License Browser Scanner Count is off

[bobgob]

It looks like the License Browser "Scanner Count" is summing the results from multiple versions of the license scanner. For example, if both version 1 and 2 of nomos produced 1 Public-domain each, then the Scanner Count would show 2 instead of 1 for the last scan. I'm only speculating here.

Sorry for the HP internal URL:
http://bobg.fc.hp.com/fossology/?mod=license&upload=257&item=21270940
I've attached a screenshot in case this helps:

[bobgob]

I just tested and the issue is not fixed. I get the same results as before.

[bobgob]

sorry. My mistake. I was looking at a cached page. Closing.

[bobgob]

Reopening. The Scanner Count did change but it does not match the number of files displayed in "List Files for License"

[wuan]

So i think we have a similar problem on the license_list_files page.

<total_count> files found (<unique_count> unique) with license <license_shortname>

Total count is again containing results of multiple agent versions.

[bobgob]

That message was changed yesterday. Do a git pull.

[bobgob]

In case there a misunderstanding of what the License Browser should be displaying, let me go over that. By default, only the latest results should be displayed. For example, say you have Monk 2.5.0, Monk 2.6.1, Nomos 2.5.0 and Nomos 2.6.1 scans. By default, the results for Monk 2.6.1 and Nomos 2.6.1 should be displayed. This should not include the 2.5.0 results. The pull down lets you recall older results.

[wuan]

The license_list_files page shows the unique file count in the beginning and the number of files listed is identical to the number of files with the particular license.
As I assume that the number of files (not unique files) is relevant here, I will try to fix this as well.

[wuan]

fixing the license_list_files page would be rather expensive, please file another issue if this is required.

[bobgob]

My tests show that List Files for License and the License Browser Scanner count now agree with each other. Thank you!

[yaobinshi]

for this commit 3128584

I have a question:
one file has multiple BSD-3-Clause license text, so monk will report multiple BSD-3-Clause for this file,
the 'Scanner Count' will count multiple times, so the file count may less than the 'Scanner Count'.
from my perspective, my check in is better on the notice description, or others?

[wuan]

The count displayed at the beginning of the license file list is the number of unique files having at least one appearance of the license. Whereas the file listing below counts unique files multiple times. Files with identical content are grouped together but counted individually.

If files with identical content exist in this view, the users may be irritated due to the different numbers.

[yaobinshi]

Ok, why I asked the question above is because, I have one archive, 'Monk Scanner Count' is 5 BSD-3-Clause on license browser page, however the file count is 2 on List Files for License page, as one file is returned 4 BSD-3-Clause by monk, another file is return a BSD-3-Clause by monk.

[wuan]

The scanner count should count multiple licenses per file individually. Would this be the expected behaviour here?
The file list does not repeat files if the same license is found at multiple location. So we can expect differences in count.

[yaobinshi]

Will talk with others to get more opinions, let us keep it. thanks.

[mcjaeger]

Andreas' (wuan) sounds reasonable now, and since there has not been an update since 12 months, I am closing