In addition to adjusting coverage data using the Annotated Source window and source code directives, you can also make adjustments by manipulating the PureCoverage adjustments file directly, either manually or using scripts. This topic provides information about the format that PureCoverage uses to define adjustments in the adjustments file.
All adjustment files must start with this line, exactly as shown:
PureCoverage Adjustments File 1.0
The line must not contain any extra spaces or tabs. It must be the first line in the file, and cannot have a comment at the end.
Each additional line in the file can be arbitrarily wide. Lines beginning with a # in the leftmost column are comments. The format of each non-comment line is:
<name> <tab> <time> <tab> <type> <tab> <style> <tab> <list of adjustments>
where:
<name> is the canonical fullpath to the file containing adjustments. This is exactly the form of the name that appears in the Viewer.
<time> is the update time of the adjustments for that file as an int (UNIX time_t).
<style> is either an uppercase L, indicating that the adjustment came from a single-line adjustment, or an uppercase B, indicating that the adjustment came from a block-style adjustment.
<list of adjustments> is a list of single line numbers or ranges of lines, each separated by a space; for example, 4 10 15-23.
If there are multiple entries for a single file, PureCoverage uses all entries with times matching the most recent time given for that file; it discards the others. It also discards incorrectly formatted lines (or sometimes only portions of them). Depending on how you are reading the file, PureCoverage displays errors either in a dialog or on stderr.
Note:
If you have already annotated your code with markers that another tool understands, you can process your code into a format that PureCoverage can read without editing all your source files. For example, you can convert lint-style NOTREACHED comments into PureCoverage deadcode adjustments.