TimeSeriesProperty: Duplicate if-else branches highlighted by cppcheck

[Russell Taylor]

Cppcheck is flagging lines 495-506 and 623-626 (link below) as a duplicate if statements. And it's right, so they're clearly wrong or nonsensical in some way.

[Russell Taylor]

Re #6152. Fix if-else construct that was identical on either branch.

Changeset: c65099a99e2d33152008dd52a20e1f4540547031

[Russell Taylor]

The code for the other highlighted place first appeared in ​https://github.com/mantidproject/mantid/commit/10c600459a977fc363155af5fbe29df1b3fc39bc and at that point the branches were ever so slightly different. I need to go through the history again to spot where they became identical (missed it the first time). The reason for adding a tolerance in the 'else' case is not clear to me (though it's always been there).

[Russell Taylor]

​https://github.com/mantidproject/mantid/commit/b9d9c3e5fe1c5c2b3f8aa32a516f1fc19944e587 is where they became identical, under #1661.

[Russell Taylor]

Re #6152. Fix if-else construct that was identical on either branch.

There's also a slight change in behaviour. If a 'good' region contains more than one log value, and log values are considered to represent the left of a bin, then the end time will be the point it went 'bad' (previously it was the time of the last good value + a tolerance).

Changeset: dd2b10fe7fa766b6cf4de8f9cedb8022cec711c8

[Russell Taylor]

Re #6152. Wiki documentation update.

Changeset: e95f127e96531bc6044cb8a49f6589b454708b9a

[Russell Taylor]

Re #6152. Fix out-of-bounds error.

Highlighted by failing FilterByLogValueTest on mac, confirmed by valgrind.

Changeset: e39192cd9fc6012332eb684b52840c3a17c6f893

[Russell Taylor]

Re #6152. In test, create algorithms on stack, not heap.

Cleans valgrind output because previously the test leaked memory.

Changeset: 203b6da23cc7560277f52f2cf1422856adc51537

[Russell Taylor]

Re #6267 & Re #6152. Fix stop time for left-aligned time boundaries.

Changeset: 12e197a233c4cc35c00ed481830ef97b7354e281

[Russell Taylor]

Re #6152. Fix if-else construct that was identical on either branch.

Changeset: c65099a99e2d33152008dd52a20e1f4540547031

[Russell Taylor]

Re #6152. Fix if-else construct that was identical on either branch.

There's also a slight change in behaviour. If a 'good' region contains more than one log value, and log values are considered to represent the left of a bin, then the end time will be the point it went 'bad' (previously it was the time of the last good value + a tolerance).

Changeset: dd2b10fe7fa766b6cf4de8f9cedb8022cec711c8

[Russell Taylor]

Re #6152. Wiki documentation update.

Changeset: e95f127e96531bc6044cb8a49f6589b454708b9a

[Russell Taylor]

Re #6152. Fix out-of-bounds error.

Highlighted by failing FilterByLogValueTest on mac, confirmed by valgrind.

Changeset: e39192cd9fc6012332eb684b52840c3a17c6f893

[Russell Taylor]

Re #6152. In test, create algorithms on stack, not heap.

Cleans valgrind output because previously the test leaked memory.

Changeset: 203b6da23cc7560277f52f2cf1422856adc51537

[Russell Taylor]

Re #6267 & Re #6152. Fix stop time for left-aligned time boundaries.

Changeset: 12e197a233c4cc35c00ed481830ef97b7354e281

[Russell Taylor]

To test: Observe that the cppcheck entries complaining of duplicate branches are gone, and check the code as well. Make note of the slight changes in functionality and make yourself confident that this is correct (note the new/changed unit tests).

[Stuart Campbell]

This ticket has been transferred to github issue ​6998