On Wed, Aug 29, 2012 at 10:20 AM, Philip Nienhuis <address@hidden
<mailto:address@hidden>> wrote:
Philip Nienhuis wrote:
Rik wrote:
- mappers.cc: assert (isprint (charset), result) differs in
column 10.
\r\n versus \n line endings?
8/28/12
Philip,
This one should be easy to check. Can you try the following
at the
command line?
isprint ("\n")
isprint ("\r")
On Unix systems these are both false. It sounds like
character 10
("\n") may be considered printable no MinGW.
Both give 0.
For completeness here are the relevant results from fntest.log:
assert (isprint (charset),result) expected
Columns 1 through 22:
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
:
but got
Columns 1 through 22:
0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0
But:
octave.exe:2> isprint ( "\t")
ans = 1
so it is TAB that is considered printable. Sorry for my miscounting.
BTW isprint ("\t") does the same on octave-3.6.2_MinGW and
3.4.3_MinGW.
Apparently the test is wrong.
That should read: on MinGW the result of isprint ("\t") seems wrong.
Indeed. I thought the reason was the Windows C runtime implementation,
but I compiled a simple C test program to check the return value of
"isprint" under MSVC, and it returns 0 for the TAB character (char 9).
This calls for further investigation.