[Debian-med-packaging] Bug#697015: Fwd: [Basics-commits] r295 - trunk

tom fogal tfogal at sci.utah.edu
Mon Sep 30 18:16:05 UTC 2013


FYI, fix has been applied upstream.

Thanks,

-tom

-------- Original Message --------
Subject: [Basics-commits] r295 - trunk
Date: Mon, 30 Sep 2013 12:09:07 -0600 (MDT)
From: tfogal at sci.utah.edu
To: basics-commits at sci.utah.edu

Author: tfogal
Date: 2013-09-30 12:09:06 -0600 (Mon, 30 Sep 2013)
New Revision: 295

Modified:
    trunk/LargeFileFD.cpp
Log:
Proper test for POSIX_FADV_WILLNEED.

According to
http://pubs.opengroup.org/onlinepubs/9699919799/functions/posix_fadvise.html

"The posix_fadvise() function is part of the Advisory Information
option and need not be provided on all implementations."

fcntl.h's specification indeed says that these constants must be
valid in #if processing directives.

Thanks Petr Salinger for the diagnosis/fix.

Modified: trunk/LargeFileFD.cpp
===================================================================
--- trunk/LargeFileFD.cpp	2013-04-30 15:22:29 UTC (rev 294)
+++ trunk/LargeFileFD.cpp	2013-09-30 18:09:06 UTC (rev 295)
@@ -75,7 +75,7 @@
    if(lseek(this->fd, offset+this->header_size, SEEK_SET) < 0) {
      throw std::ios_base::failure("could not seek to correct file 
position.");
    }
-#if _POSIX_C_SOURCE >= 200112L
+#ifdef POSIX_FADV_WILLNEED
    posix_fadvise(this->fd, offset+this->header_size, len, 
POSIX_FADV_WILLNEED);
  #endif

@@ -126,7 +126,7 @@
  void LargeFileFD::enqueue(uint64_t offset, size_t len)
  {
    if(len == 0) { return; }
-#if _POSIX_C_SOURCE >= 200112L
+#ifdef POSIX_FADV_WILLNEED
    int adv = posix_fadvise(this->fd, offset, len, POSIX_FADV_WILLNEED);
    // this should basically always succeed.  the only way it can fail 
is if we
    // gave it a bogus FD or something.  if that's the case, that points to

_______________________________________________
Basics-commits mailing list
Basics-commits at gforge.sci.utah.edu
https://gforge.sci.utah.edu/mailman/listinfo/basics-commits



More information about the Debian-med-packaging mailing list