Hi,
great work with the setup of certs and krb5 and such.
Always a pain manually and a pleasure if that would be tested in autopkgtest.
Also thanks for already Bileto-Testing them on all arches, that is really good.
I only found a few minor things, but even small things can make it better.
debian/tests/util has a few trailing whitespace.
line 86: after " } | ldapmodify -H ldapi:/// -Y EXTERNAL -Q· "
The last line (257 atm) is an empty line
In functions of util like check_getent_group check_getent_user you use var "output" I wonder if you should make that a "local" var as well as you did with most others to avoid any later issues clobbering a global one.
You might consider some sleeps, but since sleeps always suck I checked what I added in the past and found that I used this most of the time:
set send_human {.1 .3 1 .05 2}
You might give that (or sleeps) a consideration and let me know what you think about it.
Eventually there also is another style question, I realized that for
$ wc -l debian/tests/ldap-user-group-*
57 debian/tests/ldap-user-group-krb5-auth
51 debian/tests/ldap-user-group-ldap-auth
108 total
It is only
$ diff -Naur debian/tests/ldap-user-group-* | diffstat
ldap-user-group-ldap-auth | 12 +++---------
1 file changed, 3 insertions(+), 9 deletions(-)
So I wonder, should these two maybe better be ONE file with an argument?
The latter is a call up to you, I currently feel a single file would be better, but maybe you had more planned in the future that would make these differ more?
Hi,
great work with the setup of certs and krb5 and such.
Always a pain manually and a pleasure if that would be tested in autopkgtest.
Also thanks for already Bileto-Testing them on all arches, that is really good.
I only found a few minor things, but even small things can make it better.
debian/tests/util has a few trailing whitespace.
line 86: after " } | ldapmodify -H ldapi:/// -Y EXTERNAL -Q· "
The last line (257 atm) is an empty line
In functions of util like check_getent_group check_getent_user you use var "output" I wonder if you should make that a "local" var as well as you did with most others to avoid any later issues clobbering a global one.
I saw that in some tests the expect seems "too fast". /objectstorage. prodstack4- 5.canonical. com/v1/ AUTH_77e2ada1e7 a84929a74ba3b87 153c0ac/ autopkgtest- cosmic- ci-train- ppa-service- 3429/cosmic/ amd64/s/ sssd/20180923_ 232734_ 50b78@/ log.gz
For example https:/
There you see the answer of "id -un" and the "klist" on one line.
You might consider some sleeps, but since sleeps always suck I checked what I added in the past and found that I used this most of the time:
set send_human {.1 .3 1 .05 2}
You might give that (or sleeps) a consideration and let me know what you think about it.
Eventually there also is another style question, I realized that for tests/ldap- user-group- * tests/ldap- user-group- krb5-auth tests/ldap- user-group- ldap-auth tests/ldap- user-group- * | diffstat group-ldap- auth | 12 +++---------
$ wc -l debian/
57 debian/
51 debian/
108 total
It is only
$ diff -Naur debian/
ldap-user-
1 file changed, 3 insertions(+), 9 deletions(-)
So I wonder, should these two maybe better be ONE file with an argument?
The latter is a call up to you, I currently feel a single file would be better, but maybe you had more planned in the future that would make these differ more?