(
Jump to the original submission)
Fri 25 Mar 2016 03:43:03 PM UTC, comment #6:
I pushed it here:
http://hg.savannah.gnu.org/hgweb/octave/rev/936ab0fca2f7
Closing report
|
Tue 22 Mar 2016 09:20:25 PM UTC, comment #5:
Tested on Linux and Windows 7, works fine.
Thanks for fixing, Lachlan.
Will you push & close this report, please?
|
Tue 22 Mar 2016 09:48:53 AM UTC, comment #4:
Here is the rebased patch. Thanks for letting me know where things had moved to.
(file #36723)
|
Tue 22 Mar 2016 12:26:20 AM UTC, comment #3:
This patch fails to apply to the current textscan, it's been integrated into libinterp/corefcn/{file-io,oct-stream}.{cc,h} now. If you have time could you rebase your change onto the current default branch?
|
Sun 20 Mar 2016 11:42:22 PM UTC, comment #2:
Thanks, Philip and Mike.
The attached patch causes a line of emptyVal (NaN by default) if the final line didn't have any valid conversions but did match a literal before the first conversion. I assume that is the required behaviour.
(file #36703)
|
Sat 19 Mar 2016 07:06:31 AM UTC, comment #1:
Just checked after latest batch of textscan cleanups and confirm this is still the same result.
|
Fri 18 Mar 2016 07:04:33 PM UTC, original submission:
With new textscan.cc in dev:
Expected is {NaN, NaN}.
Adding any digit goes some way in the right direction:
Matlab r2016a prerelease gives:
+verbatim-
>> format compact
>> U = textscan (['Empty' char(10)], 'Empty%f %f')
U =
[NaN] [NaN]
-verbatim-
|
(Note: upload size limit is set to 16384 kB, after insertion of the required
escape characters.)
Attach File(s):
Comment:
Depends on the following items: None found
Items that depend on this one: None found
Follow 7 latest changes.