>>> Julia, Markus- can you tell me how to run coccicheck on one source file? >> >> You can choose between two methods. >> >> * Use the documented build target with appropriate parameters. >> (How many Linux software modules do consist of only a single source file?) > > This is a useless piece of information. What are "appropriate parameters". I propose to take another look at details from the discussed software documentation for the desired selection of useful settings. https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/dev-tools/coccinelle.rst?id=cd77006e01b3198c75fb7819b3d0ff89709539bb#n167 It is mentioned there that the make variable “M” can be used to apply Coccinelle to a specific directory. * Which existing directories will contain a single source file for a while? * You can create folders on demand with one source file (or equivalent symbolic links), can't you? >> * Do you get further ideas from another concrete command example? >> https://lore.kernel.org/linux-arm-kernel/5b693ee0-0cb1-7ff3-b562-bac6bcb6aae8@xxxxxx/ >> https://lkml.org/lkml/2020/6/29/357 >> >> elfring@Sonne:~/Projekte/Linux/next-patched> make J=1 C=1 CHECK='scripts/coccicheck' MODE=context COCCI=~/Projekte/Linux/next-patched/scripts/coccinelle/null/show_pointer_usage_before_null_check-20200701.cocci drivers/usb/mtu3/mtu3_gadget.o > > This is not what is intended. I am curious how corresponding intentions will be adjusted. > What is intended has never worked. I wonder about this view. Did I get inappropriate impressions from the presented test example? > The coccicheck script expects that the file name is received on its command > line, and that is not the case. I have asked the person who wrote the > code if he knows how to fix it. I am also curious how this area might evolve further. Regards, Markus
