https://bugs.koha-community.org/bugzilla3/show_bug.cgi?id=20936 --- Comment #42 from Martin Renvoize <martin.renvoize@ptfs-europe.com> --- OK.. this issue here is that the js code in bug 24455 does not distinguish between the perfectly valid 'full-date' and 'date-time' resolutions of RFC3339. Both are valid formats, but one includes a time offset whilst the other does not. The 'full-date' option does not include time resolution and as such also does not include an offset. When moment parses this, it 'upgrades' the internal representation from full-date (2020-03-25) to date-time (2020-03-25T00:00:00+00:00) internally.. When you then apply a timezone to that (say +0300 for Argentina I believe), we end up with an internal date of 2020-03-24T21:00:00-03:00.. spot the day change.. formating this back down to just 'date' will loose the day. Hope that makes sense.. so.. what we need to do is detect if we're parsing a date-time vs a full-date and only apply timezone maths to the former -- You are receiving this mail because: You are watching all bug changes.