.
A lot of consideration has been given this topic since the advent of digital. This won't be the last time the topic has come up. The "answer"—or range of tolerance—is, obviously, going to vary with the context the photos are viewed in, the intent of the image maker, etc.
Since digital workflow allows for so much a wider range of potential manipulation than analogue (film/paper printing via enlarger), there's more to talk about.
Part of what comes into play is that decades of analogue photography have set a sort of base level of what a photographic image "should" look like. And that was influenced by the available tools, techniques, and technology at the time. So, film type (with inherent grain and color-rendering characteristics), printing manipulations (burning, dodging, printing through fabric), lens filters, etc., determined the various "looks" of what photo images had the potential to look like.
Radical manipulations (collage, photograms, very high contrast, solarization, etc.) were obvious, and wouldn't be confused with a "straight" photographic rendering.
But, even for the more standard kinds of photographic images, the "enhancement" techniques, if poorly or excessively applied, called attention to themselves and made it apparent that the image had been manipulated. Compare, for example, the mostly invisible burning, dodging, etc. of an Ansel Adams print with the clumsy burning and dodging common in photojournalism of the 70s and 80s, even in some high-profile publications.
For documentary and photojournalism work, i think there still are some fairly rigid lines that can't or shouldn't be crossed—staging or recreating a scene or event, adding or removing people or objects in an image, etc.
But outside of that context, in the current world, i don't think there's any way to get around the fact that, for many, anything goes, given the range of people taking pictures, the tools available to manipulate them, the many ways to share them, etc. I suspect that there will be some generational/age differences in how people think about this.
My own view, even with digital, is that i still draw a distinction between a traditional approach that produces a more or less "straight" image, and something that moves toward photo-illustration. Neither is better than the other, but i think about them in different ways.
The former puts the emphasis on the vision and attention involved in the recognition and evaluation of the elements of the original scene—light, tonalities, colors, relationship of the elements, moment, etc. Rendering and processing decisions should enhance or emphasize, but not materially alter them. I know there are still ambiguity here (HDR), but that's my starting point.
When an image is so drastically manipulated so that the aspects of the original scene become overwhelmed, obliterated, or otherwise so diminished as to become secondary, that crosses the (admittedly, at times, blurry) line to something i'll call illustration. It still may result in a compelling or interesting image, but i can no longer appreciate it in the same way as what i'm calling a "straight" photograph where the original recognition and capture of the scene has primacy.
.