Stefan Kangas writes: > Lars Ingebrigtsen writes: > >> Stefan, was there any particular reason to switch from sha1 to md5 in >> image-dired? > > Mostly that it was marginally faster and I didn't expect any breakage. > I think we could either switch back, or make sure that we use the MD5 > version of the file name if it exists and otherwise use SHA1. If you do that you will have one call of MD5 and one of SHA1 on the same file, which defeat the purpose of using SHA1 which is presumably faster. -- Thierry