[pgpool-hackers: 246] Re: Ideas for code cleanup

[Tatsuo Ishii]

>> I have a question: you replaced strcmp to strlen in your patches.
>> Why do you prefer strlen over strcmp?
> 
> Well, I thought testing for a zero-length was clearer than comparing to
> "".  Another approach is to test if [0] == '\0'.  I had just never seen
> the "" comparison before, and it looked odd.  I think we should just
> pick a method and use it consistently, probably based on how the
> Postgres does it.

I see these in the PostgreSQL source code:
find . -name '*.c' -print|xargs grep -i strcmp|grep \"\"

./timezone/zic.c:		if (strcmp(cp, "") == 0)
grep: ./bin/pg_controldata/pg_crc.c: そのようなファイルやディレクトリはありません
./bin/initdb/initdb.c:							  (strcmp(authmethodlocal, "trust") == 0 || strcmp(authmethodhost, "trust") == 0) ? AUTHTRUST_WARNING : "");
./bin/initdb/initdb.c:	if (strcmp(xlog_dir, "") != 0)
./bin/psql/describe.c:						   strcmp(PQgetvalue(res, 0, 8), "") != 0) ?
./bin/psql/describe.c:		if (strcmp(PQgetvalue(res, i, 7), "") != 0)
./bin/pg_dump/pg_backup_custom.c:		if (AH->fSpec && strcmp(AH->fSpec, "") != 0)
./bin/pg_dump/pg_backup_custom.c:		if (AH->fSpec && strcmp(AH->fSpec, "") != 0)
./bin/pg_dump/pg_backup_custom.c:	if (AH->fSpec == NULL || strcmp(AH->fSpec, "") == 0)
./bin/pg_dump/pg_backup_tar.c:		if (AH->fSpec && strcmp(AH->fSpec, "") != 0)
./bin/pg_dump/pg_backup_tar.c:		if (AH->fSpec && strcmp(AH->fSpec, "") != 0)
./bin/pg_dump/pg_backup_archiver.c:		(AH->fSpec == NULL || strcmp(AH->fSpec, "") == 0))
./bin/pg_dump/pg_backup_archiver.c:	if (strcmp(want, "") == 0)
./bin/pg_dump/pg_dump.c:	if (strcmp("", evtinfo->evttags) != 0)
./bin/pg_dump/pg_backup_directory.c:	if (!AH->fSpec || strcmp(AH->fSpec, "") == 0)
./backend/commands/dbcommands.c:			if (strcmp(encoding_name, "") == 0 ||
--
Tatsuo Ishii
SRA OSS, Inc. Japan
English: http://www.sraoss.co.jp/index_en.php
Japanese: http://www.sraoss.co.jp