Log message:
Solaris doesn't have <paths.h>. Should fix the last traces of PR 39717.

Log message:
Note upstream tracker url (thanks to obache for filing it!)

Log message:
Undo the removal of patch-ae in the previous commit, to avoid the
insecure-temporary-files problem patch-ae was supposed to fix.
Provide a new patch-ae that does something more like the right thing.

(The old one tried to open and write to a directory created with
mkdtemp(), so even if this code doesn't fully work it's still an
improvement.)

XXX: compile-tested only, don't have time to do it properly right now...

PKGREVISION++.

Log message:
Remove patch-ae.
It replace tmpnam() with mkdtemp(), but
 * It exists since initial import, but no reason.
 * mkdtemp(3) is not portable, but used unconditionally, reported by PR 39717.
 * tmpnam(3) is used to get temp filename, but mkdtemp(3) create temp directory,
   and return the path.  So, the replacement is completely mistaken.

Bump PKGREVISION.

While here, marked as DESTDIR ready.

Log message:
Changes 0.11.57:
* Updated to automake/autoconf/aclocal/a* 1.10.x, with the usual
  pains. I hope it builds for people other than me ...
* Updated to a new included version of GNU gettext (0.16.1).
* Made gentoo's start-up more friendly to GTK+, so that standard
  options supported by the toolkit (like --sync) work, now.
* Added new Mexican Spanish translation.
* Added a "Case" page to the RenameRE command, with options to
  easily convert filenames to all-lower or all-upper case. Far
  easier than manually entering all characters for "Map". Also
  refactored code of RenameRE quite a bit, it is shorter now.
* Fixed bugs where the GTK+ event-flushing loop used in various
  places in the code generated a spurious error 11
  ("Resource temporarily unavailable").
* Added a new setting ("Buffer Size for mmap()" under Commands/
  Options/ViewText) that controls how much data is passed from
  the mmap()ed text file to the GTK+ text widget at a time. Before
  this release, this was always "all of it", which seems to cause
  problems sometimes. If you get an empty viewer when you try
  looking at a large-ish (>64 KB) text file and you use mmap(),
  lower this setting.

Log message:
Fix this to not use the compile-time $HOME to look for the gtkrc file.
(Since that is set to work/.home, it was compiling in the pkgsrc work
directory and failing for PKG_DEVELOPERs.)
Use getenv("HOME") instead.
PKGREVISION++.

Log message:
Update gentoo to 0.11.56.
Patch provided by Martin Wilke via PR 34341.

0.11.56
* Fixed buglet that caused copying of protection bits to always
  be a fatal error, regardless of the "Ignore Failure to Copy
  Attributes" option setting.
* Fixed a silly case of errno "leaking" out of the window title
  setting code (which needs to mess with NLS). This had gentoo
  break totally (couldn't even copy files) on Ubuntu. Big thanks
  to M. Williams for letting me ssh and vnc to his Ubuntu box so
  I could hunt this one down.
* Rewrote core of regular file copying function. It seems when
  writing to a Samba-mounted (network) share, write() doesn't
  like large buffer sizes. So we loop to empty the buffer of read
  data, doing as many writes() per read() as is necessary. Odd.
* Fixed bug that made it impossible to use RenameSeq on volumes
  using a VFAT filesystem. This was due to the use of invalid
  characters in temporary filenames.
* Made RenameSeq's start value settable through a spinbutton in
  the dialog, rather than a plain entry. Faster small changes.
* Added a new page to the RenameRE dialog, 'Map', which does
  simple character mappings on filenames. Also supports removing
  a set of character (think 'tr', applied twice).
* Added information about missing options (--locale-info, --run)
  to the manual page.

Log message:
Consistently use statfs on DragonFly.

Log message:
Forgot patchsum.

Log message:
Add DragonFly support. Replace the static FILE magic with a different
(messy) magic, which doesn't relay on sizeof(FILE), but on two integer
variables having separate addresses and pointer casts not changing
this fact.