search

csutils / csdiff

Utilities for processing results of static analyzers, dynamic analyzers, and formal verification tools
GNU General Public License v3.0
18 stars 16 forks source link

RFE: eliminate repeated file paths/URL in plain-text/HTML output #109

Open kdudka opened 1 year ago

kdudka commented 1 year ago

While working on issue #90, Cedric Buissart suggested to eliminate repeated URLs in the plain-text/HTML output of csdiff. When dealing with alerts reported by OWASP ZAP, the URLs can be too long, which makes csdiff's output difficult to read by people.

Ideally, we should implement it in a way that the plain-text or rendered HTML output is still readable by csdiff's coverity parser.

We should also be careful with changing the default output for tools like GCC because the plain-text output is often used by IDEs like vim's quickfix. So either we should use the simplification for http(s):// URLs only or make it configurable by an option.

cedricbu commented 1 year ago

IMHO : "make it configurable by an option" sounds like a better approach to me. e.g. "--human-readable" which would enable something that's explicitly not-backward compatible (ie. future csdiff versions may make changes in the layout), made to be understood by human interpretation, not programs

kdudka commented 1 year ago

@cedricbu Thanks for the input! The problem is that end-users often have no idea what csdiff is, yet they want to copy paste the human readable output to their own lists of findings and expect such lists to be machine readable again when using them later on. For example, commit 7a56fbbdd0fbf01c622b17fcf41598cd3ae2de7a makes sure that HTML output rendered by a web browser is properly recognized as valid input data by csdiff.