Several notes:
For recording voice, people generally prefer a mic that
doesn't have a flat frequency response. But that's not really the question you asked (or even one of the questions).
For recording voice, people generally also prefer a directional mic. This is actually another facet of the preceding point, since cardiod mics, by their nature, typically have a presence peak when you sing into them at a fairly close distance. Again, not one of your questions, but it's information.
Those two mics almost certainly have the same
capsule. It's also highly likely that it's the same capsule that's used in various other small diaphragm mics. The XLR version you're interested in looks almost exactly like
the Behringer ECM 8000 and some other similarly-priced measurement mics wearing several different brandnames (Nady, Apex, etc.). There's a good possibility they all come out of the same factory and are the same in all respects save for the logo. That's not definite, though. They also look similar (though not identical) to some higher-priced measurement mics from Audix, Earthworks, etc. The various cheapies may just be separate knockoffs based on the same
originals. Pretty good chance they use the same capsule, though.
Though the list for the Behringer version is a little higher, it's readily available at a cheaper price than the one you've linked, on eBay or wherever.
Don't rely on frequency-response charts as truth, absent some independent reason to do so. It's pretty well established, I think, that individual ECM 8000s are different from one another, and (of course) different from the supposed precision response plots. That may not matter a
ton to you, as it might if you were using it as a measurement mic. If your goal is really just to be approximately flat at a low price, you're not likely to do a lot better than those choices. If you have a different goal, those may or may not be good choices.
it seems strange that the frequency response line is at -40db; what does this mean?
It doesn't mean anything. The output level they measured is just 40 dB lower than something. In the absence of knowing what that "something" is, -40 dB means nothing. If you knew the something, it still wouldn't really mean much standing by itself. FWIW, -40 dB just mean 1/10,000 the power of whatever the reference level is. It's a reasonable guess that the reference level is standard consumer line-level (-10 dBV), as that would be consistent with a fairly typiczal mic output. But that really doesn't have anything to do with the frequency response.
There are no differences (unless I missed something), other than the connection.
As I mentioned, the capsule is very likely the same. The electronics probably are different, though. To the extent the specs can be trusted (which, in this instance, they probably can), the XLR version requires at least 15V of phantom power. The TRRS version couldn't possibly expect that, or it wouldn't work when plugged into an iPhone. So it's highly likely they're different inside. The impedance, level, etc. all may be different.
I want to use the mic with a digital recorder; so which one of them do you thing will be better?
See prior post. It depends on the recorder. If it's an iPhone, the one with
the iPhone plug will work better, in that the other one won't work at all. As quite appropriately noted in the prior post, most little handheld recorders have built-in mics, which are likely as good or better in some sense than those (though probably not flatter), and which are free once you pay for the recorder.