[pve-devel] [PATCH pmg-api] fix #2622: include all spem levels

[Dominik Csapak]

On 3/3/20 10:02 AM, Thomas Lamprecht wrote:
> subject: s/spem/spam/
> 
> On 3/3/20 9:33 AM, Dominik Csapak wrote:
>> by using 'LIMIT 10' for the spamlevels, we only got the first
>> 10 spamlevels back from the database (whatever they were)
> 
> "whatever they were" is misleading, as this is ordered, so the
> top 10 should get returned, or?

with that i meant that it is possible to have spamlevels 0,1,10,20,30, 
etc. not all spamlevels have to be in the database, only those
for which mails were received/sent

> 
>>
>> this is only ok if there are only <= 10 different spamlevels in the database
>> but not if there are more
>>
>> the call site of this uses the combined spam count of this sub
>> for calculating the 'rest' (meaning the mails with spam level 0)
>>
>> but this is obviously wrong if not all spamlevels are counted
>> so simply return all available levels
> 
> hmm, sounds OK, this code is a classic imported from other git,
> then imported from svn case - was this always an issue or did something
> else change to make this one?

looking at the old code, it seems this issuea always existed..

> 
>>
>> Signed-off-by: Dominik Csapak <d.csapak at proxmox.com>
>> ---
>>   src/PMG/Statistic.pm | 2 +-
>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/src/PMG/Statistic.pm b/src/PMG/Statistic.pm
>> index b00fed4..6d27930 100755
>> --- a/src/PMG/Statistic.pm
>> +++ b/src/PMG/Statistic.pm
>> @@ -431,7 +431,7 @@ sub total_spam_stat {
>>   
>>       my $sth = $rdb->{dbh}->prepare("SELECT spamlevel, COUNT(spamlevel) AS count FROM CStatistic " .
>>   				   "WHERE virusinfo IS NULL and time >= ? AND time < ? AND ptime > 0 AND spamlevel > 0 " .
>> -				   "GROUP BY spamlevel ORDER BY spamlevel LIMIT 10");
>> +				   "GROUP BY spamlevel ORDER BY spamlevel");
>>       $sth->execute($from, $to);
>>   
>>       my $res = $sth->fetchall_arrayref({});
>>
>