Back to the main page.
Bug 2610 - ft_timelockanalysis fails for component input data
| Status | CLOSED FIXED |
| Reported | 2014-06-12 14:08:00 +0200 |
| Modified | 2019-08-10 12:28:52 +0200 |
| Product: | FieldTrip |
| Component: | core |
| Version: | unspecified |
| Hardware: | PC |
| Operating System: | Windows |
| Importance: | P5 major |
| Assigned to: | Eelke Spaak |
| URL: | |
| Tags: | |
| Depends on: | |
| Blocks: | 2518 |
| See also: |
Eelke Spaak - 2014-06-12 14:08:19 +0200
Eelke Spaak - 2014-06-12 14:15:12 +0200
Well, that was an easy one. ft_timelockanalysis still used keyval-syntax with ft_selectdata, and the bug was (and remains) in ft_selectdata_old. I now changed line 125: data = ft_selectdata(data, 'rpt', cfg.trials); to: tmpcfg = []; tmpcfg.trials = cfg.trials; data = ft_selectdata(tmpcfg, data); [cfg, data] = rollback_provenance(cfg, data); To be fair, I find the old syntax much easier on the eye, but I'm afraid that the new one is how it's going to be :) bash-4.1$ svn commit ft_timelockanalysis.m Sending ft_timelockanalysis.m Transmitting file data . Committed revision 9617.
Robert Oostenveld - 2014-06-12 19:32:20 +0200
(In reply to Eelke Spaak from comment #1) the new one provides the updated cfg, which is important for provenance. If it were a low-level function, the old one would be ok and could be extended to also provide output. But it being a high-level function imposes the end-user function call interface format with cfg as first input on it. We don't want to branch of to different functions having different user interfaces.
Eelke Spaak - 2014-06-12 22:00:02 +0200
(In reply to Robert Oostenveld from comment #2) Yes, sorry, I totally see the rationale for the new one and agree that it's better, my remark was just an aside.