Changes have been made to block rules for footnotetext and footnote.
Previously, we accumulated content to find opening { and closing } tags.
This resulted in slower performance for large documents.
Now, instead, we search for these tags for the current line and store the result about the number of open parentheses.
Next, for the next line, we search for opening and closing tags based on this value.
This significantly speeded up the parser.
branch: dev/olga/6637-Hanging-conversions
Why:
Fixes [ocr-api] Hanging conversions #6637
What's being changed:
Changes have been made to block rules for footnotetext and footnote.
Previously, we accumulated content to find opening { and closing } tags. This resulted in slower performance for large documents.
Now, instead, we search for these tags for the current line and store the result about the number of open parentheses. Next, for the next line, we search for opening and closing tags based on this value. This significantly speeded up the parser.