Subcase/mode output fixes

[Bruno02468]

You know how, before each data block in the F06, you usually see something like: OUTPUT FOR SUBCASE 2 (or EIGENVECTOR), often followed by the deck title and, if applicable, subcase title and label? Yeah? Well, turns out that:

1. Some of it was inconsistent from the start (see: no headers at all during LINK4),
2. None of that code was not prepared for our separations of subcases and modes, and
3. There was a lot of code repetition.

This PR aims to fix all that:

• For almost all blocks (i.e. LINK 9 ones), writing the headings has been centralised into the aptly-named WRITE_SUBCASE_EIGENVEC_HEADER subroutine, increasing consistency and getting rid of reduncancies (hence the negative diff);
• For the remaining (non-LINK 9) blocks I could find, I just fixed it locally, since there were too few and it was all in the same subroutine already (EIG_SUMMARY). I couldn't use the LINK 9 one because not everything's in the same allocation state, so to speak.

That fixes the listed issues, makes our F06 far more consistent and, honestly, it also looks better.

Any downsides? Yeah, it probably breaks your F06 parser -- it broke mine!

P.S.: oh, and there's a bonus fix, I got tid of a hardcoded JTSUB in LINK9.f90 that could produce incorrect output for buckling decks with more than one static case.