Replace special characters with normal characters - 2018-May Core STU #144 #25 (IG-Publisher)

lmckenzi commented 3 years ago

(from https://jira.hl7.org/browse/FHIR-16953)

Change the IG publisher to spit out a warning if it sees any of these codes and prompt the user to replace with the 'standard' representation for the particular code (regular quote, dash, etc.). If an author genuinely needs one of these codes, they can suppress the warning and the FMG and others can review the suppression for appropriateness.

Candidate characters:

Single left quotation mark: U+2018 -> U+0027
Single right quotation mark: U+2019 -> U+0027
Double left quotation mark: U+201C -> U+0022
Double right quotation mark: U+201D -> U+0022
Hyphen: U+2010 -> U+002D
Non-breaking Hyphen: U+2011 -> U+002D
Figure Dash: U+2012 -> U+002D
En dash: U+2013 -> U+002D
Em dash: U+2014 -> U+002D (2 reps)
Horizontal Bar: U+2015 -> U+002D (2 reps)
2-Em dash: U+2E3B -> U+002D (4 reps)
3-Em dash: U+2E3A -> U+002D (6 reps)
Small em dash: U+FE58 ->U+002D
Small hyphen minus: U+FE63 -> U+002D
Ellipses: U+2026 -> U+002E (3 reps)
Horizontal midline ellipses: U+22EF -> U+002E (3 reps)

lmckenzi commented 2 years ago

Lloyd will raise this on the IG creation stream

lmckenzi commented 2 years ago

https://chat.fhir.org/#narrow/stream/179252-IG-creation/topic/Expectations.20around.20'smart'.20punctuation

Discussion on the call suggests that a configuration option might be best. Trick is how to catch everything. Current proposal:

run the check once after resource load (and narrative generation) but before generating fragments
run the check as well against generated HTML pages
if warnings are raised, they should show the characters surrounding the problematic character

@grahamegrieve will think about where the right places in the code might be to do this

HL7 / fhir-ig-publisher

Replace special characters with normal characters - 2018-May Core STU #144 #25 (IG-Publisher) #275