On 20/3/21 6:51 pm, Lukas Bulwahn wrote: > On Sat, Mar 20, 2021 at 1:45 PM Aditya Srivastava <yashsri421@xxxxxxxxx> wrote: >> >> On 20/3/21 12:23 pm, Aditya wrote: >>> On 18/3/21 11:48 pm, Jonathan Corbet wrote: >>>> Lukas Bulwahn <lukas.bulwahn@xxxxxxxxx> writes: >>>> >>>>> Yeah, and as this line-counting is really just a poor man's >>>>> heuristics, we might just be better to really turn this heuristics >>>>> into a dedicated cleanup warning script, then we can check for more >>>>> indicators, such as "does it contain the word Copyright" somewhere in >>>>> the kernel-doc comment, which tells us even more that this is not a >>>>> kernel-doc as we would expect it. >>>> >>>> I really don't think we need that kind of heuristic. The format of >>>> kerneldoc comments is fairly rigid; it shouldn't be too hard to pick out >>>> the /** comments that don't fit that format, right? Am I missing >>>> something there? >>>> >>>> Thanks, >>>> >>>> jon >>>> >> >> Hi Lukas and Jon! >> I have a question, should I clean up the files with '/**' like >> comments in only header lines? Or as we are planning for making it >> generic, for other lines as well? >> > > Aditya, of course, if you can detect and come across some unintended > '/**' comments in some files, clean them in the same go (as you did > with ecryptfs). > > I am just worried that if you extend it to the fully generic case, > that the list of cases simply explodes: showing many 1,000 cases > across various 1,000 files that need to be cleaned up, and such > clean-up work is just too much to get done by yourself. > > The current list limited to comments in header lines seems to be a set > of patches that you can probably get done. > Sounds good, Lukas. Thanks Aditya
