I don't have the page number reference handy but in his book The Technique and Spirit of Fugue George Oldroyd gave a bit of advice that I have found valuable that I think anyone attempting to write a fugue should consider. He wrote that too many students think they have done their work in writing a fugue if they simply have a bunch of melodies that all happen to keep the rules for counterpoint. He said that this is the difference between a merely competent writer of fugues and an actually good one.
His advice is that if you have yourself a subject and two countersubjects then lay them all out in sequence in the order in which they appear in your exposition as a continuous melodic line. Then sing that. If you have written a good fugue exposition then you will yourself happy with that melodic line. If you have ostensibly good counterpoint but you find that your three of our successive tunes don't form a musically pleasing single line then you had probably better go back to the drawing board. Somewhere along the way you wrote a merely perfunctory countersubject or maybe even have some problems in your subject itself. As someone who was steeped in choral music in college and even a bit in high school I find this test prescribec by Oldroyd to be immensely valuable. In fact I would say its value is impossible to overstate.
As a guitarist I feel it's even more important to put the whole issue of developing contrapuntal music in these terms because it can seem as though guitarists, when it comes to counterpoint, can be a bit easily pleased. By that I mean that we have a history of occasionally writing fugues while not working on aspects such as the interchangeable aspect of the voices in the exposition. If we compose expositions with no countersubjects and our middle entries employ long strips of free material then, yes, we've written ourselves some fugues because multiple melodies are moving around but my hope is that we guitarists can, over time, expand our approach to counterpoint as both a discipline and an art to incorporate more than simply having a few melodies going on at the same time that, were we to switch around their relationships, would reveal parlalel fifths and octaves. Our instrument is not a particularly forgiving one, I know! But we can still try.
That said, go get the fugal cycles of Castelnuovo-Tedesco, Rekhin, and Koshkin's (when those get published). When you stop to think about it we guitarists have seen only in the last fifty years the creation of a systematic approach to the art of counterpoint that has been with keyboard literature for centuries. We have reason for cautious optimism. Our instrument's limitations may always remain severe but it is good to know there are guitarists and composers willing to tackle this as yet rarely explored possibility for our instrument. In fact Koshkin is the first guitarist-composer I know of to directly tackle the project. He's the first because he finished his cycle and is getting it published. I may be second or third. We'll see. I still am hunting for a regular job and am not a professional so I have a few different circumstances than Koshkin. I am, however, very much looking forward to studying and listening to his set!