Hi Christoph, On Thu, Dec 19, 2019 at 11:13 AM Christoph Biedl <[email protected]> wrote: > > Control: tags 913867 pending
Thank you for your prompt response. > But for the moment > disabling such an error-prone pattern is the way to go. No rush, please. Norbert may encounter the issue in other respects, but for Lintian the matter is closed until we examine COFF files again. I will observe lintian.d.o to see how many of the purported Windows executables remain. The may be other false positives, such as the .fon files in allure or angband [1]. As an aside, the Lintian team could provide you with extensive feedback on how your tool works across the archive. We could issue classification tags for all path we see in the archive and attach the output from file(1). You could construct extremely exact statistics about how many files there are of each type (via UDD), and how much space they occupy in the archive. You could also easily locate exotic file types for your test suite. Please let us know if that is of interest to you. > Version 1:5.38-1 will hit experimental very soon (hours - I'm still in > the middle of regression tests). Many important users of Lintian run on stable (or oldstable). The natural propagation of your package will work just fine. Lintian is backported, and the issue has been mitigated. > do you want this sniplet become part of > file/libmagic? I really have no opinion. Please exercise your best judgment. My inclination would be leave things as they are. Kind regards Felix Lechner [1] https://lintian.debian.org/tags/source-contains-prebuilt-windows-binary.html
