I have used both lenses with a FujiFilm X-H1, which has a 1.5x crop sensor. Because I have used only a single copy of each lens, this response should be considered anecdotal. Also, my comments may not apply to uncropped, full-frame images.
Both of my copies of the lenses appear to have normally functioning focus, zoom, and aperture. They are free of fungus, haze, and balsam separation. The Canon optics are mark free. The front element of the Pentax has a small scratch about 1 cm off center, which I do not expect to be significant.
Build and Handling
Canon FD 35-105/3.5: Weighs in at 571g, with more external plastic components. Macro mode is accessible only at 35mm by sliding a switch to move the zoom ring past the 35mm mark. Focus, aperture, and zoom all turn counterclockwise. Filter threads are 72mm, and rotate while focusing.
Pentax-A SMC 35-105/3.5: Feels more solid at 585g, but the focus ring rattles a bit because to switch into macro mode involves a clutch mechanism on the focusing ring. This mechanism is easier to use than the switch on the Canon. It is also available at every focal length. Focus, aperture, and zoom turn opposite the direction of the Canon lens. Filter threads are 67mm and rotate while focusing.
On my camera, the Canon lens consistently meters about a third-stop slower than the Pentax lens at the same F-stops, including wide-open. I wouldn't have checked light transmission if the difference hadn't been noticeable.
Both lenses have 15 elements in 13 groups. (Canon, Pentax)
My cutoff for "Good Enough" is at 60-65 lp/mm. Images taken with "good enough" lenses still look good when pixel peeped, as long as they are not of resolution charts or brick walls. There's also (what I consider) "good enough for film" around 45 lp/mm, which sharpens up to 65 lp/mm when used with a focal reducer. Otherwise, "Excellent", "Very Good", and "Good" are relative, and I use some XF lenses as references (18-55/2.8-4, 18-135/3.5-5.6, 35/1.4). I check sharpness only wide open because a lens that is good enough wide open will be even better stopped down.
Canon: Over 100 lp/mm – "Excellent". Corners stay sharp. Central sharpness almost as good as the XF 18-55/2.8-4 R LM OIS.
Pentax: About 95 lp/mm – "Very Good". Corners soften a bit. Central sharpness better than the XF 35/1.4 R.
Quick comparison with some other lenses, shortest focal length, wide open. Imagine these are 1"x1" crops from a 40"x27" image.
Canon EF 24-105/4L @ 24/4
Nikon 24-120/4G @ 24/4 -and- Nikon 24-70/2.8E @ 24/2.8
Pentax-A SMC 35-105/3.5 @ 35/3.5 and 35/4.5
Canon FD 35-105/3.5 not present at the time. I'd expect it to look a little bit better than the Pentax because of its increased resolution and contrast.
Both lenses have mild pincushion distortion throughout the zoom range. I likely wouldn't notice in ordinary photographs, even if they include brick walls and other grid structures.
Aberration, Glow, Color, and Contrast
Neither lens has any significant color aberrations that I can see. Both lenses "glow" wide open. The Canon lens is more mild and resolves with a half-stop. The Pentax lens resolves with a full-stop. Colors from both lenses look good, and I can't tell the lenses apart based on color alone. The Canon lens has more contrast.
Bokeh from both lenses look good. The aperture of the Canon has a pentagonal shape, while the Pentax has a hexagonal shape. The following images were both taken wide open. (Canon, Pentax)
Veiling Glare, Flare Ghosts, and Sun Stars
Canon: Sun stars have ten points with prominent flare ghosts and some veiling glare. Usually, veiling glare is not a significant problem with this lens. However, I have had images turn out to be unusable because of veiling glare.
Pentax: Sun stars have six points with subtle flare ghosts and minimal veiling glare. There was no noticeable veiling glare while photographing in a church with large windows. This particular environment is particularly challenging for many lenses.
The Canon lens seems sharper with more contrast, but can't handle veiling glare and flare ghosts as well as the Pentax lens. The lens has enough difficulty with glare and flare that another copy would be unlikely to be significantly improved as a result of sample variation.
The Pentax lens is more fun to use because it feels and handles better, especially with respect to macro mode. However, the control rings turn in the "wrong" direction. Pentax SMC used in this lens is clearly superior to Canon SSC used in "New FD" lenses. The one-third stop better light transmission is also more useful than I'd expected when ISO is maxed out in low light.
The main problem with both lenses is they aren't 28-135mm.