[PATCH] test: say_color() in one write(2)

[PATCH] test: say_color() in one write(2)

[Tomi Ollila]

say_color() used to call (builtin) printf (and tput(1) to stdout)
several times, which caused attempts to write messages with color
to have partial content (e.g. escape sequences) often intermixed
with other tests when parallel tests were run.

Now, with all output collected, then written out using one
printf, all strings with color print out correctly
((at least short) write(2)'s appear to write out "atomically").

While at it, used only one tput(1) execution to determine whether
color output works, and made bold/colors/sgr0 to tput(1) their
values once per test.
---
 test/test-lib.sh | 61 +++++++++++++++++++++++-------------------------
 1 file changed, 29 insertions(+), 32 deletions(-)

diff --git a/test/test-lib.sh b/test/test-lib.sh
index d46bb4c3..d7c96a93 100644
--- a/test/test-lib.sh
+++ b/test/test-lib.sh
@@ -201,56 +201,53 @@ do
 done
 
 if test -n "$debug"; then
-    print_subtest () {
-	printf " %-4s" "[$((test_count - 1))]"
-    }
+	fmt_subtest () {
+		printf -v $1 " %-4s" "[$((test_count - 1))]"
+	}
 else
-    print_subtest () {
-	true
-    }
+	fmt_subtest () {
+		printf -v $1 ''
+	}
 fi
 
 test -n "$COLORS_WITHOUT_TTY" || [ -t 1 ] || color=
 
-if [ -n "$color" ] && [ "$ORIGINAL_TERM" != 'dumb' ] && (
-		TERM=$ORIGINAL_TERM &&
-		export TERM &&
-		tput bold
-		tput setaf
-		tput sgr0
-	) >/dev/null 2>&1
+if [ -n "$color" ] && [ "$ORIGINAL_TERM" != 'dumb' ] &&
+	tput -T "$ORIGINAL_TERM" -S <<<$'bold\nsetaf\nsgr0\n' >/dev/null 2>&1
 then
 	color=t
 else
 	color=
 fi
 
-if test -n "$color"; then
+if test -n "$color"
+then
+	# _tput run in subshell (``) only
+	_tput () { exec tput -T "$ORIGINAL_TERM" "$@"; }
+	unset BOLD RED GREEN BROWN SGR0
 	say_color () {
-		(
-		TERM=$ORIGINAL_TERM
-		export TERM
 		case "$1" in
-			error) tput bold; tput setaf 1;; # bold red
-			skip)  tput bold; tput setaf 2;; # bold green
-			pass)  tput setaf 2;;            # green
-			info)  tput setaf 3;;            # brown
-			*) test -n "$quiet" && return;;
+			error)	b=${BOLD=`_tput bold`}
+				c=${RED=`_tput setaf 1`}   ;; # bold red
+			skip)	b=${BOLD=`_tput bold`}
+				c=${GREEN=`_tput setaf 2`} ;; # bold green
+			pass)	b= c=${GREEN=`_tput setaf 2`} ;; # green
+			info)	b= c=${BROWN=`_tput setaf 3`} ;; # brown
+			*) b= c=; test -n "$quiet" && return ;;
 		esac
-		shift
-		printf " "
-		printf "$@"
-		tput sgr0
-		print_subtest
-		)
+		f=$2
+		shift 2
+		sgr0=${SGR0=`_tput sgr0`}
+		fmt_subtest st
+		printf " ${b}${c}${f}${sgr0}${st}" "$@"
 	}
 else
 	say_color() {
 		test -z "$1" && test -n "$quiet" && return
-		shift
-		printf " "
-		printf "$@"
-		print_subtest
+		f=$2
+		shift 2
+		fmt_subtest st
+		printf " ${f}${st}" "$@"
 	}
 fi
 
-- 
2.25.1

Re: [PATCH] test: say_color() in one write(2)

[David Bremner]

Tomi Ollila <tomi.ollila@iki.fi> writes:

> say_color() used to call (builtin) printf (and tput(1) to stdout)
> several times, which caused attempts to write messages with color
> to have partial content (e.g. escape sequences) often intermixed
> with other tests when parallel tests were run.

I still see things like:

T568-lib-thread: Testing API tests for notmuch_thread_*
 PASS   PASS   count messages on closed db
 get indexopts from closed database

Is that expected? Both "PASS" are coloured OK.

Re: [PATCH] test: say_color() in one write(2)

[Tomi Ollila]

On Sat, May 22 2021, David Bremner wrote:

> Tomi Ollila <tomi.ollila@iki.fi> writes:
>
>> say_color() used to call (builtin) printf (and tput(1) to stdout)
>> several times, which caused attempts to write messages with color
>> to have partial content (e.g. escape sequences) often intermixed
>> with other tests when parallel tests were run.
>
> I still see things like:
>
> T568-lib-thread: Testing API tests for notmuch_thread_*
>  PASS   PASS   count messages on closed db
>  get indexopts from closed database
>
> Is that expected? Both "PASS" are coloured OK.

Yes, that is (unfortunately) possible, say_color only writes that PASS
(with escape sequences coloring and reseting around it) (and subtest
message if $debug is set). The messages following are written in separate
lines, after say_color() returns (echo's if I remember correctly).

Much more overhaul would be required to get the full line written in
one write(1) -- I did think a few options when doing that change but
nothing seems clear enough at this moment...

Tomi

Re: [PATCH] test: say_color() in one write(2)

[David Bremner]

Tomi Ollila <tomi.ollila@iki.fi> writes:

> say_color() used to call (builtin) printf (and tput(1) to stdout)
> several times, which caused attempts to write messages with color
> to have partial content (e.g. escape sequences) often intermixed
> with other tests when parallel tests were run.

applied to master.

d