It really all depends on a lot of factors. First off, how well was everything tracked. Secondly, why are you using the compressor? Is it just because you have it, or is it because you hear something on the track that could benefit from a compressor. To expand on that, there are several reasons for using compression. The most common around here seems to be for volume control (dynamics). I prefer to not use compressors for that most of the time. I prefer to automate volumes... basically mixing. Sometimes I do use compressors for dynamic reduction though. It really depends on how the other tracks sit around it and how much the compressor is affecting this. The problem with compressors for dynamic volume control is that they are linear in the sense that they always compress when the threshold is hit. I prefer automation because it is easier to allow parts of a track to pop out or sink in without some machine deciding when and if it is appropriate.
Compressors however can also be used to shape the tone of a track, expand dynamics, or to shape the attack and decay envelopes of a track. Personally, this is how I prefer to use compression. I like using compression as a tool that becomes a part of the desired sound, and not as a tool to fix some sort of problem or symptom with a track.
In general, if you have to use a compressor on every channel, than there was probably some sort of problem during tracking. Most likely it would be due to poor musicianship (which is not the mix engineers fault, but does become his/her problem). It could however be due to poor tracking technique. Often this is caused by poor mic selection and por mic placement. Once again however, this may become the mix engineers problem to fix as many times retracking is just not a viable option.
Another thing that would change the way I personally use compressors is the quality of the compressors themselves. I have no problem throwing a compressor on the m ix bus, as long as it is a nice compressor that handles what I am asking it to do in a way that makes things sound better. If you are using a nice Manley, API, SSL or Drawmer compressor, I see no reason why it should not be used if desired. This is also assuming that it is being used in a way that will not be detrimental to future issues (such as mastering) by being applied in such a way that it is an enhancement, and not just a change in dynamics. Mastering is not just about compressing the 2 buss master. If thats all your "mastering engineer" thinks is his/her job, then you should be in the market for a better mastering engineer that actually understands the intricacies of their job. Mastering deals with volume control (be it dynamically or static), phasing issues, Minor (hopefully not major or the mix should be redone) EQ issues, fades, and overall sonic flow of an album. That is just a short list, but it can hopefully give you the overall impression of mastering. If a mix is done right, the master will be easy to achieve and the mastering will actually have a greater impact on the material in an enhancing sort of way.