Rather than chancing calling strchr when @err_noinfo == NULL or
calling virErrorTestMsgFormatInfoOne when @err_info == NULL, both
of which would fail, let's just return -1 immediately when we
encounter an error.

Signed-off-by: John Ferlan <jferlan@redhat.com>
---
 tests/virerrortest.c | 11 +++++------
 1 file changed, 5 insertions(+), 6 deletions(-)

diff --git a/tests/virerrortest.c b/tests/virerrortest.c
index eebf259343..48b4848b44 100644
--- a/tests/virerrortest.c
+++ b/tests/virerrortest.c
@@ -58,7 +58,6 @@ virErrorTestMsgs(const void *opaque ATTRIBUTE_UNUSED)
     const char *err_noinfo;
     const char *err_info;
     size_t i;
-    int ret = 0;
 
     for (i = 1; i < VIR_ERR_NUMBER_LAST; i++) {
         err_noinfo = virErrorMsg(i, NULL);
@@ -66,25 +65,25 @@ virErrorTestMsgs(const void *opaque ATTRIBUTE_UNUSED)
 
         if (!err_noinfo) {
             VIR_TEST_VERBOSE("\nmissing string without info for error id %zu\n", i);
-            ret = -1;
+            return -1;
         }
 
         if (!err_info) {
             VIR_TEST_VERBOSE("\nmissing string with info for error id %zu\n", i);
-            ret = -1;
+            return -1;
         }
 
         if (strchr(err_noinfo, '%')) {
             VIR_TEST_VERBOSE("\nerror message id %zu contains formatting characters: '%s'\n",
                              i, err_noinfo);
-            ret = -1;
+            return -1;
         }
 
         if (virErrorTestMsgFormatInfoOne(err_info) < 0)
-            ret = -1;
+            return -1;
     }
 
-    return ret;
+    return 0;
 }
 
 
-- 
2.17.2

--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list

On Mon, Dec 17, 2018 at 09:55:46 -0500, John Ferlan wrote:
> Rather than chancing calling strchr when @err_noinfo == NULL or
> calling virErrorTestMsgFormatInfoOne when @err_info == NULL, both
> of which would fail, let's just return -1 immediately when we
> encounter an error.
> 
> Signed-off-by: John Ferlan <jferlan@redhat.com>
> ---
>  tests/virerrortest.c | 11 +++++------
>  1 file changed, 5 insertions(+), 6 deletions(-)

This would make the test hard to diagnose as it spits out one failure at
time. I prefer you add the checks explicitly.
--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list

On 12/18/18 5:34 AM, Peter Krempa wrote:
> On Mon, Dec 17, 2018 at 09:55:46 -0500, John Ferlan wrote:
>> Rather than chancing calling strchr when @err_noinfo == NULL or
>> calling virErrorTestMsgFormatInfoOne when @err_info == NULL, both
>> of which would fail, let's just return -1 immediately when we
>> encounter an error.
>>
>> Signed-off-by: John Ferlan <jferlan@redhat.com>
>> ---
>>  tests/virerrortest.c | 11 +++++------
>>  1 file changed, 5 insertions(+), 6 deletions(-)
> 
> This would make the test hard to diagnose as it spits out one failure at
> time. I prefer you add the checks explicitly.
> 

To avoid wasting steps/time w/ multiple iterations, you would rather
have the "err_noinfo &&" be a gate to strchr and "err_info &&" be a gate
to virErrorTestMsgFormatInfoOne?

So that the only diffs are:

-        if (strchr(err_noinfo, '%')) {
+        if (err_noinfo && strchr(err_noinfo, '%')) {

and

-        if (virErrorTestMsgFormatInfoOne(err_info) < 0)
+        if (err_info && virErrorTestMsgFormatInfoOne(err_info) < 0)


Tks -

John

--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list

On Tue, Dec 18, 2018 at 08:21:08 -0500, John Ferlan wrote:
> 
> 
> On 12/18/18 5:34 AM, Peter Krempa wrote:
> > On Mon, Dec 17, 2018 at 09:55:46 -0500, John Ferlan wrote:
> >> Rather than chancing calling strchr when @err_noinfo == NULL or
> >> calling virErrorTestMsgFormatInfoOne when @err_info == NULL, both
> >> of which would fail, let's just return -1 immediately when we
> >> encounter an error.
> >>
> >> Signed-off-by: John Ferlan <jferlan@redhat.com>
> >> ---
> >>  tests/virerrortest.c | 11 +++++------
> >>  1 file changed, 5 insertions(+), 6 deletions(-)
> > 
> > This would make the test hard to diagnose as it spits out one failure at
> > time. I prefer you add the checks explicitly.
> > 
> 
> To avoid wasting steps/time w/ multiple iterations, you would rather
> have the "err_noinfo &&" be a gate to strchr and "err_info &&" be a gate
> to virErrorTestMsgFormatInfoOne?

Yes exactly.

> 
> So that the only diffs are:
> 
> -        if (strchr(err_noinfo, '%')) {
> +        if (err_noinfo && strchr(err_noinfo, '%')) {
> 
> and
> 
> -        if (virErrorTestMsgFormatInfoOne(err_info) < 0)
> +        if (err_info && virErrorTestMsgFormatInfoOne(err_info) < 0)
> 
> 
> Tks -

ACK to that version
--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list