this change is feeding lots of low end to your compressor for a really strong effect)
[*]
Boost EQ after Comp:[newline ]
This might be a more polished approach, as here, your comp controls your signal making it easier to safely cut or boost the desired frequency ranges, as you will get an even and more dependable response.
[*]
A typical workflow:[newline ]
For a vox track, I often hipass up to 120-160Hz before compressing, then maybe tame a specific freq, eg. -1db at 700hz, then compress again before boosting for air, eg. +1db at 20K.
Even given my user name, I'm not the expert on EQ or compression, but this is my approach. It's important to learn the nuances between each approach and experiment to fine-tune your own, unique style. good luck