Hi
> On 24 Jun 2025, at 13:36, Dan Carpenter <[email protected]> wrote:
>
> I'm going to respond to you email out of order based on the easiest
> thing first.
First things first: thanks for doing that.
>
> On Tue, Jun 24, 2025 at 10:16:54AM +0000, Miguel Luis wrote:
>>> The SQL
>>> is mostly pulled into smatch_warns.txt.sql and smatch_warns.txt.call_info.
>>
>> Got that but I’m still a bit lost here, so in my script >
>> ‘create_db.sh’ is using smatch_warns.txt.sql solely, should
>> ‘smatch_warns.txt.call_info’ be passed to ‘create_db.sh’ too or
>> how does that information gets into the db?
>>
>
> The script just expects the file to be named that way. Whatever
> the "smatch_warns.txt" is called plus ".caller_info".
>
> ${bin_dir}/fill_db_caller_info.pl "$PROJ" ${info_file}.caller_info $db_file
>
Got it! Thanks to that I’ve fixed my script to pass ${info_file} which in my
case is ’smatch_warns.txt’ and not the .sql one.
>
> For the return states in smatch_warns.txt.sql it doesn't really
> matter but it turns out that this does matter for caller_info. The
> reload_partial.sh script assumes it's all in one file and then
> separates the caller_info out into a tmp file for processing.
>
I see. At first, entries are being deleted as those will be filled with new
data via fill_db_{caller_info|sql}.pl.
Lastly but definitely not the least the fixup script will filter and insert
already identified portions of what matter to be kept in the db, for instance,
in order to reduce the cause of false positives.
> regards,
> dan carpenter
>
>
Regards,
Miguel
