Can you make a quick list of the dna members that could be affected by this change?

There are 43 total char * members in DNA*types.h files (including const char *). Here I'll categorize them:

• Affected by change (Exposed to RNA with no custom get and set), null checks everywhere: SpreadsheetRowFilter::value_string, NodeInputString::string, bNodeSocket::default_attribute_name, Curves::surface_uv_map
• Affected by change (Exposed to RNA with no custom get and set), null checks in most places: SpreadsheetContextNode::node_name, SpreadsheetContextModifier::modifier_name, SpreadsheetColumnID::name
• Exposed to RNA with custom get and set, does use null: TextLine::format, NodeCryptomatte::matte_id, ShaderNodeScript::bytecode, TexPaintSlot:;uvname, Curve::str, AssetMetaData::author, AssetMetaData::description, KS_Path::rna_path, FCurve::rna_path, DriverTarget::rna_path
• Exposed to RNA with custom get and set, does not use null: Text::filepath, TextLine::line, ConsoleLine::line
• Exposed to RNA with custom get, no set: FileDirEntry::relpath, FileDirEntry::name, TexPaintSlot::attribute_name
• Not exposed to RNA, runtime data: WorkSpace::status_text, Report::typestr, Report::message, FileDirEntry::redirection_path, ShaderFxData::error, ARegion::headerstr, SpreadsheetColumn::display_name, ARegion_Runtime::category, ModifierData::error, GpencilModifierData::error, AnimOverride::rna_path
• Not exposed to RNA, other: RE_engine_id_BLENDER_EEVEE, RE_engine_id_BLENDER_WORKBENCH, RE_engine_id_CYCLES, SDNA::data, PointCache::cached_frames, Object::matbits, LightCacheTexture::data
• Totally unused: uiList::custom_drag_opname, uiList::custom_activate_opname

That was interesting! This is good because the only areas that are affects are code that we're responsible for already. The change looks safe to me.
Many of the items in the "Exposed to RNA with custom get and set, does use null" list could be simplified to avoid defining RNA getters and setters manually after this. I'll leave that for another time though.