On 9/10/25 16:04, Alyssa Ross wrote:
> Demi Marie Obenour <demiobenour@gmail.com> writes:
> 
>> Use builtin string manipulation instead.
>>
>> Signed-off-by: Demi Marie Obenour <demiobenour@gmail.com>
>> ---
>>  scripts/make-erofs.sh | 14 +++++++++++++-
>>  1 file changed, 13 insertions(+), 1 deletion(-)
>>
>> diff --git a/scripts/make-erofs.sh b/scripts/make-erofs.sh
>> index cf942972910c76e1835dc5b0084c2d04bf084a9d..93cb3245f409b24c24be05e9307a1b2e12c867fe 100755
>> --- a/scripts/make-erofs.sh
>> +++ b/scripts/make-erofs.sh
>> @@ -78,7 +78,19 @@ while read -r arg1; do
>>  
>>  	check_path "$arg1" "$arg2"
>>  
>> -	parent=$(dirname "$arg2")
>> +	# The below simple version of dirname(1) can only handle
>> +	# a subset of all paths, but this subset includes all of
>> +	# the paths that check_path doesn't reject.
>> +	case $arg2 in
>> +	(*/*)
>> +		# Create the parent directory if it doesn't already
>> +		# exist.
>> +		parent=${arg2%/*}
>> +		;;
>> +	(*)
>> +		parent=.
>> +		;;
>> +	esac
>>  	mkdir -p -- "$root/$parent"
>>  	cp -RT -- "$arg1" "$root/$arg2"
>>  done
> 
> Saves about 600ms for me, and the improvement is just outside the margin
> of error.  What do we think?  Worth it?
> 
> (6aab79a is with patches 1–3 from this series applied;
>  7cc01b6 is patches 1–5.)
> 
> % hyperfine -w 1 -L commit 6aab79a,7cc01b6 --prepare 'git checkout {commit} && make clean && make build/etc/s6-rc' 'make'
> Benchmark 1: make (commit = 6aab79a)
>   Time (mean ± σ):     13.205 s ±  0.282 s    [User: 2.007 s, System: 6.397 s]
>   Range (min … max):   12.934 s … 13.698 s    10 runs
>  
> Benchmark 2: make (commit = 7cc01b6)
>   Time (mean ± σ):     12.662 s ±  0.290 s    [User: 1.675 s, System: 6.151 s]
>   Range (min … max):   12.371 s … 13.127 s    10 runs
>  
> Summary
>   make (commit = 7cc01b6) ran
>     1.04 ± 0.03 times faster than make (commit = 6aab79a)

Yup, worth it in my opinion!
-- 
Sincerely,
Demi Marie Obenour (she/her/hers)