Changeset 2669

Timestamp:

Mar 19, 2006, 5:30:13 AM (19 years ago)

Author:

bird

Message:

synced with current libc-0.6 changes.

Location:

trunk/doc

Files:

• MailingLists.os2 (modified) (1 diff)
• ReleaseNotes.os2 (modified) (1 diff)

trunk/doc/MailingLists.os2

@@ -1 +1 @@
 
                                                                                                                 
-G C C   a n d   L I B C   m a i l i n g   l i s t ( s )
-------------------------------------------------------------
+ L I B C   m a i l i n g   l i s t
+ +-------------------------------+
 
 
-A mailing list primarily intended for development of LIBC and GCC is crated on
-netlabs. If there is too much usage questions and problem reports we'll create
-a new list, but let's try with one list for a start.
+A mailing list for LIBC & GCC users has been created on netlabs.org.
+Bugs, user questions, development questions, new ticket notifications,
+patches, etc. can be posted on that list.
 
-Netlabs is using majormajor, and the procedure for subscribing is to send a
-mail to majormajor@netlabs.org with this body:
-        subscribe gcc
-        end
+Netlabs is using ezmlm, and the procedure for subscribing is simply to
+send a mail to libc-user-subscribe@netlabs.org.
         
-The list gets archived at http://www.mail-archive.com/
+The list can also be accessed using gmane:
+     http://dir.gmane.org/gmane.org.netlabs.libc.user
 
 

trunk/doc/ReleaseNotes.os2

@@ -2 +2 @@
 
     R e l e a s e    N o t e s    G C C   f o r   O S / 2
-    +----------------------------------------------------+
-
-
-(C)'2003-2005 InnoTek Systemberatung GmbH
-http://www.innotek.de
-(C)'2004-2005 Knut St. Osmundsen
+    +---------------------------------------------------+
+
+
+(C)'2004-2006 Knut St. Osmundsen (bird-gccspam@anduin.net)
+(C)'2003-2005 InnoTek Systemberatung GmbH (http://www.innotek.de)
+
+
+GCC v3.3.5 CSD 1 / LIBC v0.6.1:
+-------------------------------
+
+    First, thanks to Yuri, nickk, Paul, Andy and Michal for patches,
+    bugs and other kind of feedback.
+
+
+    Fix list:
+
+        #19     gcc: gcc limits.h overrides limits.h from libc
+        #20     libc: random panic SIGSEGV when doing parallel builds
+        #48     libc: freopen(NULL, "rb", stdio) panics (doesn't work
+                  as expected)
+        #60     libc: filehandle inheritance doesn't work right!
+        #22     libc: path resolver fails on open file
+        #23     libc: symlink is broken
+        #50     libc: umask isn't inherited
+        #51     libc: unlink fails on readonly file
+        #53     libc: freopen("missing",,) causes panic
+        #54     libc: fchdir(open('.', O_RDONLY)) may corrupt path
+                  resolving
+        #58     libc: closed standard handles cause uninitialized
+                  stdout/stdin/stderr streams
+        #21     libc: limit Unix EAs to specific file systems
+        #24     libc: fts crashes because it's still using chdir()
+        #32     libc: readdir_r returns -1 instead of errno
+        #33     libc: panic when using timers
+        #38     libc: when debugged default to breakpoint on panic
+        #40     libc: open("/", RDONLY) fails.
+        #41     libc: incorrect __libc_back_fsUnixAttribsGet() usage
+                  caused st_dev and st_ino to be 0 in some stat()s
+                  and chmod & fchmod to set MODE EA without type
+        #46     ilink map file name is wrong when multiple dots are in
+                  target name
+        #49     libc: st_ctime is incorrect and st_birthtime isn't
+                  converted to UTC
+        #52     libc: path resolver fails resolving a directory
+                  symlink for lstat
+        #61     libc: symlink(,existing) returns the wrong errno
+        #62     libc: getdelim / getline returns failure when the last
+                  line doesn't end with the delimiter / newline
+        #65     libc: sys/stat.h doesn't work in non BSD source mode
+        #67     libc: Remove ncurses headers
+        #68     libc: getopt prototype clashes (getopt.h/unistd.h)
+        #69     libc: basename & dirname return incorrect results
+                  for paths with drive
+        #71     libc: Incorrect fmutex padding in FILE
+        #27     ld bug when map files are requested
+        #34     weakld: generate duplicate weak aliases bird
+        #30     setlocale / ctype is broken
+        #70     emxomf: Demangle function names for omf debug info
+        #55     emxomfld doesn't delete temporary files
+        #56     emxomfld doesn't properly allocate space for map file name
+        #28     emxbind doesn't read library name from .def file
+        #36     emxbind: export truncation bug and allow export & imports
+                     up to 255 chars
+
+    Full details can be found at:
+
+        http://svn.netlabs.org/libc/query?status=closed&milestone=libc-0.6.1
+
+
+
+GCC v3.3.5 / LIBC v0.6:
+-----------------------
+
+    There was a build problem with the first attempt at releasing.
+    This is the 2nd attempt, hope I got it right this time...
+
+    WARNINGS
+
+    There are several incompatible interface changes. So, you'll have
+    to rebuild static libraries and it's a good idea to do the same
+    thing for dynamic libs and programs too.
+
+    The build is not tested as much as I would've liked. I'm very
+    pressed on time as usual, but I really wanted to get the 0.6 stuff
+    out now. I can only hope everything still builds and works.
+
+
+    The Project is Moving
+
+    The source code has been migrated to subversion and now resides
+    on the svn.netlabs.org server. The server also provides an issue
+    tracker and project management service called trac. Check it
+    out at http://svn.netlabs.org/libc/.
+
+    The source code for this release (and any bugfix releases on
+    the 0.6 code base) can be checked out using this url:
+        http://svn.netlabs.org/repos/libc/branches/libc-0.6
+
+    Developers who wish to contribute to the 0.7 code will have to
+    diff against the trunk. There are several pending restructuring
+    tasks on 0.7, so don't expect everything to build fine or work
+    properly while those are still open. Anyway, the url:
+        http://svn.netlabs.org/repos/libc/trunk
+
+    Users are encouraged to join the gcc-user@netlabs.org mailing list
+    and use that list for questions rather than the InnoTek form. If
+    you think you've found a bug, you should first ask on the mailing
+    list, and if I decide it's a *real*, *new*, bug I'll open a
+    ticket for it and address it.
+
+    Developers which have patches for real bugs or enhancements can
+    create trac tickets and drop me or the gcc-user list a mail (we
+    haven't figure out how to get email notifications from trac yet).
+
+
+    New features:
+
+        o Added __libc_Back_gfProcessHandlePCBatchScripts and
+          __libc_Back_gfProcessHandleHashBangScripts (globale
+          variables) to turn on/off script handling in exec/spawn.
+          Very useful for shells.
+
+        o Changed the default stack of emxomfld and emxbind to 1MB,
+          it used to be 8KB and 8MB.
+
+        o open() can now open directories like on unix, and fchdir()
+          can chdir to the opened directories.
+
+        o Implemented much of the unix-attributes-in EA stuff.
+
+        o Implemented getdirents(), will be used to replace
+          opendir/readdir/seekdir/telldir/closedir in 0.7.
+
+        o Implemented the fcntl interface for setting and clearing
+          the O_NONBLOCK flag on sockets.
+
+
+    Bug fixes:
+
+        o Fixed bug preventing libc from working on legacy systems.
+          Kudos to Andrew Below for spotting it.
+
+        o Fixed ld bustage in rc1.
+
+        o Fixed fork bug in shm.
+
+        o Fixed spawn/exec bug where .exe was appended too early
+          preventing configure and other scripts without suffix
+          from being executed.
+
+        o Fixed several signal bugs (deadlocks, sigsuspend, pause).
+
+        o Fixed symbol truncation problem in emxomf.
+
+        o Fixed a generic error handling bug in ioctl, write, read,
+          dup and dup2 which was causing incorrect errno values.
+
+
+    Known Issues, see open ticket in trac:
+        http://svn.netlabs.org/libc/report/1