Straight and curly φ are swapped with IBM Plex Serif

[slotThe]

Hi,

the unicode characters φ (U+03C6, φ in HTML) and ϕ (U+03D5, ϕ in HTML) seem to be swapped in the IBM Plex Serif font: the latter is displayed as a curly phi, while the former has a straight appearance:

IBM Plex Sans does not seem to have this problem (again, the following image shows U+03C6, followed by U+03D5)

(I have only checked these two characters, so there may be others showing the same behaviour)

[moyogo]

U+03C6 GREEK SMALL LETTER PHI can have several shape but U+03D5 GREEK PHI SYMBOL should only have the technical symbol shape with the tall ascender-height to descender-height straight stroke.

They don't need to be swapped to be fixed, they could both have the tall stroke like the technical symbol. If they look different then only U+03C6 GREEK SMALL LETTER PHI doesn't have the tall stroke.

[slotThe]

The swap was merely suggested for consistency with the sans serif typeface. If both end up having the same appearance then I suppose that would be—while unconventional—fine as well.

[kamekku14]

Erm, looks like the lowercase Phi in the Serif variant used the Porson style rather than using the Didot style like in the Sans variant, but according to the Wikipedia article:

Because Unicode represents a character in an abstract way, the choice between glyphs is purely a matter of font design. While some Greek typefaces, most notably those in the Porson family (used widely in editions of classical Greek texts), have a "stroked" glyph [...], most other typefaces have "loopy" glyphs. This also applies to the "Didot" (or "apla") typefaces employed in most Greek book printing, as well as the "Neohellenic" typeface often used for ancient texts.
It is necessary to have the stroked glyph available for some mathematical uses, and U+03D5 GREEK PHI SYMBOL is designed for this function. Prior to Unicode version 3.0 (1998), the glyph assignments in the Unicode code charts were the reverse, and thus older fonts may still show a loopy form at U+03D5.