The Gettext file format supports flags on individual strings, which can mark a string as, say, c-format, depending on whether that string is used with printf (or a special comment xgettext:c-format is used). This allows gettext tools to perform printf validations, without getting false positives every time a percent sign "%" appears in a string. See https://www.gnu.org/software/gettext/manual/html_node/c_002dformat-Flag.html

We should allow Zanata to store similar information about individual TextFlows, so that we can eg, hard-enforce validations for the right strings (ZNTA-301), and avoid spurious warnings for others (ZNTA-944).

Some formats we should know about: HTML (tag validation), XML (tag/entity validation), Java Formatter, C-printf, XSI-printf, Python-printf(ZNTA-1212), EL (ZNTA-136). See the implementations of org.zanata.webtrans.shared.model.ValidationAction for all our current validators. We should make it easy to add more formats, even if the validators come later.