stable 5.1 contains the fix, which is the addition of -w to the grep on line 5428. that particular fix works for the example names given (100001 as the basename, 100001_S0 and 100001_S1 in the base-tps file).
can someone give me the basename and base-tps contents so that i can replicate the problem?
n.
On Wed, 2011-08-10 at 14:45 -0400, Martin Reuter wrote:
Hi Nick, can you look into this? I think we fixed this bug in 5.0 or is it something different? Is it still not fixed in 5.1?
Thanks, Martin
On Wed, 2011-08-10 at 11:19 -0700, Irwin, William wrote:
Hi-
Note, this bug still exists in 5.1. Below is my exchange with Martin and Nick a several months ago.
-Wil
|-----Original Message----- |From: Nick Schmansky [mailto:nicks@nmr.mgh.harvard.edu] |Sent: Wednesday, May 25, 2011 3:52 PM |To: Irwin, William; Martin Reuter |Subject: RE: 5.1.0 Longitudial syntax question | |Martin, | |attached is his log. it seems like it will fail if the basename |partially matches a timepoint name. | | | |Wil, there are two possible workarounds: | |1) edit line 5421 of recon-all to change: |if ( ! $status) then | |to | |if ( 0 ) then | | |so that this block is never run | |or | |2) change the name of the basename so that its not a sub-match, for |instance change to xxxxxxx_base | | |we'll have to figure out how this happened. | |n.
|-----Original Message----- |From: Martin Reuter [mailto:mreuter@nmr.mgh.harvard.edu] |Sent: Wednesday, August 10, 2011 10:10 AM |To: Jeff Sadino |Cc: freesurfer@nmr.mgh.harvard.edu |Subject: Re: [Freesurfer] Longitudinal Subject/Base name errors | |Hi Jeff, | |this was a known bug in 5.0 and there are fixes on the web (release notes), |just download the recon all from there (and did you grab the robust template |binary)?. | |http://surfer.nmr.mgh.harvard.edu/fswiki/ReleaseNotes | |Best, Martin | |On Mon, 2011-08-08 at 17:43 -1000, Jeff Sadino wrote: |> Hello Experts, |> |> |> I am using fs 5.0.0. My naming convention is similar to 100001_S01 |> and 100001_S02 for subject, and then 100001 for the base. Now when I |> process the runs longitudinally, I get: ERROR: longitudinal base ID |> cannot be the same as a timepoint. I believe this is because of line |> 4845 in recon-all: grep $longbaseid |> ${longbasedir}/${BaseSubjsListFname} >& /dev/null. I can just as |> easily cut out this part of the code, or I could manipulate the |> base-tps file, or rename the base folder. But I do not know if there |> are other spots in the code that rely on the original naming |> conventions. Can anyone recommend one approach over the other? |> |> |> Thank you very much, |> Jeff Sadino |> _______________________________________________ |> Freesurfer mailing list |> Freesurfer@nmr.mgh.harvard.edu |> https://mail.nmr.mgh.harvard.edu/mailman/listinfo/freesurfer |