On Mon, 2020-09-28 at 19:11 +0200, Julia Lawall wrote: > > On Mon, 28 Sep 2020, Joe Perches wrote: > > > On Sun, 2020-09-27 at 21:12 +0200, Julia Lawall wrote: > > > Replace commas with semicolons. What is done is essentially described by > > > the following Coccinelle semantic patch (http://coccinelle.lip6.fr/): > > > > Hi Julia. > > > > How did you decide on this patch subject header line? > > > > It's certainly reasonable, but not what I would expect > > from any automation. > > Most of the patches on this file touch other files as well. There are > four that only touch this file. One has no subsystem. One has dma as the > subsystem. Two have pata_icside, so that was chosen. So did you select the patch subject without any automation or did you create a tool that looks only at commits for individual files? > > $ git log --no-merges --format='%s' drivers/ata/pata_icside.c | \ > > cut -f1 -d":" | sort | uniq -c | sort -rn > > 10 libata > > 5 libata-sff > > 2 [ARM] pata_icside > > 2 [ARM] ecard > > 2 [ARM] dma Yeah... Something like the script below and some examples of its use? $ cat find_best_header.sh #!/bin/bash # Show possible patch subject prefixes for a file in git # use commits that modify only the single file argument and # emit up to the 5 most common commit prefixes git log --no-merges --format='%h' -- $1 | while read commit ; do if [[ $(git log --format='%h' --name-only -1 $commit | wc -l) == 3 ]] ; then git log --format='%s' -1 $commit fi done | cut -f1 -d":" | cut -f1-3 -d" " | sort | uniq -c | sort -rn | head -5 $ ./find_best_header.sh drivers/ata/pata_icside.c 2 [ARM] pata_icside 1 Fix pata_icside build 1 [ARM] dma $ ./find_best_header.sh kernel/fork.c 36 fork 14 kernel/fork.c 6 mm 4 pidns 4 [PATCH] unshare system $ ./find_best_header.sh kernel/sched/core.c 150 sched 110 sched/core 12 sched/uclamp 12 sched/debug 10 sched/deadline $ ./find_best_header.sh drivers/iio/pressure/mpl115.c 3 iio
