fix parse issues in https://github.com/spm/spm12

[florianschanda]

Separate out the parse issues indirectly reported in #218

• [x] In spm_eeg_review_callbacks.m, line 749

  |                 try rotate3d off;end
  |                              ^^^ error: expected end of statement, found IDENTIFIER instead

• [x] In spm_fx_hdm.m, line 76

  | try, global dt, f  = f*dt; end
  |               ^ error: expected IDENTIFIER, found COMMA instead

• [x] @file_array/Contents.m: error: expected KEYWORD, reached EOF instead

• [x] In external/fieldtrip/fileio/ft_read_data.m, line 1540

  |     case'chans_samples_trials'
  |         ^ lex error: unable to distinguish between string and transpose operation

• [x] In external/fieldtrip/preproc/ft_preproc_online_downsample_init.m, line 27

  | if factor < 1 or factor~=round(factor)
  |                  ^^^^^^ error: expected end of statement, found IDENTIFIER instead

• [x] In toolbox/DEM/spm_cost_SHC_fx.m, line 22

  | global A; X   = A.x;
  |           ^ error: expected NEWLINE, found IDENTIFIER instead

• [x] In toolbox/DEM/spm_voice_fundamental.m, line 30

  | global VOX, try VOX.formant; catch, VOX.formant = 0; end
  |           ^ error: expected IDENTIFIER, found COMMA instead

• [x] In toolbox/FieldMap/pm_invert_phasemap.m, line 111

  |          for i=1:indx(1)-1 y(i) = y(indx(1)); end
  |                            ^ error: expected end of statement, found IDENTIFIER instead

[florianschanda]

@Remi-Gau I have fixed all of these now, except for one:

In external/fieldtrip/preproc/ft_preproc_online_downsample_init.m, line 27
| if factor < 1 or factor~=round(factor)
|                  ^^^^^^ error: expected end of statement, found IDENTIFIER instead

I believe this is actually a syntax error on their side, and indeed in upstream this has been fixed: https://github.com/fieldtrip/fieldtrip/blob/master/preproc/ft_preproc_online_downsample_init.m#L30

if factor < 1 || factor~=round(factor)

I have reported this to the SPM12 authors.

[florianschanda]

@Remi-Gau This error is actually hilarious:

In external/fieldtrip/preproc/ft_preproc_online_downsample_init.m, line 27
| if factor < 1 or factor~=round(factor)
|                  ^^^^^^ error: expected end of statement, found IDENTIFIER instead

This "parses" in MATLAB, but not really. Because what it means is this:

if factor < 1
   or('factor~=round(factor)');

Command form is evil. And the language design that permits an expression after a bracket-less expression is either dumb or careless.

So I think in the original library factor was just never less than 1, otherwise they would have gotten a run-time error.