From: Alexander Turenko <alexander.turenko@tarantool.org>
To: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: tml <tarantool-patches@dev.tarantool.org>
Subject: Re: [Tarantool-patches] [PATCH 7/7] test/unit: add popen test
Date: Wed, 11 Mar 2020 23:22:13 +0300 [thread overview]
Message-ID: <20200311202213.pugz24653u7zlk4f@tkn_work_nb> (raw)
In-Reply-To: <20200302201227.31785-8-gorcunov@gmail.com>
There are two good properties, which can be easily achieved for the
test:
* better fit TAP13 format;
* set a non-zero exit code at fail.
See below (comments 2 and 4) if you're interested or skip if not.
There also two comments around (1 and 3), see them below too.
WBR, Alexander Turenko.
> +static int
> +main_f(va_list ap)
> +{
> + int rc = 0;
> +
> + rc = popen_write_exit();
> + if (rc == 0)
> + rc = popen_read_exit();
> + if (rc == 0)
> + rc = popen_kill();
1. I would run all tests even if some of them fails: this way we can get
more info what is healthy and what is broken. And it is usual for
TAP13 tests as I see.
> +
> + ev_break(loop(), EVBREAK_ALL);
> + return 0;
> +}
2. Since our unit.[ch] module supports test nesting I would start the
function from plan() + header() prologue and end it with footer() +
check_plan() epilogue. See swim.c or uri.c unit tests for example.
This will make the output almost TAP13 compliant (now it contains
several plans on one nesting level) and I hope we'll make it fully
compliant by a change in unit.[ch] and remove most of *.result files
in test/unit directory.
I propose also save the check_plan() return value in a global static
variable like in swim.c unit test to report it in the exit code of
the process.
> +int
> +main(int argc, char *argv[])
> +{
> + //say_logger_init(NULL, S_DEBUG, 0, "plain", 0);
> + memory_init();
> +
> + fiber_init(fiber_c_invoke);
> + popen_init();
> + coio_init();
> + coio_enable();
> +
> + if (!loop())
> + panic("%s", "can't init event loop");
> +
> + struct fiber *test = fiber_new("coio_stat", main_f);
3. It seems the name was copied from coio.cc and forgotten to change.
AFAIS, it is usually "main" for such tests.
> + fiber_wakeup(test);
> +
> + ev_now_update(loop());
> + ev_run(loop(), 0);
> + popen_free();
> + fiber_free();
> + memory_free();
> +
> + return 0;
> +}
4. I would return a return value from check_plan() in main_f() here,
like in swim.c unit test.
I think that it is good property to report a non-zero exit code from
a test if it fails in some way, because this way we may relax
requirements for a testing harness if we'll need it for some reason.
----
In fact I experimented around this and made the proposed changes. So
I'll share it with you in order to better show what I propose.
I also verified that after small changes in test/unit/unit.[ch] and the
changes below the result file for popen test can be deleted at all.
diff --git a/test/unit/popen.c b/test/unit/popen.c
index 7656ebe73..a40ca514c 100644
--- a/test/unit/popen.c
+++ b/test/unit/popen.c
@@ -16,6 +16,11 @@
POPEN_FLAG_SHELL | \
POPEN_FLAG_RESTORE_SIGNALS)
+/**
+ * A real return value of main_f(), see a comment in swim.c.
+ */
+static int test_result = 1;
+
static int
wait_exit(struct popen_handle *handle, int *state, int *exit_code)
{
@@ -30,7 +35,7 @@ wait_exit(struct popen_handle *handle, int *state, int *exit_code)
return 0;
}
-static int
+static void
popen_write_exit(void)
{
struct popen_handle *handle;
@@ -95,10 +100,10 @@ out_kill:
out:
footer();
- return check_plan();
+ check_plan();
}
-static int
+static void
popen_read_exit(void)
{
struct popen_handle *handle;
@@ -152,10 +157,10 @@ out_kill:
out:
footer();
- return check_plan();
+ check_plan();
}
-static int
+static void
popen_kill(void)
{
struct popen_handle *handle;
@@ -205,21 +210,24 @@ out_kill:
out:
footer();
- return check_plan();
+ check_plan();
}
static int
main_f(va_list ap)
{
- int rc = 0;
+ plan(3);
+ header();
- rc = popen_write_exit();
- if (rc == 0)
- rc = popen_read_exit();
- if (rc == 0)
- rc = popen_kill();
+ popen_write_exit();
+ popen_read_exit();
+ popen_kill();
ev_break(loop(), EVBREAK_ALL);
+
+ footer();
+ test_result = check_plan();
+
return 0;
}
@@ -237,7 +245,7 @@ main(int argc, char *argv[])
if (!loop())
panic("%s", "can't init event loop");
- struct fiber *test = fiber_new("coio_stat", main_f);
+ struct fiber *test = fiber_new("main", main_f);
fiber_wakeup(test);
ev_now_update(loop());
@@ -246,5 +254,5 @@ main(int argc, char *argv[])
fiber_free();
memory_free();
- return 0;
+ return test_result;
}
next prev parent reply other threads:[~2020-03-11 20:22 UTC|newest]
Thread overview: 23+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-03-02 20:12 [Tarantool-patches] [PATCH 0/7] popen: various fixes and a test Cyrill Gorcunov
2020-03-02 20:12 ` [Tarantool-patches] [PATCH 1/7] core/say: Export logger fd Cyrill Gorcunov
2020-03-02 20:12 ` [Tarantool-patches] [PATCH 2/7] popen: allow accessing environ variable Cyrill Gorcunov
2020-03-02 20:12 ` [Tarantool-patches] [PATCH 3/7] popen: close_inherited_fds - add support for macos/freebsd Cyrill Gorcunov
2020-03-02 20:12 ` [Tarantool-patches] [PATCH 4/7] popen: log errors if popen creation failed Cyrill Gorcunov
2020-03-02 20:12 ` [Tarantool-patches] [PATCH 5/7] popen: add logging in child process Cyrill Gorcunov
2020-03-02 20:12 ` [Tarantool-patches] [PATCH 6/7] popen: handle setsid os specifics Cyrill Gorcunov
2020-03-03 11:38 ` Alexander Turenko
2020-03-03 11:45 ` Cyrill Gorcunov
2020-03-10 15:49 ` Alexander Turenko
2020-03-10 16:36 ` Alexander Turenko
2020-03-10 16:41 ` Cyrill Gorcunov
2020-03-10 17:12 ` Alexander Turenko
2020-03-10 17:40 ` Cyrill Gorcunov
2020-03-11 7:55 ` [Tarantool-patches] [PATCH v5 6/7] popen: handle sid on macos Cyrill Gorcunov
2020-03-06 14:30 ` [Tarantool-patches] [PATCH v2 6/7] popen: handle setsid os specifics Cyrill Gorcunov
2020-03-10 8:02 ` [Tarantool-patches] [PATCH v3 6/7] popen: use ioctl on macos Cyrill Gorcunov
2020-03-02 20:12 ` [Tarantool-patches] [PATCH 7/7] test/unit: add popen test Cyrill Gorcunov
2020-03-11 20:22 ` Alexander Turenko [this message]
2020-03-12 10:38 ` [Tarantool-patches] [PATCH v5 " Cyrill Gorcunov
2020-03-12 11:58 ` [Tarantool-patches] [PATCH 0/7] popen: various fixes and a test Alexander Turenko
2020-03-12 12:18 ` Cyrill Gorcunov
2020-03-16 15:58 ` Kirill Yukhin
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20200311202213.pugz24653u7zlk4f@tkn_work_nb \
--to=alexander.turenko@tarantool.org \
--cc=gorcunov@gmail.com \
--cc=tarantool-patches@dev.tarantool.org \
--subject='Re: [Tarantool-patches] [PATCH 7/7] test/unit: add popen test' \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox