A PDF page stores neither letters nor words, only codes that pick glyphs, so the file must say separately which character each glyph stands for. Unicode’s glossary describes the same split from the other side: “In displaying Unicode character data, one or more glyphs may be selected to depict a particular character.”

Can one character have several glyphs, or one glyph several characters?

Both. Every typeface and weight draws its own a; the demo cover letter made for PDF Textfix’s screenshots has one in each of its two weights. The other way round, a ligature is “A glyph representing a combination of two or more characters,” such as fi. One code draws it, and the file’s ToUnicode map must say it means f and i, or copying gives a stray symbol.

What does it mean when a font lacks a character?

That it has no glyph for it. Many PDFs embed subsets holding only the glyphs the document used: the cover letter’s SemiBold subset gives widths to just 21 codes, with no s, m or y, and none for é. Type é into its heading in PDF Textfix and it turns orange; the editor bar names the font that draws it, and the Issues pane lists the line.

PDF Textfix editing the heading Renée Okafor, the writer's own name, in the demo cover letter, set in IBM Plex Serif SemiBold. Only the letter é is orange: the font subsets in the file have no glyph for it, so another font draws it.
No subset here has a glyph for é, so a fallback font draws it in orange.

PDF Textfix’s own wording says “character”, never “glyph”. Red marks a character no font can draw. The fallback font entry gives the search order.

Where do you see glyphs and characters in a PDF?

In the content stream, which qpdf --show-object=N --filtered-stream-data prints for any PDF (brew install qpdf poppler adds qpdf and Poppler’s pdftotext to a Mac). This piece of the demo cover letter draws its heading, the writer’s name:

/TT1 1 Tf [ (R) 22 (enee Oka) 1 (f) 21 (or) ] TJ

Each byte is a code. TT1, the SemiBold subset, uses the Mac Roman encoding, so its code for R selects the glyph of R, and pdftotext prints the name with no ToUnicode map. An “é” here would need code 8E, for which the subset holds no glyph.

Does an edit remove the replaced glyphs?

No, fonts are copied as they are. PDF Textfix’s Help says: “Fonts keep the shapes of all their characters, those of the text you replaced included.” It is one reason an edit is not a redaction.

Questions about glyphs and characters

What is a glyph ID?

The number of a glyph inside one font; another font may number the same glyph differently.

What is the .notdef glyph?

The shape a font shows for a code it has no glyph for, often an empty box.

Why does copied text come out wrong when the page looks right?

Drawing needs glyphs; copying needs characters. A missing or wrong ToUnicode map breaks only the copy.