To avoid information loss, introduce a difference between unset stats
and 0 for statistics that are supposed to be set once per entry point.
Now, if the statistic is not set for an entry point, the corresponding
CSV cell will be empty, and not 0.
Thanks to this differentiation, I noticed that `PathRunningTime` was
actually never set, and fixed that.
Additionally, this patch enables the timers if
`DumpEntryPointStatsToCSV` is set, because in most cases you dump these
stats to get a detailed view on analyzer performance.
Finally, I added a dedicated debug checker that demonstrates the use of
a statistic and tested the set and unset scenarios explicitly.
--
CPP-7097
---------
Co-authored-by: Donát Nagy <donat.nagy@ericsson.com>
The complexity of maintaining an extra kind of metrics have not
justified itself, as it is not used upstream, and we have only a single
use of boolean stats per entrypoint downstream.
As I will do downstream, you can use an unsigned statistic type with
values 0 and 1 to model a boolean flag.
--
CPP-7097
Also remove the gratuitous spaces after "," that break strict CSV
compliance.
As the debug function name does not uniquely identify an entry point,
add the main-TU name and the USR values for each entry point snapshot to
reduce the likelihood of collisions between declarations across large
projects.
While adding a filename to each row increases the file size
substantially, the difference in size for the compressed is acceptable.
I evaluated it on our set of 200+ open source C and C++ projects with 3M
entry points, and got the following results when adding these two
columns:
- Raw CSV file increased from 530MB to 1.1GB
- Compressed file (XZ) increased from 54 MB to 78 MB
--
CPP-7098
---------
Co-authored-by: Balazs Benics <balazs.benics@sonarsource.com>
So far CSA was relying on the LLVM Statistic package that allowed us to
gather some data about analysis of an entire translation unit. However,
the translation unit consists of a collection of loosely related entry
points. Aggregating data across multiple such entry points is often
counter productive.
This change introduces a new lightweight always-on facility to collect
Boolean or numerical statistics for each entry point and dump them in a
CSV format. Such format makes it easy to aggregate data across multiple
translation units and analyze it with common data-processing tools.
We break down the existing statistics that were collected on the per-TU
basis into values per entry point.
Additionally, we enable the statistics unconditionally (STATISTIC ->
ALWAYS_ENABLED_STATISTIC) to facilitate their use (you can gather the
data with a simple run-time flag rather than having to recompile the
analyzer). These statistics are very light and add virtually no
overhead.
Co-authored-by: Balazs Benics <benicsbalazs@gmail.com>
CPP-6160
