Hi all,
You must know that we parse norally all source code with Coverity Scan service and fix step by step the issues detected by the static analyzer. Since January, git/master cannot be processed by Coverity. The build is complete, but the report is never commited and is send in somewhere in /dev/nulll (:=)))... Coverity Scan service was acquired by a new company in 2018, and i suspect a side effect to commit report to the remote server. I contacted the Coverity team, who respond that investiguation are under progress, please wait. So, i finally try to found a new solution to parse week by week all source code to review by another static analyzer. I tried Clang one, and reports are really excellent. I written a script in project/reports/, but it's not yet perfect. The first report that clang generate is really interesting. I shared the files (web pages) in this archive : Please take a look and feel free to apply patches is necessary. I will try to finalize the script while this week end to be able to run the analyzer locally. Best Gilles Caulier |
Hi all, My Clang static analyzer script is working well now. It publish in digiKam.org static area the report automatically The contents is currently and older one. I will run again the script soon to update the contents. Best Gilles 2018-05-04 17:48 GMT+02:00 Gilles Caulier <[hidden email]>:
|
The url has a little bit changed : We have now clang and cppcheck reports posted to digiKam.org... Gilles 2018-05-05 14:19 GMT+02:00 Gilles Caulier <[hidden email]>:
|
Hi all, Now 3 reports are published by my BASH scripts located in project/reports : All use the same configuration about to ignore directories to parse while scanning : ".krazy" file on root directory. This mailing list is CC automatically when a new report is online with the right url to take a look.- clang : this analyzer do not have an option to ignore directory. So i parse all, and i filter the output HTML file before to publish. The task is hard to complete, especially to update the analysis statistics accordingly. Currently it's not the case, and statistics include dropped items. I must code more and more with BASH to achieve a complete filtering. - krazy : i only scan with "extra" checks not published to EBN (http://ebn.kde.org/krazy/reports/extragear/graphics/digikam/) For this last one, i passed 3 weeks and 300 commits to fix all reports. The extra checks still under development and can generate false errors. Take a care... - cppcheck : very verbose, probably about 'style' code analyze. It just an option to tune if style are not suitable. Anyway, some reports are interesting to investigate. To conclude : now we have suitable reports to detect wrong coding done by contributors as students or patches. Best Gilles 2018-05-06 15:39 GMT+02:00 Gilles Caulier <[hidden email]>:
|
In reply to this post by Gilles Caulier-4
Hi Gilles,
Thanks for setting these checks up! I'd rather not have automatical emails sent to the digikam-devel mailing list for the same reason it is now separate from digikam-bugzilla: Emails in this list should be actual discussions by humans, not generated mail. Otherwise the important email may get overlooked. So I propose to send the updates to digikam-bugzilla or not send them at all. What do you think? Cheers, Simon On 06/05/18 15:39, Gilles Caulier
wrote:
signature.asc (879 bytes) Download Attachment |
You has right. Do you mean to create bugzilla bug reports accordingly with clang/cppcheck reports ? this will be an hot task todo... Gilles 2018-05-11 11:08 GMT+02:00 Simon Frei <[hidden email]>:
|
No, I think that's too much work for not much gain. It was just an
idea to "misuse" the bugzilla mailing list for these emails too. But
my personal opinion is, that the static check commit emails can be
disabled fully, because they don't have actual information about the
new report. So you need to look at the actual report anyway.
On 11/05/18 11:38, Gilles Caulier
wrote:
|
i changed to email notification which use now [hidden email] The idea is to CC later the students when a new report is available. But i'm not sure if the notification work as expected as this email is prbably used internally from bugzilla. Gilles 2018-05-11 12:04 GMT+02:00 Simon Frei <[hidden email]>:
|
Free forum by Nabble | Edit this page |