|
From: | Jacob Bachmeyer |
Subject: | bug#67391: dejagnu report card and multiple rows |
Date: | Wed, 22 Nov 2023 20:41:40 -0600 |
User-agent: | Mozilla/5.0 (X11; U; Linux x86_64; en-US; rv:1.8.1.22) Gecko/20090807 MultiZilla/1.8.3.4e SeaMonkey/1.1.17 Mnenhy/0.7.6.0 |
Tom Tromey wrote:
I tried "dejagnu report card" today for gdb (well, the subset of gdb tests that I most recently ran). It shows: prentzel. dejagnu report card gdb.sum __________________________________________________ / PASS FAIL ?PASS ?FAIL UNSUP UNRES UNTEST |-------------------------------------------------- gdb | 8599 0 0 37 0 0 0 |-------------------------------------------------- | 8599 0 0 37 0 0 0 \__________________________________________________ I couldn't figure out why there would be two rows here. The documentation doesn't describe it either. Is the second row the sum?
Correct. If multiple passes were tested there is an additional group of per-pass total lines before the grand total at the end. The grand total is always produced and was expected to be self-documenting, as recognizing it as the sum of the tool lines was expected to be obvious. It turns out to be a bit less obvious when there is only one tool...
If so (1) if there's a single tool, maybe it can be omitted; and (2) perhaps this could be mentioned in the documentation, maybe with some example output.
There are examples in the dejagnu-report-card(1) manpage. The "dejagnu help report card" command should display it. I seem to recall omitting them from the Texinfo manual for reasons of style, but manpages traditionally have an EXAMPLES section, so there they are.
Given that the pass totals are already omitted if not relevant, omitting the grand totals if only one row was emitted is a future possibility.
The original use case for "dejagnu report card" was to collect results from testing multiple tools in a convenient summary display. If you are only testing one tool, DejaGnu itself produces a similar summary, although not quite as compact.
In any case, a documentation patch has been pushed on the PR67391 branch. Does it resolve the confusion?
-- Jacob
[Prev in Thread] | Current Thread | [Next in Thread] |