1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
|
Wed Jan 5 00:18:45 1994 Guido van Rossum (guido@voorn.cwi.nl)
* {Parser,Objects,Python,Makefile}/Makefile*in: made depend target
work with VPATH
* README: describe new build procedure; added section on building
for multiple architectures
* acconfig.h: fix (reversed!) comment for SYS_SELECT_WITH_SYS_TIME
* Modules/Makefile.pre.in: copy Setup from $(srcdir) if not
present, and use local Setup as input for makesetup script
Tue Jan 4 12:32:16 1994 Guido van Rossum (guido@voorn.cwi.nl)
* Note that there are no functional changes below -- just changes
to the build process and changes to avoid compiler warnings
* Modules/Setup: disable nis as well by default, change the
pertaining comments, and change the comments about the multimedia
modules to be default on
* fixed all warnings about function pointer initializations, and
miscellanous other warnings (e.g. about extern forward references
to static variables); touched random bits of code as a consequence
* changed configuration process and Makefiles to support VPATH;
for this, config.h(.in) now lives to the toplevel directory, the
toplevel Makefile is now created by configure as well, and various
improvements to it have been made (e.g. working tags and TAGS
targets), the makesetup script follows configure instead of
preceding it, it understands srcdir and has an exception for
glmodule.c, the intermediate file is called Makefile.pre, the
Makefiles don't use TOP any more and are much more careful about
the difference between .. and the toplevel directory, and I've
improved my understanding of how configure handles srcdir
* Modules/threadmodule.c: refuse to compile when WITH_THREAD is not
defined
* configure.in, acconfig.h, config.h.in, Include/ceval.h,
Modules/{stdwin,thread}module.c, Python/{ceval,pythonrun}.c:
renamed USE_THREAD to WITH_THREAD
* configure.in: add AC_PROG_INSTALL
* README, Python/version.c: version set to 1.0.0 BETA 3
* Demo, Include, Lib: added Makefile with clean/clobber targets
* README: added remarks on --with-svr4; unnumber special cases
* configure.in: only look for -lnsl and -lsocket if --with-svr4 is
specified, to avoid linking with them on IRIX 5
========================================================================
Release of 1.0.0 BETA 2 (Jan 3 1994)
========================================================================
Mon Jan 3 22:21:24 1994 Guido van Rossum (guido@voorn.cwi.nl)
* Include/myselect.h: fix typo in name of SYS_SELECT_WITH_SYS_TIME
* Parser/pgen.h: moved here from Include; removed extern
definition of 'gram'
* Parser/acceler.c: include node.h, now needed by parser.h
* README: added paragraph on testing
* Misc/python.man: changed date and add 1994 copyright
* Makefile: added test target
* Python/thread.c: include config.h if needed
* Parser/parser.h: remove references to struct _grammar and
similar things
* Modules/rotormodule.c (RTR_e_char, RTR_d_char): avoid warnings
by picky compilers about unsigned % signed
* README: added a section on building it for non-UNIX systems
* Makefile (configure): call autoheader when calling autoconf
* Include/config.h.in: now generated by autoheader
* acconfig.h: new file (input for autoheader)
* configure.in: added AC_REVISION call to top
* Modules/flmodule.c (form_setattr): one NULL should be 0
* Include/myselect.h: this now implies mytime.h and attempts to
work around systems where sys/select.h and sys/time.h can't be
included together
* Modules/socketmodule.c, Doc/libsocket.tex: remove socket avail()
method -- you can use select instead
* Modules/Setup: disable dbm, it is not truly portable
* Lib/sunau.py: incorporate one-line fix by Sjoerd
* Include/pgenheaders.h: include <stdlib.h> if its symbol
defined, not just on the mac
* Include/grammar.h: remove redundant structure tags
* Include/cgensupport.h: avoid possible macro argument
substitution inside string literal
* configure.in, Include/config.h.in: add test whether sys/select.h
and sys/times.h can be included by the same program
* Include/config.h.in: add lines for HAVE_SYS_UN_H and
HAVE_GETPEERNAME
* Extensions/mkext.py: copy change in library order from
Modules/Makefile.in.in
* Modules/Makefile.in.in: change library order subtly so -ltermcap
follows instead of precedes -lgl_s on SGI systems; this solves
(hides?) problems with clashing entry points
* configure.in: added sys/un.h to list of tested header files;
added getpeername to list of tested functions (both for
Modules/socketmodule.c)
* Modules/socketmodule.c: conditionally include sys/un.h and
change tests for AF_UNIX to tests for HAVE_SYS_UN_H; test for
HAVE_GETPEERNAME instead of NO_PEERNAME
* Modules/config.c.in: add marshal and __main__ built-in modules
* Python/sysmodule.c (list_builtin_module_names): sort the list
* Doc/Makefile: remove 'qua' from default targets
* Doc/README: add reference to ext.tex, change reference to
lib*.tex, explain that qua isn't built by default
* README: explain DESTDIR, clarify install procedure, add more
explanation to some options, add description of ChangeLog, add
wuarchive.wustl.edu to list of mirror sites
* Modules/socketmodule.c: make AF_UNIX code dependent on existence
of AF_UNIX (SCO ODT 3.0 doesn't support it -- let's hope it
doesn't define the symbol either)
* Makefile: attempt to fix install targets (added separate
libinstall and maninstall)
* Doc/libregex.tex: documented Tracy Tims' changes
* Modules/regexpr.c: redid Tracy Tims' changes to minimize diffs
(only two added lines now)
* Modules/regexmodule.c: fix core dump when asking a plain regex
object for a named group
Sun Jan 2 23:10:44 1994 Guido van Rossum (guido@voorn.cwi.nl)
* README, Python/version.c: changed version string to 1.0.0 BETA 2
* Modules/{regexpr.{c,h},regexmodule.c}: merged in Tracy Tims'
mods for named subexpressions
* Include/regexpr.h: moved to Modules/regexpr.h
* Modules/timingmodule.c: change tests for no arguments
* configure.in: remove strtoul from AC_REPLACE_FUNCS; remove
initial blank line (which got copied into configure so it wouldn't
start with #!/bin/sh as required)
* Python/compile.c: call mystrto(u)l instrad of strto(u)l
* Python/Makefile.in: add mystrtoul.c to OBJS
* Python/mystrtoul.c: renamed from strtol.c; renamed functions to
mystrto(u)l; this is now a standard source file (since some
systems have a strto(u)l that doesn't report errors properly)
* Modules/Setup: added entry for timing module
* Modules/{timing.h,timingmodule.c}: new files implementing GNN's
timing module
========================================================================
Release of 1.0.0 BETA (Jan 1 1994)
========================================================================
|
