https://bugs.dpdk.org/show_bug.cgi?id=238
Bug ID: 238
Summary: [tree-wide] enhance getopt_long usage
Product: DPDK
Version: unspecified
Hardware: All
OS: All
Status: CONFIRMED
Severity: enhancement
Priority: Low
Component: other
Assignee: dev@dpdk.org
Reporter: david.march...@redhat.com
Target Milestone: future
We have quite a couple of the following pattern in the apps and examples
(quoting examples/fips_validation/main.c as a bad example):
struct option lgopts[] = {
{REQ_FILE_PATH_KEYWORD, required_argument, 0, 0},
{RSP_FILE_PATH_KEYWORD, required_argument, 0, 0},
{FOLDER_KEYWORD, no_argument, 0, 0},
{CRYPTODEV_KEYWORD, required_argument, 0, 0},
{CRYPTODEV_ID_KEYWORD, required_argument, 0, 0},
{CRYPTODEV_ST_KEYWORD, no_argument, 0, 0},
{CRYPTODEV_BK_ID_KEYWORD, required_argument, 0, 0},
{CRYPTODEV_BK_DIR_KEY, required_argument, 0, 0},
{NULL, 0, 0, 0}
};
while ((opt = getopt_long(argc, argvopt, "s:",
lgopts, &option_index)) != EOF) {
switch (opt) {
case 0:
if (strcmp(lgopts[option_index].name,
REQ_FILE_PATH_KEYWORD) == 0)
env.req_path = optarg;
else if (strcmp(lgopts[option_index].name,
RSP_FILE_PATH_KEYWORD) == 0)
env.rsp_path = optarg;
else if (strcmp(lgopts[option_index].name,
FOLDER_KEYWORD) == 0)
env.is_path_folder = 1;
else if (strcmp(lgopts[option_index].name,
...
We can get rid of all those strcmp by directly binding each longopt with an int
enum.
eal has been using this convention for quite some time, see:
https://git.dpdk.org/dpdk/tree/lib/librte_eal/common/eal_options.h?h=v19.02
https://git.dpdk.org/dpdk/tree/lib/librte_eal/common/eal_common_options.c?h=v19.02
--
You are receiving this mail because:
You are the assignee for the bug.
