bug in the timealarm library

unfortunately, it's not the only issue when using Dow:

1st the 'value' of the alarm is calculated using this function
(7 + DOW - dayOfWeek(now())) %7) * SECS_PER_DAY + AlarmHMS(H,M,S)
this value is only valid for the next alarm to be set and works well with alarmonce for exemple
if you are on monday morning and set an alarm for (and each) tuesday 12:00AM, the 'value' is roughly 1 day and a half.
This is OK for the first alarm but the next one is also set 1 day and a half after midnight of the first ie: for wednesday (which is not the goal)

Concerning the id returned by the alarm.alarmrepeat, it works very well and it's very usefull from outside the alarmhandling, but sometimes you want the alarmhandling funtion to do different things base on which alarmid you are handling, and this is not doable, (or I missed something)

best regards

Viknet