On Mon, 22 Jul 2019, Marion & Christophe JAILLET wrote: > Hi, > > Attached is a WIP script I've just written which tries to spot typos in a > file, based on the filename itself. > > Yesterday, I've posted some finding done with this script. > Today, I share if someone finds it useful, want to improve it or just want to > take the idea. > > As, I'm not a bash guru, it is neither optimal, nor well written. > But it seems to work as I expect. > > > The name of a file can be a good source of information to spot typo in the > code itself. This can help spot typo in comments or strings, but also wrongly > named functions or constant. > 3 checks are implemented. They can be disabled individually. > > The filename should sometime be tweaked a bit to only take the part before or > after a '-' or a '_'. (some regex patterns are in the script for that, just > comment/un-comment) > > The 2 last checks generate lot of false positives. > It can find some few things, but honestly, the semantic should be improved. > > > Just in case s.o. find it useful and want to use it to clean-up a few things. It seems like a nice idea. Based on another patch you sent, perhaps something could be done with non-English words in general. If there are a few occurrences of XYZ in a file, but only one occurrence of XXZ, it might be worth highlighting XXZ as a possible typo. (I don't know what string length and string distance would give the best results). julia
