http2: track available data as it changes instead of recomputing it
With a large amount of available data, this computation can become
costly.
$ check_code="$TESTDIR"/../contrib/check-code.py
$ cd "$TESTDIR"/..
$ if hg identify -q > /dev/null 2>&1; then :
> else
> echo "skipped: not a Mercurial working dir" >&2
> exit 80
> fi
New errors are not allowed. Warnings are strongly discouraged.
$ hg manifest 2>/dev/null \
> | xargs "$check_code" --warnings --nolineno --per-file=0 \
> || false
Check Python files without py extension
$ cp \
> hg \
> hgweb.cgi \
> contrib/convert-repo \
> contrib/dumprevlog \
> contrib/hgweb.fcgi \
> contrib/hgweb.wsgi \
> contrib/simplemerge \
> contrib/undumprevlog \
> "$TESTTMP"/
$ for f in "$TESTTMP"/*; do cp "$f" "$f.py"; done
$ "$check_code" --warnings --nolineno --per-file=0 "$TESTTMP"/*.py \
> || false