users > bug in streamxform?
Showing 1-5 of 5 posts
Apr 18, 2013 08:04 PM | Greg Jefferis
bug in streamxform?
I can't seem to get the --inverse switch to behave in cmtk 2.2.6
jefferis:JFRC2_IS2.list (master)$ /opt/local/bin/streamxform /Volumes/JData/JPeople/Common/Neuroanatomy/BridgingRegistrations/JFRC2_IS2.list
368.250092 112.890015 107.061127
199.998448 135.002517 148.002742
jefferis:JFRC2_IS2.list (master)$ /opt/local/bin/streamxform --inverse /Volumes/JData/JPeople/Common/Neuroanatomy/BridgingRegistrations/JFRC2_IS2.list
WARNING: the following command line arguments were not used:
--inverse /Volumes/JData/JPeople/Common/Neuroanatomy/BridgingRegistrations/JFRC2_IS2.list
Unknown option: --inverse [argument #0]
Incorrect syntax or bug?
Thanks,
Greg.
jefferis:JFRC2_IS2.list (master)$ /opt/local/bin/streamxform /Volumes/JData/JPeople/Common/Neuroanatomy/BridgingRegistrations/JFRC2_IS2.list
368.250092 112.890015 107.061127
199.998448 135.002517 148.002742
jefferis:JFRC2_IS2.list (master)$ /opt/local/bin/streamxform --inverse /Volumes/JData/JPeople/Common/Neuroanatomy/BridgingRegistrations/JFRC2_IS2.list
WARNING: the following command line arguments were not used:
--inverse /Volumes/JData/JPeople/Common/Neuroanatomy/BridgingRegistrations/JFRC2_IS2.list
Unknown option: --inverse [argument #0]
Incorrect syntax or bug?
Thanks,
Greg.
Apr 18, 2013 08:04 PM | Torsten Rohlfing
RE: bug in streamxform?
Ah, if all problems were but this simple... :)
The "--inverse" parameter is not parsed the same way as other command line options (because it is, strictly speaking, part of the xform sequence).
So what you need to do is insert "--" before it -
/opt/local/bin/streamxform -- --inverse /Volumes/JData/JPeople/Common/Neuroanatomy/BridgingRegistrations/JFRC2_IS2.list
In general, "--" denotes the end of the non-positional arguments, regardless of whether anything that follows also starts with a hyphen (that's actually pretty POSIX I think).
Best,
Torsten
The "--inverse" parameter is not parsed the same way as other command line options (because it is, strictly speaking, part of the xform sequence).
So what you need to do is insert "--" before it -
/opt/local/bin/streamxform -- --inverse /Volumes/JData/JPeople/Common/Neuroanatomy/BridgingRegistrations/JFRC2_IS2.list
In general, "--" denotes the end of the non-positional arguments, regardless of whether anything that follows also starts with a hyphen (that's actually pretty POSIX I think).
Best,
Torsten
Apr 22, 2013 05:04 PM | Greg Jefferis
RE: bug in streamxform?
Hi Torsten,
Originally posted by Torsten Rohlfing:
That works fine. Thanks for the tip. On the other hand, there is no indication from --help or --help-all that this extra -- is specifically required when using the inverse switch. I've been using CMTK for years and didn't get it (POSIX or not). Would you have a suggestion to clarify the docs for that?
Best,
Greg.
Originally posted by Torsten Rohlfing:
The "--inverse" parameter is not parsed the same
way as other command line options (because it is, strictly
speaking, part of the xform sequence).
So what you need to do is insert "--" before it -
/opt/local/bin/streamxform -- --inverse /Volumes/JData/JPeople/Common/Neuroanatomy/BridgingRegistrations/JFRC2_IS2.list
In general, "--" denotes the end of the non-positional arguments, regardless of whether anything that follows also starts with a hyphen (that's actually pretty POSIX I think).
So what you need to do is insert "--" before it -
/opt/local/bin/streamxform -- --inverse /Volumes/JData/JPeople/Common/Neuroanatomy/BridgingRegistrations/JFRC2_IS2.list
In general, "--" denotes the end of the non-positional arguments, regardless of whether anything that follows also starts with a hyphen (that's actually pretty POSIX I think).
That works fine. Thanks for the tip. On the other hand, there is no indication from --help or --help-all that this extra -- is specifically required when using the inverse switch. I've been using CMTK for years and didn't get it (POSIX or not). Would you have a suggestion to clarify the docs for that?
Best,
Greg.
Apr 22, 2013 07:04 PM | Torsten Rohlfing
RE: bug in streamxform?
Well, I certainly agree it would be a good idea to have this
documented explicitly. I am just not entirely sure of the best way,
technically, to do that.
Maybe you can file a feature request and I'll see whether there's a good solution when I have some time.
Thanks.
TR
Maybe you can file a feature request and I'll see whether there's a good solution when I have some time.
Thanks.
TR
Apr 25, 2013 10:04 PM | Torsten Rohlfing
RE: bug in streamxform?
Okay, so it turns out the syntax for the xform sequence was
actually coded separately in each tool that uses this construct.
So I just added a note regarding the use of "--" before the first xform in the sequence to each of the tools.
This is in svn r/4701; not sure when I'll ever get to releasing CMTK 2.3.0, but it will be in there as well.
TR
So I just added a note regarding the use of "--" before the first xform in the sequence to each of the tools.
This is in svn r/4701; not sure when I'll ever get to releasing CMTK 2.3.0, but it will be in there as well.
TR
