SUMMARY

When formatting IDs, non-word characters (like dashes, colons, etc...) should be used as word separators. The current behaviour is to split words at whitespace, resulting in non-word characters ending up as part of words in the IDs, which can break autosuggestion in editors and generally doesn't look nice. IDs should be concise and not contain special characters as they are used for in-code referencing. If a user wants special characters in between words in IDs, there is a configuration option already provided in the ID suggestion editor.

STEPS TO REPRODUCE

1. Use the following example bibtex entry:

@article{testarticle,
	author = {Doe, John},
	title = {{Long-Term Measurements: A Better Technique}},
	year = {2020}
}


2. Create an ID suggestion like "Alw00|Y|Tlw01" (taken from ~/.config/kbibtexrc)  (first lowercased author, 4-digit year, then all title words with small words removed) and set it as default.

3. Format the ID of the entry.

OBSERVED RESULT

The ID is formatted as "doe2020long-termmeasurements:bettertechnique"

EXPECTED RESULT

ID gets formatted as "doe2020longtermmeasurementsbettertechnique"

SOFTWARE/OS VERSIONS

up-to-date Manjaro XFCE 

ADDITIONAL INFORMATION

kbibtex-git built from the AUR with this PKGBUILD fix: https://aur.archlinux.org/packages/kbibtex-git/#comment-754938