28/02/2019 23:09, Michael Santana Francisco: > On 2/28/19 6:21 AM, Thomas Monjalon wrote: > > 14/02/2019 20:35, Michael Santana: > >> Enable codespell by default. > >> codespell is a feature by checkpatch.pl that > >> checks for common spelling mistakes in patches. > > What is the difference between codespell and spelling.txt included > > with checkpatch? > > Is it just a different dictionary? > codespell has a larger dictionary, about 15000 word fixes whereas > spelling.txt has about 1000. > That's really the only big difference
OK > >> This feature is disabled by default. To enable it one must add > >> the '--codespell' flag to the $options variable in > >> checkpatches.sh. > > We need also to specify the dictionary path if not in > > /usr/share/codespell/dictionary.txt > > In my case, it is in > > /usr/lib/python3.7/site-packages/codespell_lib/data/dictionary.txt > > > >> With this change codespell is enabled by default. > > It seems it is not enabled by default, > > because we need DPDK_CHECKPATCH_CODESPELL=enable > V2 sets DPDK_CHECKPATCH_CODESPELL=enable at the beginning of > checkpatches, right before reading in the config files. > If DPDK_CHECKPATCH_CODESPELL is set in one of the config files it > overwrites the enabled by default. > This way a user can disable it via a config file OK I missed it. > >> The user can decide to turn off codespell from a one of the config > >> files read by checkpatches.sh. > > [...] > >> # override default Linux options > >> options="--no-tree" > >> +if [ "$DPDK_CHECKPATCH_CODESPELL" == "enable" ]; then > > What about allowing either "enable" or a path? > > If it is a path (have some slash), then we can add --codespellfile option. > I like your thinking. We can use `if [ -f <file> ]` to see if the path > given is an existing file. > so, if DPDK_CHECKPATCH_CODESPELL is set to enable, then enable it with > default path (the way it is right now) > if DPDK_CHECKPATCH_CODESPELL is set to a valid path to a file then > enable codespell and set --codespellfile to said file > otherwise if it's not set to enable or set to a valid path file, then > assume it's disabled. > Missed anything? Well described. Would be nice to see in v3. Thanks
