The symbolizer markup syntax is structured such that fields require only
previous fields for their interpretation; this was originally intended
to make adding new fields a natural extension mechanism for existing
elements. This codifies this into the spec and makes the behavior of the
llvm-symbolizer match. Extra fields are now warned about, but ignored,
rather than ignoring the whole element.
Details
Diff Detail
- Repository
- rG LLVM Github Monorepo
Event Timeline
llvm/lib/DebugInfo/Symbolize/MarkupFilter.cpp | ||
---|---|---|
136–137 | I don't know of a reason to print anything at all when a reset element is encountered. | |
llvm/test/DebugInfo/symbolize-filter-markup-reset.test | ||
24–26 | A trailing colon is an odd case, though I see no reason not to accept it. But it's never what an expected future extension would look like. Probably the test should cover reset:this=1:that, etc. |
Update test cases.
llvm/lib/DebugInfo/Symbolize/MarkupFilter.cpp | ||
---|---|---|
136–137 | This is mostly just to provide context for error messages; it would be highly surprising to see a "no mmap covers address" error not too long after a mmap that appears to cover the address, but where a reset was present in-between. | |
llvm/test/DebugInfo/symbolize-filter-markup-reset.test | ||
24–26 | Good point; the tool is agnostic to what the field looks like, and these tests are fairly white-box, so I've changed these to add an "ext" field to make the purpose of the case clearer. |
I don't know of a reason to print anything at all when a reset element is encountered.