I'm with you but this is definitely more and more of a battle each year. It's usually for a variety of reasons:
1) The artist/band didn't tell the mixing engineer they were going to use a dedicated mastering engineer, so the final mixes were sent with extreme peak-limiting which of course can't be undone in mastering. Sometimes it's easy to get non-limited versions to master from if you ask, or sometimes it's a big ordeal or not possible.
2) The mix engineer subscribes to "top down" mixing which is popular with the younger generation but I don't particularly believe in this. Yes, it's not a bad idea to throw a temporary limiter on your mix to see how it might
react in mastering, but then of course remove this limiter prior to mastering, or send both version to the mastering engineer. I think if you want to create a great mix that will stand the test of time, and sound great on all formats and situations, mixing into a limiter is not a good idea.
If I do ask for a non-limited version to master from, there is a 98% chance the mix engineer will complain that the mix "falls apart" without the limiter. I usually don't say this to them but in my opinion, if the mix falls apart without peak limiting, it's probably not a great mix. But also, it's going to come right "back together" when peak-limiting is appleid AT THE END of the mastering process, after we've corrected any issues to the best of our abilities. Starting the mastering process from peak-limited material is likely to go downhill fast. Limiting is not exactly a tricky process. It's similar to pouring in a nail, so recreating what they've done is not exactly hard but again, AFTER any processing to correct issues and get all the songs on the same page.
The other complaint when asked to remove the peak-limiter is that there are a few stray overs, and they don't want to have to trim down their mix channels and screw things up, so that's why there is a limiter on there prior to mastering.
This is the easiest problem in the world to solve and requires no mixer changes. Remove the limiter and simply save the "in the box" bounce as 32-bit float instead of 24-bit and the peaks are preserved. This is possible assuming there isn't a peak limiter or other plugin that creates a ceiling at digital 0.
On more than one occasion a mix that was said to have "no limiting" turned into, well, there is a limiter but it's not doing much. My answer is to turn it off, save the bounce as 32-bit float, and that is something I can work with. So engineers have thought that putting the threshold to 0 was the same as removing the limiter. That would be a no unless the peaks are always below 0dBFS in which case, then you don't need the limiter. So just remove it completely and save as 32-bit float is my advice.
There are some people out there that think you can just turn down the file no matter what if you need/want more headroom which is true, but you can't undo the brick-wall processing and some people just don't get it no matter how you explain it.
Then they ask "how much" headroom I need. I tell them it doesn't matter, just don't use a peak-limiter, try not to hit 0dB so your monitoring DAC isn't potentially distorting, and save the bounce as 32-bit floating point and that will be ok.
It drove me to write this article:The 6 dB of Headroom for Mastering Myth Explained