To Mend A Broken Heart

It's a very beautiful hymn melody. And the harmonies are ok, I think. But I'm not that sure about how the voices are treated. I have a classical education and I find myself disturbed when the rules are not followed. I find myself asking if the arranger deliberately broke the rules or did he just not know them. The worst case is when an arranger (or composer) wants to write something that sounds "classic", but doesn't know the rules. A better case is when he knows the rules and knows how to break them so that it will be obvious. And maybe even better would be if he just would write something completely free from associations to classic rules.
Of course, if one would pick this beautiful melody and maybe the harmonies, too, and would arrange it according to the book, it might sound too real, too old school.
I might be completely lost here and can't see the value of diverging from the rules at some spots in your score. On the other hand, some of the diverging I like a lot. I try to pick them all, after listening one more time to the score. I did listen to the earlier version, too, and that might have opened some things for me, too.
Here they are:
Bar 3: three voices, B, B and F#. I miss the D, which would complete the triad. The open chord might sound good, but it might also sound like one player is missing.
Bar 4: same thing. G, D and D. And here melody plays D D C# B, while one voice plays D B. This parallel voicing sounds empty. Again, it might be on purpose or based on some other idea, but it might also sound like someone not knowing what he is doing.
Bar 5: Between same voices we have the D C# sustention while the other voice plays a C#.
Bars 7-9: All four voices are present, but it sounds like the bass line is missing. A missing bass line can sound good if you have less voices. The same effect with four voices, especially in the chords with the 5th in the lowest voice, gives the feeling that it's not finished. The open ending chord in bar 9 sounds good. The missing 3rd (F#) is ok this time.
Bars 10-: Here the harmonies are a bit more complete, kind of setting the style, which would promise that all the rest should follow the same style. But bar 14 is again a bit empty. A C# D sustention together with a D. A C# D sustention together with a B (4-3 sustention) would support the modulation to the key A better. According to the old school book again.
Bar 28-29: The F#m chord is a bit too static. You got it in 28, 29 and 30 on the 1st beat. Though the tonality is a bit free in the whole tune, this static feel disturbes me a bit. It goes back from F#m to A, through this very sustained F#m, then suddenly F#m E A in 30-31.
The crescendo in 35 would perhaps be supported by a complete triad. (Missing the G# here).
Bars 36-37: You go from the D on an up beat to the D on a downbeat in 37. Again, forbidden in the book.

I'm not quite sure if my software plays all parts in their right octaves. According to the sscore it does. I never checked if it were possible to get rid of the chords with the 5th in the bass, by transposing another part one octave down.
An important note is of course that all these "empty" spots and rule breaking spots might sound much better, when played by real musicians knowing what it's all about.

This took longer than I expected. And I went nit picking. I really should return to this and point out the good things I found, things where you broke the rules in a way that appealed to me.

Changed the ending for flute choir