1 |
|
---|
2 |
|
---|
3 | R e l e a s e N o t e s
|
---|
4 | +------------------------+
|
---|
5 |
|
---|
6 | v3.2.2 Beta 3:
|
---|
7 | --------------
|
---|
8 |
|
---|
9 | A few new features and fixes.
|
---|
10 |
|
---|
11 | o Fixed a bunch of debug info problems.
|
---|
12 |
|
---|
13 | o Added large file file support.
|
---|
14 |
|
---|
15 | o Enabled high memory suppor for LIBC. Two new option to GCC -Zhigh-mem
|
---|
16 | and -Zhigh-low. If the program is enabled for high memory use
|
---|
17 | -Zhigh-mem when linking.
|
---|
18 |
|
---|
19 | Note. that socket is not enabled yet.
|
---|
20 |
|
---|
21 | o Made LIBC preserve the FS selector.
|
---|
22 |
|
---|
23 | o Fixed crash with link386 and unresolved symbols.
|
---|
24 |
|
---|
25 | o A few other problem which is mentioned in the change log.
|
---|
26 |
|
---|
27 |
|
---|
28 | v3.2.2 Beta 2:
|
---|
29 | --------------
|
---|
30 |
|
---|
31 | A few bugfixes and a some new features (as usual).
|
---|
32 |
|
---|
33 | o Added file locking support to LIBC.
|
---|
34 |
|
---|
35 | o Fixed _System and _Optlink bugs and made _Optlink support floating
|
---|
36 | point arguments passing in registers.
|
---|
37 |
|
---|
38 | o Header updates related to limits and file locking primarily.
|
---|
39 |
|
---|
40 | o Timezone fixes.
|
---|
41 |
|
---|
42 | o Fixed some CRT init trouble in regards to getenv and putenv.
|
---|
43 |
|
---|
44 | o And some other minor stuff..
|
---|
45 |
|
---|
46 |
|
---|
47 | v3.2.2 Beta 1:
|
---|
48 | --------------
|
---|
49 |
|
---|
50 | A few bugfixes and a some new features.
|
---|
51 |
|
---|
52 | o Locale support in libc.
|
---|
53 |
|
---|
54 | o Implmented a different approach to OMF and weak symbols. This new hack
|
---|
55 | does a pre-linking before invoking the OMF linker generating new .def
|
---|
56 | if needed and an extra object file with aliases to resolv the weak
|
---|
57 | symbols correctly.
|
---|
58 | The old approach used when GCC_WEAKSYMS is present in the environment.
|
---|
59 |
|
---|
60 | Please note that due to a bug in BFD weak symbols doesn't work 100%
|
---|
61 | correctly. This will be addressed when we now move on to a later
|
---|
62 | version of binutils.
|
---|
63 |
|
---|
64 | o A couple of new functions in libc, and a few correction of existing
|
---|
65 | in respect to namespace.
|
---|
66 |
|
---|
67 | o Addressed issues with calling conventions. We now detect incompatible
|
---|
68 | calling conventions, emitting warnings and errors where appropritate.
|
---|
69 |
|
---|
70 | o _Optlink is much better, but there is still some issues left. Passing
|
---|
71 | prototyped floating point parameters to _Optlink doesn't work. Mixing
|
---|
72 | of parameters which goes into registers (eax,edx,ecx) and parameters
|
---|
73 | which doesn't - void foo(int eax, struct sss s, int edx) - don't work.
|
---|
74 | All these issues are being addressed and is expected to be fixed in
|
---|
75 | the next build.
|
---|
76 |
|
---|
77 | o Changed GCC to link with shared libc by default. This means that in
|
---|
78 | order to link with static libc one have to specify -static.
|
---|
79 | It also means that -Zmtd or -Zcrtdll is no longer required for linking
|
---|
80 | with shared libc and can be removed from all makescripts.
|
---|
81 |
|
---|
82 | o Changed EMXOMFLD defaults to ilink from Visual Age for C++ v3.6.5 or
|
---|
83 | later (preferably later). The emxomfld usage output will instruct you
|
---|
84 | how how to use older linkers. For VAC308 and VAC365 linkers /dbgpack
|
---|
85 | is specified if debuginfo is to be generated.
|
---|
86 | Emxomf checks for EMXOMFLD_TYPE=link386 when choosing debug info level.
|
---|
87 |
|
---|
88 | o Shipping three extra libc dlls:
|
---|
89 | libc02.dbg: Unstripped version of libc02.dll. This is identical
|
---|
90 | except from the presense of debug info.
|
---|
91 | libc02.prf: Linked together from profiled enabled libraries
|
---|
92 | using kProfiler.
|
---|
93 | libc02.elh: Standard libc but linked with an external eletrical
|
---|
94 | fence heap from kLibDbg.
|
---|
95 | These libc02.* dlls can be used instead of libc02.dll. kLibDbg and
|
---|
96 | kProfiler is available in separate zip files at same location as the
|
---|
97 | rest of the Innotek GCC suite. However these external tools are
|
---|
98 | provided as is. (I might actually not find time to put them out yet.)
|
---|
99 |
|
---|
100 | o Distribution is splitted into several zips:
|
---|
101 | -core.zip: All you really need.
|
---|
102 | -dbg.zip: Map files, unstripped tools and libs, and libc02.elh.
|
---|
103 | -prf.zip: Profiling libraries and libc02.prf.
|
---|
104 | libc.zip: The runtime libraries. (included in core.)
|
---|
105 | The big zip is everything.
|
---|
106 |
|
---|
107 | o I'm sure I've foregotten something this time....
|
---|
108 |
|
---|
109 |
|
---|
110 | v3.2.2 Alpha 3:
|
---|
111 | ---------------
|
---|
112 |
|
---|
113 | Primarily bugfixes.
|
---|
114 |
|
---|
115 | o Changed code generator for floating point truncations (double to int
|
---|
116 | casts and such) so PM can freely mess up the fpu status.
|
---|
117 |
|
---|
118 | o We were using the wrong float.h in two sense. We're now using the
|
---|
119 | correct one which includes all the PC defines and prototypes.
|
---|
120 |
|
---|
121 | o Fixed a few issues with error defines and tcpip.
|
---|
122 | Note that some of the error defines have changed value in the process
|
---|
123 | so we could become compatible with the tcpip stack and FreeBSD.
|
---|
124 |
|
---|
125 | o Corrected bugs in a number of headers.
|
---|
126 |
|
---|
127 | o Added more OS/2 prototypes and imports. There is still work left on
|
---|
128 | the header part,
|
---|
129 |
|
---|
130 | o Optimized emxomf debug info conversion.
|
---|
131 |
|
---|
132 | o Wrapped bsdselect for BSD 4.4 header mode to implement bsdselect().
|
---|
133 |
|
---|
134 | o .... what did we forget...
|
---|
135 |
|
---|
136 |
|
---|
137 | v3.2.2 Alpha 2:
|
---|
138 | ---------------
|
---|
139 |
|
---|
140 | A few bugfixes and a couple of new features.
|
---|
141 |
|
---|
142 | o The TCP/IP headers and libraries scheme have been changed.
|
---|
143 | The default is the current OS/2 tcpip toolkit (BSD 4.4 based).
|
---|
144 | To target the older OS/2 tcpip stack (BSD 4.3 based) you must
|
---|
145 | define TCPV40HDRS before including any TCP/IP headers and make
|
---|
146 | sure usr/lib/tcpipv4 is searched before usr/lib (this is to get
|
---|
147 | the right libsocket). It is recommended to use the -D compiler
|
---|
148 | option for the define and either the LIBRARY_PATH or the -L
|
---|
149 | compiler/linker option for the library.
|
---|
150 |
|
---|
151 | o There are a couple of new and updated based on FreeBSD 5.1.
|
---|
152 | sys\stdint.h and inttypes.h (and a few internal once) are new.
|
---|
153 | sys\types.h and unistd.h are the ones most heavily updated.
|
---|
154 | We're working on getting the headers and libc in a better shape
|
---|
155 | standardwise to easy compatability and portability.
|
---|
156 |
|
---|
157 | o A infinite loop in emxomf has been fixed.
|
---|
158 |
|
---|
159 | o Linker error on sopen() has been corrected.
|
---|
160 |
|
---|
161 | o .... what did we forget...
|
---|
162 |
|
---|
163 |
|
---|
164 | v3.2.2 Alpha 1:
|
---|
165 | ---------------
|
---|
166 |
|
---|
167 | This is the first release from Innotek, and it contains the Innotek LIBC.
|
---|
168 | In addition to the new LIBC there is a number of other major changes:
|
---|
169 |
|
---|
170 | o Name mangling in OMF mode have changed.
|
---|
171 | Until this release emxomf used to pretty much strip one leading '_'
|
---|
172 | from symbols when converting them to omf. This is now longer done
|
---|
173 | chiefly in order to correctly support mangling calling conventions
|
---|
174 | like __stdcall and __cdecl. This change have the following side
|
---|
175 | effects:
|
---|
176 | - _System and __cdecl no longer produces the same names and hence
|
---|
177 | all OS/2 API which are _System must be declared as _System in
|
---|
178 | order to link.
|
---|
179 | - Assembly code might need an extra underscore to work, or
|
---|
180 | perhaps change the prototypes to _System.
|
---|
181 | - Global variables are prepended and underscore, so mixing VAC and
|
---|
182 | GCC, or accessing globals from asm, take care.
|
---|
183 | (MSC v6.0 does this too, so it's not any uncommon practise.)
|
---|
184 |
|
---|
185 | The OS/2 headers which are included are all updated to use _System.
|
---|
186 |
|
---|
187 | o TCPIP headers are updated a quite a bit to match the toolkit ones
|
---|
188 | and to reflect the fact that the new libc does not a having common
|
---|
189 | handle space for files and sockets. These changes aren't 100%
|
---|
190 | completed yet, and usage probably will uncover some issues left in
|
---|
191 | there. Also for the programmer there is the following changes in
|
---|
192 | the apis provided:
|
---|
193 | - TCPV40HDRS are supported and should work like with the toolkit
|
---|
194 | headers. It will alias all functions to tcpip40_ editions
|
---|
195 | (using #defines) so we get the right imports thru -llibsocket.
|
---|
196 | - There is no support for sockets in select(). To do select on
|
---|
197 | sockets you will have to use bsdselect() or os2_select().
|
---|
198 | bsdselect() is only available for TCPV40HDRS. For
|
---|
199 | !defined(TCPV40HDRS) fd_set is in the emx/bsd format and not the
|
---|
200 | OS/2 tcpip v5.0 format.
|
---|
201 | - Socket IOCTLs are not supported by the ioctl() libc function
|
---|
202 | because of the separate handle space for sockets. For sockets
|
---|
203 | so_ioctl() and os2_ioctl() are used. os2_ioctl() is not available
|
---|
204 | when TCPV40HDRS is defined. According to the tcpref in v5.0
|
---|
205 | so_ioctl() only supports BSD ioctls.
|
---|
206 | - writev() and readv() also clashes with libc. For socket handles
|
---|
207 | use so_writev() and so_readv().
|
---|
208 |
|
---|
209 | o The prefered OMF linker is now ilink. link386 will fail with an odd
|
---|
210 | error message if you tell gcc/g++ to make debug info. Set the
|
---|
211 | EMXOMFLD_LINKER environment variable to value ilink in order to use
|
---|
212 | ilink as OMF linker.
|
---|
213 | It likely that we will change the default OMF linker to ilink in an
|
---|
214 | later version of the GCC suite.
|
---|
215 |
|
---|
216 | o .... what did we forget...
|
---|