On Fri, 23 Sep 2022 22:02:25 +0200 Jochen Sprickerhof wrote: > Hi Francesco,
Hello Jochen, thanks for your reply. > > * Francesco Poli (wintermute) <[email protected]> [2022-08-13 19:08]: > >I applied the following patch to the Makefile: > > > > -CEX := po/ > > +CEX := po/.*\.mo > > This includes the po/*.po in the file list which where excluded, > previously. Yes, that's intentional: the goal was to let decopy scan po/*.po files too, in order to automatically pick updates (for instance, when their copyright years change). Is this the wrong way to achieve this result? > > >Honestly, I expected that decopy would abide by this new base copyright > >file and regenerate an identical debian/copyright file. > > It works just fine if you do a debclean before generating it. I checked and I confirm that it indeed works, after a debclean. However, I added all the generated files to the --exclude argument, precisely in order to avoid the need for a debclean. I wanted decopy to act as if the generated files were absent. Apparently, this is not what's happening. Why? Is the syntax I used for the --exclude option incorrect? > > >Well, this looks (almost) technically correct, > > Yes and I think that's all you can ask for for such a tool. To me decopy > is a good first step to create a d/copyright file but it always need > some human eyes. Do you mean that you only use decopy for a first rough draft of the debian/copyright file and then you modify it by hand? That's what I thought to do myself. But then, what do you do, when the source tree changes and the debian/copyright file has to be updated? Do you update it by hand? Or do you re-run decopy with the outdated debian/copyright file as a base for the processing? My intention was to follow the latter strategy, hence my Makefile target 'debian-copyright'... Did I misunderstand how decopy should be used? > > >Firstly, decopy decided that the license for po/* applies to * !!! > > I don't know all the decopy code but I think it has some heuristics for > which copyright block should be the top one. Well, I suspect the heuristics should be improved a bit... I think that, when a preexisting debian/copyright file is used as a base for processing, decopy should not change which copyright block is used as the top one, unless there is a really strong reason to do so. Do you agree? > > >Then, I see that apt-listbugs.1 is listed among the special cases. > >But that file is generated, and I thought it was excluded through > >the --exclude option of decopy. Apparently, it is not! Why? > > I guess exclude is only for the content not for the file list. Maybe > that's what meant in #997814.. Wait a second: do you mean that the --exclude option only prevents decopy from looking inside the excluded files, but does not make decopy ignore their existence?!? If this is confirmed, I think a new option should be added (we could perhaps call it "--ignore") that makes decopy act as if the ignored files were absent. > > >Finally, I fail to understand where the "2002-2020, Masato Taruishi" > >additional Copyright come from. > >It seems to me that every copyright notice referring to Masato Taruishi > >in the source tree is either followed by his e-mail address or by "et al.". > >So where does this additional Copyright come from? > > decopy strips the et al.: > > https://sources.debian.org/src/decopy/0.2.4.7-0.2/decopy/res.py/#L509 > > So with the Makefile change above the .po files are searched again and > the entry is added. Ah, I see, thanks for the explanation. But why does decopy strip "et al."? And what's the correct way to specify that a copyright is owned by one main owner plus many other co-owners? I mean, the correct way for decopy. Please let me know what you think about my doubts. Thanks for your time! -- http://www.inventati.org/frx/ There's not a second to spare! To the laboratory! ..................................................... Francesco Poli . GnuPG key fpr == CA01 1147 9CD2 EFDF FB82 3925 3E1C 27E1 1F69 BFFE
pgpPDXa_hrCOY.pgp
Description: PGP signature
