The old code was creating a datetime from maybeDate + maybeTime
but that meant that it could be creating a datetime that never
existed if that datetime falls in the time that jumps to the
future when doing dst (i.e 2->3 am)
Details
Details
Added new test that fails without my patch and works with it
Diff Detail
Diff Detail
- Repository
- R180 PIM: KMime
- Lint
Automatic diff as part of commit; lint not applicable. - Unit
Automatic diff as part of commit; unit tests not applicable.
Comment Actions
+1, but as discussed on IRC, the test is unfortunately local-timezone-dependent (the old code only fails if 31 Mar 2013 02:29 is a non-existent timestamp in the local timezone). Does anyone have an idea to avoid that? The test without the fix should fail regardless of local environment...
Comment Actions
Looks good to me. Regarding the autotest timezone, that's a common issue in kcalcore tests as well, many of those force the timezone via something like qputenv("TZ", "Europe/Zurich");.
Comment Actions
guys, i've pushed this by mistake.
Was running the script to update versions of repos to 17.12.2 and it ended up commiting this too.
Do you want me to revert it or is it ok?