Reimplement missing commands

[Juergen Reuter]

One that is still missing is:

1. printf
2. show
3. scan
4. include
5. scale
6. ren_scale
7. fac_scale
8. tolerance
9. quit
10. exec
11. slha (separate ticket #519)
12. weight
13. selection
14. reweight
15. clear
16. open_out/close_out (separate tickets #571, #542)

[Juergen Reuter]

[1. printf done in r5090] [2. show done in r5048 ... and finally in r5107] [3. scan done in r5085] [4. include was already implemented] [5. scale done in r4977] [6. ren_scale done in r4977] [7. fac_scale done in r4977] [8. tolerance done in r5064] [9. quit done in r4968] [10. exec done in r4992] [11. slha done in r5046 (separate ticket #519)] [12. weight done in r4977] [13. selection done in r4982] [14. reweight done in r4978] [15. clear done in r5056] [16. open_out/close_out done in r5090 (separate tickets #571, #542)] [17. me_test done in r5113] [18. beam_polarization done in r5161 (separate ticket #559)]

[kilian]

I see the cause of the problem, but don't yet have a solution.

[I admit that my variable handling module is _really_ broken (partly due to the absence of proper type extension). But I won't open another bottle by rewriting it now ;) ... rather try to find a workaround, again.]

[kilian]

tolerance parameter (r5064): this was actually ok, but I unified the comparison operators in Sindarin according to the discussion on the phone.

Comparison (==, <>, <, >, <=, >=) now takes into account the tolerance parameter if a real number is involved in the comparison. By default, the tolerance is zero, of course. The operators (==~ and <>~) have disappeared, they are redundant.

[Juergen Reuter]

Also show before had screen output. This vanished now. Why?

[Juergen Reuter]

Ok, it seems magically both is present. I don't know why. Adding the missing cases (as no one reacted).

[Juergen Reuter]

WK, can you take over again, please? (you know what I want: as many of the things in vars.sin should work again)

[Juergen Reuter]

Again, my thinking here was to cover as much as possible from the vars.sin test by the implemented commands. So, WK: 1) if there is stuff in vars.sin you think should never be reenabled, please remove it from

that SINDARIN file

2) if there is stuff that will be disabled for now but be covered in the planned year 2063

revision of SINDARIN () then comment it out (and make a comment in vars.sin)

Thanks again and sorry for the weekend disturbances.

[kilian]

status for command scan: done, but fighting a segfault that occurs with gfortran only (arghh...)

[Juergen Reuter]

WK, please, please, PLEASE, let me know whether you will do the show(real), show(int), show(intrinsic), show(logical) or not?

[kilian]

Hmm - I'm still not convinced. Could you suggest 'show' variants that are *actually* useful?

In any case, I would do it (easy), but I prefer to do the other missing commands first, and collect wishes for 'show' in the meantime.

[Juergen Reuter]

Show stopper: show done finally in r5107 (all that JRR wanted to have into). Actually, show (alias XXX) was broken in cmd_show_compile. Fixed that, and added a new test show_4 that tests for it, and the other features.

[Juergen Reuter]

Actually, in r5108 I added logical and string test cases. WK, please reassign now for the missing tests.

[Juergen Reuter]

Ok, JRR will fix that issue (making the test intrinsic). Give me 10-15 mins. Sorry.

[kilian]

r5113: removed the timing me_test command (matrix_element_test), superseded by providing n_calls_test to the integrate command. (If phs_only is set, only the test will be run, after phase-space configuration.)

[kilian]

As of r5161, all commands have been either deleted, replaced, or resurrected. Any further additions would count as new features.