I'm always a bit worried to use sed -E anywhere, because the canonifiers
give the impression it won't work everywhere consistently. My manpage says
sed -E should be preferred for portability, so lets remove the
sed -r / sed -E differentiation assuming it's just a thing from the past.
All changes in this patch were performed automatically with `shfmt` with
configuration flags specified in `.pre-commit-config.yaml`.
In addition to fixing whitespace the roundtrip through shfmt's AST also
transforms command substitutions
`cmd`
# becomes
$(cmd)
and some redirects
>&2 echo "msg"
# becomes
echo >&2 "msg"
On Alpine (BusyBox `sed`), the previous `sed -e '$a\'` invocation always
added a newline, breaking most every diff. There doesn't seem to be a
need to attempt normalizing EOF newlines at the moment and doing it that
way doesn't seem to be portable anyway. If canonifiers need to be
portable, `sed` should be treated as a text-processing tool and POSIX
definition of text-file is zero or more newline-terminated
character-sequences, so if canonification of Baselines via `sed` is
required, those Baselines should always end with a newline to be
considered text files. I.e. that's not the job of this canonifier, and
changing it also doesn't necessarily generalize since it could be
considered coincidental that diff-remove-timestamps in particular is the
default canonifier that's commonly used while there's still others that
also make use of `sed`.
By default all baslines are run through diff-remove-timestamp. On a BSD
sed implementation, this means that a newline is added to the end of the
file, if no newline was there originally. This behavior differs from GNU
sed, which does not add a newline.
In this commit we unify this behavior by always adding a newline, even
when using GNU sed. This commit also disables the canonifier for a bunch
of binary baselines, so we do not have to change them.
This pattern got mislead by matching suffixes of other numbers, and
noramlizing exact 0-timestamps isn't really required.
- Remove eplicit "0.000000" number pattern from timestamp normalization
- Require beginning of line or non-numeric character before the
beginning of the number replacement
Unit tests sometimes use current_time(), which occassionally lands on
a perfect tenth of a second and end up fooling the old canonification
regexp and resulting in output which differs from what's in the
test baseline.
I edited the text little bit, reorganized the structure somewhat and
extended some parts. I've also simplified the tests a bit, using some
of the BTest tweaks commited in parallel.
frameworks.
There were a number of cases that weren't thread-safe. In particular,
we don't use std::string anymore for anything that's passed between
threads (but instead plain old const char*, with manual memmory
managmenet).
This is still a check-point commit, I'll do more testing.
Turns out the finish methods weren't called correctly, caused by a
mess up with method names which all sounded too similar and the wrong
one ended up being called. I've reworked this by changing the
thread/writer/reader interfaces, which actually also simplifies them
by getting rid of the requirement for writer backends to call their
parent methods (i.e., less opportunity for errors).
This commit also includes the following (because I noticed the problem
above when working on some of these):
- The ASCII log writer now includes "#start <timestamp>" and
"#end <timestamp> lines in the each file. The latter supersedes
Bernhard's "EOF" patch.
This required a number of tests updates. The standard canonifier
removes the timestamps, but some tests compare files directly,
which doesn't work if they aren't printing out the same
timestamps (like the comm tests).
- The above required yet another change to the writer API to
network_time to methods.
- Renamed ASCII logger "header" options to "meta".
- Fixes#763 "Escape # when first character in log file line".
All btests pass for me on Linux FC15. Will try MacOS next.
- The Makefiles now run btest with "-f diag.log" so that diangnostics
output will always be written into that file.
- Makefiles now hardcode path to btest to avoid picking up the wrong version
if in PATH.
- The canonifier scripts now live in testing/scripts, and they are
used from both btest/ and external/.
- There's a new diff-remove-uids scripts that removed connection UIDs
for diffing. The external/* tests now use that by default.
- Timestamp removal now has its own script: diff-remove-timestamps.
diff-canonifier calls that.
- All Makefile have a "brief" target that runs btest with -b.
- The higher-level directories have Makefile to call the subdirs.
2011-07-05 18:47:08 -07:00
Renamed from testing/btest/Scripts/diff-canonifier (Browse further)
