Wed 30 Nov 2011 04:48:51 AM UTC, comment #9:
With the changeset
http://hg.savannah.gnu.org/hgweb/octave/rev/efa658122cc9
src\sysdep.cc .......................................... PASS 13/13
FAIL is disappeared! Thanks.
|
Wed 30 Nov 2011 04:39:24 AM UTC, comment #8:
Try a new tip after this changeset (http://hg.savannah.gnu.org/hgweb/octave/rev/efa658122cc9). Hopefully that changeset will fix the problem for you.
|
Wed 30 Nov 2011 04:16:29 AM UTC, comment #7:
The same fail still exits on recent development source.
However, this error is not critical.
|
Tue 11 Oct 2011 02:26:16 AM UTC, comment #6:
I have built the recent dev. source.
(changeset 13311:d590d9df5596
date Mon Oct 10 15:08:48 2011 -0400)
|
Thu 06 Oct 2011 11:23:09 PM UTC, comment #5:
One possibility is to change the left-hand side of the test rather than the right-hand side.
|
Thu 06 Oct 2011 11:10:38 PM UTC, comment #4:
I think the issue may lie in liboctave/file-ops.cc in function tilde_expand_word. The excerpt below is at line 220.
So it looks like "~/foobar" is immediately parsed as "~" and filename.substr (1) is "/foobar". Thus, no file separator is ever inserted.
|
Thu 06 Oct 2011 10:46:14 PM UTC, comment #3:
I think it is still a jumble. fullfile() and filesep both use '\' while tilde_expand is using '/'. Note that this is only for the last element of the path. One would think that tilde_expand would use '\' for all of the file separation locations but that would be wrong.
Tatsuro, could you run the following 3 lines of code to show clearly what each function returns
|
Thu 06 Oct 2011 09:42:12 PM UTC, comment #2:
I don't understand the fix.
So fullfile is using \ on Windows systems? That I would expect. But why is tilde_expand not producing a file with \ on Windows systems? And why is filesep returning /? I thought all of this worked in the past, with evertything producing \ on Windows systems. What changed?
|
Thu 06 Oct 2011 09:32:26 PM UTC, comment #1:
Fixed in this changeset (http://hg.savannah.gnu.org/hgweb/octave/rev/4cca76e15876).
|
Tue 26 Jul 2011 04:08:53 AM UTC, original submission:
I found a new fail on the test of sysdep.cc
This fail is not so serious but I report it.
|