• Philip Daniels (23/23) Aug 06 2012 FYI, I believe that bugs 8370 and 8375, to do with std.path.sep
• Jonathan M Davis (4/9) Aug 06 2012 -inline should have _zero_ impact on the semantics of any code. So, if

"Philip Daniels" <Philip.Daniels1971 gmail.com> writes:

FYI, I believe that bugs 8370 and 8375, to do with std.path.sep
being deprecated, are fixed in the latest version.



Does anybody have any clues how to proceed with bug 8250
"dirEntries fails to generate file list when compiled with
-inline". I also hit this bug, on my second day of playing with D
:-( I wondered if I would be able to track it down myself, so I
wrote a trivial test harness:


import std.stdio;
import std.file;

void main()
{
	auto filenames = dirEntries("/home/phil/", "*.d",
SpanMode.shallow);
	foreach (filename; filenames)
	{
		writeln(filename);
	}
}

This exhibits the problem when compiled with -inline. I then
removed the "import std.file" line and copied and pasted the
entire contents of file.d into this test harness, so that it is
self contained. dirEntries now works correctly. Compiler bug or
library bug?

"Jonathan M Davis" <jmdavisProg gmx.com> writes:

On Tuesday, August 07, 2012 00:00:49 Philip Daniels wrote:
 This exhibits the problem when compiled with -inline. I then
 removed the "import std.file" line and copied and pasted the
 entire contents of file.d into this test harness, so that it is
 self contained. dirEntries now works correctly. Compiler bug or
 library bug?

-inline should have _zero_ impact on the semantics of any code. So, if 
including it or not including it changes behavior, it's a compiler bug.

- Jonathan M Davis