Commit | Line | Data |
a0d0e21e |
1 | # hints/dec_osf.sh |
8e964347 |
2 | |
b971f6e4 |
3 | # * If you want to debug perl or want to send a |
8e964347 |
4 | # stack trace for inclusion into an bug report, call |
5 | # Configure with the additional argument -Doptimize=-g2 |
6 | # or uncomment this assignment to "optimize": |
7 | # |
8 | #optimize=-g2 |
9 | # |
8a019ce7 |
10 | # If you want both to optimise and debug with the DEC cc |
11 | # you must have -g3, e.g. "-O4 -g3", and (re)run Configure. |
12 | # |
b971f6e4 |
13 | # * gcc can always have both -g and optimisation on. |
8a019ce7 |
14 | # |
b971f6e4 |
15 | # * debugging optimised code, no matter what compiler |
8a019ce7 |
16 | # one is using, can be surprising and confusing because of |
17 | # the optimisation tricks like code motion, code removal, |
18 | # loop unrolling, and inlining. The source code and the |
19 | # executable code simply do not agree any more while in |
20 | # mid-execution, the optimiser only cares about the results. |
21 | # |
b971f6e4 |
22 | # * Configure will automatically add the often quoted |
f2c69087 |
23 | # -DDEBUGGING for you if the -g is specified. |
24 | # |
b971f6e4 |
25 | # * There is even more optimisation available in the new |
f2c69087 |
26 | # (GEM) DEC cc: -O5 and -fast. "man cc" will tell more about them. |
27 | # The jury is still out whether either or neither help for Perl |
28 | # and how much. Based on very quick testing, -fast boosts |
29 | # raw data copy by about 5-15% (-fast brings in, among other |
30 | # things, inlined, ahem, fast memcpy()), while on the other |
31 | # hand searching things (index, m//, s///), seems to get slower. |
32 | # Your mileage will vary. |
8e964347 |
33 | # |
b971f6e4 |
34 | # * The -std is needed because the following compiled |
e60a08f8 |
35 | # without the -std and linked with -lm |
36 | # |
37 | # #include <math.h> |
38 | # #include <stdio.h> |
39 | # int main(){short x=10,y=sqrt(x);printf("%d\n",y);} |
40 | # |
41 | # will in Digital UNIX 3.* and 4.0b print 0 -- and in Digital |
42 | # UNIX 4.0{,a} dump core: Floating point exception in the printf(), |
43 | # the y has become a signaling NaN. |
44 | # |
b971f6e4 |
45 | # * Compilation warnings like: |
46 | # |
47 | # "Undefined the ANSI standard macro ..." |
48 | # |
49 | # can be ignored, at least while compiling the POSIX extension |
50 | # and especially if using the sfio (the latter is not a standard |
51 | # part of Perl, never mind if it says little to you). |
52 | # |
8e964347 |
53 | |
e60a08f8 |
54 | # If using the DEC compiler we must find out the DEC compiler style: |
55 | # the style changed between Digital UNIX (aka DEC OSF/1) 3 and |
56 | # Digital UNIX 4. The old compiler was originally from Ultrix and |
57 | # the MIPS company, the new compiler is originally from the VAX world |
58 | # and it is called GEM. Many of the options we are going to use depend |
59 | # on the compiler style. |
60 | |
a68015de |
61 | cc=${cc:-cc} |
62 | |
b691c02f |
63 | # do NOT, I repeat, *NOT* take away the leading tabs |
64 | # Configure Black Magic (TM) |
b971f6e4 |
65 | # reset |
e60a08f8 |
66 | _DEC_cc_style= |
8d5a0f9f |
67 | case "`$cc -v 2>&1 | grep cc`" in |
2cae8c0d |
68 | *gcc*) _gcc_version=`$cc -v 2>&1 | grep "gcc version" | sed 's%^gcc version \([0-9]*\)\.\([0-9]*\) .*%\1 \2%'` |
69 | set $_gcc_version |
70 | if test "$1" -lt 2 -o \( "$1" -eq 2 -a "$2" -lt 95 \); then |
71 | cat >&4 <<EOF |
72 | |
73 | Your cc seems to be gcc and its version seems to be less than 2.95. |
74 | This is not a good idea since old versions of gcc are known to produce |
75 | buggy code when compiling Perl (and no doubt for other programs, too). |
76 | |
77 | Therefore, I strongly suggest upgrading your gcc. (Why don't you |
78 | use the vendor cc is also a good question. It comes with the operating |
79 | system and produces good code.) |
80 | |
81 | Note that as of gcc 2.95 (19990728) and Perl 5.6.0 (end of March 2000) |
82 | if the said Perl is compiled with the said gcc the lib/sdbm test will |
83 | dump core. As this doesn't happen with the vendor cc, this is |
84 | most probably a lingering bug in gcc. Therefore unless you have |
85 | a better gcc you are still better off using the vendor cc. |
86 | |
87 | Cannot continue, aborting. |
88 | |
89 | EOF |
90 | exit 1 |
91 | fi |
92 | ;; |
e60a08f8 |
93 | *) # compile something small: taint.c is fine for this. |
94 | # the main point is the '-v' flag of 'cc'. |
52c7d5b6 |
95 | case "`cc -v -I. -c taint.c -o taint$$.o 2>&1`" in |
e60a08f8 |
96 | */gemc_cc*) # we have the new DEC GEM CC |
dfa3a3d3 |
97 | _DEC_cc_style=new |
8a019ce7 |
98 | ;; |
e60a08f8 |
99 | *) # we have the old MIPS CC |
dfa3a3d3 |
100 | _DEC_cc_style=old |
8a019ce7 |
101 | ;; |
e60a08f8 |
102 | esac |
103 | # cleanup |
52c7d5b6 |
104 | rm -f taint$$.o |
e60a08f8 |
105 | ;; |
106 | esac |
107 | |
b971f6e4 |
108 | # be nauseatingly ANSI |
8d5a0f9f |
109 | case "`$cc -v 2>&1 | grep gcc`" in |
9607fc9c |
110 | *gcc*) ccflags="$ccflags -ansi" |
b971f6e4 |
111 | ;; |
112 | *) ccflags="$ccflags -std" |
113 | ;; |
114 | esac |
115 | |
116 | # for gcc the Configure knows about the -fpic: |
117 | # position-independent code for dynamic loading |
118 | |
e60a08f8 |
119 | # we want optimisation |
120 | |
121 | case "$optimize" in |
8d5a0f9f |
122 | '') case "`$cc -v 2>&1 | grep gcc`" in |
e60a08f8 |
123 | *gcc*) |
124 | optimize='-O3' ;; |
125 | *) case "$_DEC_cc_style" in |
313489a2 |
126 | new) optimize='-O4' |
127 | ccflags="$ccflags -fprm d -ieee" |
128 | ;; |
e60a08f8 |
129 | old) optimize='-O2 -Olimit 3200' ;; |
130 | esac |
b971f6e4 |
131 | ccflags="$ccflags -D_INTRINSICS" |
e60a08f8 |
132 | ;; |
8e964347 |
133 | esac |
134 | ;; |
1fc4cb55 |
135 | esac |
28757baa |
136 | |
313489a2 |
137 | # Make glibpth agree with the compiler suite. Note that /shlib |
138 | # is not here. That's on purpose. Even though that's where libc |
139 | # really lives from V4.0 on, the linker (and /sbin/loader) won't |
140 | # look there by default. The sharable /sbin utilities were all |
141 | # built with "-Wl,-rpath,/shlib" to get around that. This makes |
142 | # no attempt to figure out the additional location(s) searched by |
143 | # gcc, since not all versions of gcc are easily coerced into |
144 | # revealing that information. |
b0362887 |
145 | glibpth="/usr/shlib /usr/ccs/lib /usr/lib/cmplrs/cc" |
146 | glibpth="$glibpth /usr/lib /usr/local/lib /var/shlib" |
313489a2 |
147 | |
8e964347 |
148 | # dlopen() is in libc |
149 | libswanted="`echo $libswanted | sed -e 's/ dl / /'`" |
150 | |
b971f6e4 |
151 | # libPW contains nothing useful for perl |
8a019ce7 |
152 | libswanted="`echo $libswanted | sed -e 's/ PW / /'`" |
153 | |
723e14d4 |
154 | # libnet contains nothing useful for perl here, and doesn't work |
155 | libswanted="`echo $libswanted | sed -e 's/ net / /'`" |
156 | |
b971f6e4 |
157 | # libbsd contains nothing used by perl that is not already in libc |
8a019ce7 |
158 | libswanted="`echo $libswanted | sed -e 's/ bsd / /'`" |
159 | |
b971f6e4 |
160 | # libc need not be separately listed |
8a019ce7 |
161 | libswanted="`echo $libswanted | sed -e 's/ c / /'`" |
162 | |
b971f6e4 |
163 | # ndbm is already in libc |
164 | libswanted="`echo $libswanted | sed -e 's/ ndbm / /'`" |
8a019ce7 |
165 | |
166 | # the basic lddlflags used always |
167 | lddlflags='-shared -expect_unresolved "*"' |
168 | |
b971f6e4 |
169 | # Fancy compiler suites use optimising linker as well as compiler. |
170 | # <spider@Orb.Nashua.NH.US> |
b8f0c030 |
171 | case "`uname -r`" in |
b971f6e4 |
172 | *[123].*) # old loader |
173 | lddlflags="$lddlflags -O3" |
174 | ;; |
a07564da |
175 | *) if $test "X$optimize" = "X$undef"; then |
176 | lddlflags="$lddlflags -msym" |
177 | else |
baa8820a |
178 | case "`/usr/sbin/sizer -v`" in |
e3159d07 |
179 | *4.0D*) |
180 | # QAR 56761: -O4 + .so may produce broken code, |
181 | # fixed in 4.0E or better. |
182 | ;; |
183 | *) |
184 | lddlflags="$lddlflags $optimize" |
185 | ;; |
186 | esac |
187 | # -msym: If using a sufficiently recent /sbin/loader, |
188 | # keep the module symbols with the modules. |
baa8820a |
189 | lddlflags="$lddlflags -msym -std" |
a07564da |
190 | fi |
b971f6e4 |
191 | ;; |
192 | esac |
193 | # Yes, the above loses if gcc does not use the system linker. |
194 | # If that happens, let me know about it. <jhi@iki.fi> |
195 | |
8a019ce7 |
196 | |
b971f6e4 |
197 | # If debugging or (old systems and doing shared) |
198 | # then do not strip the lib, otherwise, strip. |
199 | # As noted above the -DDEBUGGING is added automagically by Configure if -g. |
8a019ce7 |
200 | case "$optimize" in |
201 | *-g*) ;; # left intentionally blank |
b8f0c030 |
202 | *) case "`uname -r`" in |
b971f6e4 |
203 | *[123].*) |
204 | case "$useshrplib" in |
205 | false|undef|'') lddlflags="$lddlflags -s" ;; |
206 | esac |
207 | ;; |
8a019ce7 |
208 | *) lddlflags="$lddlflags -s" |
b971f6e4 |
209 | ;; |
210 | esac |
211 | ;; |
8a019ce7 |
212 | esac |
8e964347 |
213 | |
214 | # |
313489a2 |
215 | # Make embedding in things like INN and Apache more memory friendly. |
216 | # Keep it overridable on the Configure command line, though, so that |
217 | # "-Uuseshrplib" prevents this default. |
218 | # |
219 | |
2bf2710f |
220 | case "$_DEC_cc_style.$useshrplib" in |
221 | new.) useshrplib="$define" ;; |
222 | esac |
313489a2 |
223 | |
85ab1d1d |
224 | # The EFF_ONLY_OK from <sys/access.h> is present but dysfunctional for |
225 | # [RWX]_OK as of Digital UNIX 4.0[A-D]?. If and when this gets fixed, |
226 | # please adjust this appropriately. See also pp_sys.c just before the |
227 | # emulate_eaccess(). |
5ff3f7a4 |
228 | |
baa8820a |
229 | # Fixed in V5.0A. |
230 | case "`/usr/sbin/sizer -v`" in |
f60a5d46 |
231 | *5.0[A-Z]*|*5.[1-9]*|*[6-9].[0-9]*) |
baa8820a |
232 | : ok |
233 | ;; |
234 | *) |
235 | # V5.0 or previous |
5ff3f7a4 |
236 | pp_sys_cflags='ccflags="$ccflags -DNO_EFF_ONLY_OK"' |
baa8820a |
237 | ;; |
238 | esac |
5ff3f7a4 |
239 | |
6b8eaf93 |
240 | # The off_t is already 8 bytes, so we do have largefileness. |
241 | |
f60a5d46 |
242 | cat > UU/usethreads.cbu <<'EOCBU' |
104d25b7 |
243 | # This script UU/usethreads.cbu will get 'called-back' by Configure |
244 | # after it has prompted the user for whether to use threads. |
104d25b7 |
245 | case "$usethreads" in |
246 | $define|true|[yY]*) |
75d72f2c |
247 | # Threads interfaces changed with V4.0. |
8d5a0f9f |
248 | case "`$cc -v 2>&1 | grep gcc`" in |
75d72f2c |
249 | *gcc*)ccflags="-D_REENTRANT $ccflags" ;; |
250 | *) case "`uname -r`" in |
251 | *[123].*) ccflags="-threads $ccflags" ;; |
252 | *) ccflags="-pthread $ccflags" ;; |
253 | esac |
104d25b7 |
254 | ;; |
75d72f2c |
255 | esac |
256 | case "`uname -r`" in |
257 | *[123].*) libswanted="$libswanted pthreads mach exc c_r" ;; |
258 | *) libswanted="$libswanted pthread exc" ;; |
259 | esac |
e883c329 |
260 | |
baa8820a |
261 | case "$usemymalloc" in |
262 | '') |
263 | usemymalloc='n' |
264 | ;; |
265 | esac |
104d25b7 |
266 | ;; |
267 | esac |
268 | EOCBU |
269 | |
f60a5d46 |
270 | cat > UU/uselongdouble.cbu <<'EOCBU' |
271 | # This script UU/uselongdouble.cbu will get 'called-back' by Configure |
272 | # after it has prompted the user for whether to use long doubles. |
273 | case "$uselongdouble" in |
274 | $define|true|[yY]*) d_Gconvert='sprintf((b),"%.*Lg",(n),(x))' ;; |
275 | esac |
276 | EOCBU |
277 | |
313489a2 |
278 | # |
e60a08f8 |
279 | # Unset temporary variables no more needed. |
280 | # |
281 | |
282 | unset _DEC_cc_style |
283 | |
284 | # |
8e964347 |
285 | # History: |
286 | # |
85ab1d1d |
287 | # perl5.005_51: |
288 | # |
289 | # September-1998 Jarkko Hietaniemi <jhi@iki.fi> |
290 | # |
291 | # * Added the -DNO_EFF_ONLY_OK flag ('use filetest;' support). |
292 | # |
313489a2 |
293 | # perl5.004_57: |
294 | # |
295 | # 19-Dec-1997 Spider Boardman <spider@Orb.Nashua.NH.US> |
296 | # |
85ab1d1d |
297 | # * Newer Digital UNIX compilers enforce signaling for NaN without |
313489a2 |
298 | # -ieee. Added -fprm d at the same time since it's friendlier for |
299 | # embedding. |
300 | # |
301 | # * Fixed the library search path to match cc, ld, and /sbin/loader. |
302 | # |
303 | # * Default to building -Duseshrplib on newer systems. -Uuseshrplib |
304 | # still overrides. |
305 | # |
306 | # * Fix -pthread additions for useshrplib. ld has no -pthread option. |
307 | # |
308 | # |
723e14d4 |
309 | # perl5.004_04: |
310 | # |
311 | # 19-Sep-1997 Spider Boardman <spider@Orb.Nashua.NH.US> |
312 | # |
313 | # * libnet on Digital UNIX is for JAVA, not for sockets. |
314 | # |
315 | # |
b971f6e4 |
316 | # perl5.003_28: |
317 | # |
318 | # 22-Feb-1997 Jarkko Hietaniemi <jhi@iki.fi> |
319 | # |
320 | # * Restructuring Spider's suggestions. |
321 | # |
322 | # * Older Digital UNIXes cannot handle -Olimit ... for $lddlflags. |
323 | # |
324 | # * ld -s cannot be used in older Digital UNIXes when doing shared. |
325 | # |
326 | # |
327 | # 21-Feb-1997 Spider Boardman <spider@Orb.Nashua.NH.US> |
328 | # |
329 | # * -hidden removed. |
330 | # |
331 | # * -DSTANDARD_C removed. |
332 | # |
333 | # * -D_INTRINSICS added. (that -fast does not seem to buy much confirmed) |
334 | # |
335 | # * odbm not in libc, only ndbm. Therefore dbm back to $libswanted. |
336 | # |
337 | # * -msym for the newer runtime loaders. |
338 | # |
339 | # * $optimize also in $lddflags. |
340 | # |
341 | # |
e60a08f8 |
342 | # perl5.003_27: |
343 | # |
344 | # 18-Feb-1997 Jarkko Hietaniemi <jhi@iki.fi> |
345 | # |
346 | # * unset _DEC_cc_style and more commentary on -std. |
347 | # |
348 | # |
dfa3a3d3 |
349 | # perl5.003_26: |
350 | # |
351 | # 15-Feb-1997 Jarkko Hietaniemi <jhi@iki.fi> |
352 | # |
e60a08f8 |
353 | # * -std and -ansi. |
dfa3a3d3 |
354 | # |
355 | # |
f2c69087 |
356 | # perl5.003_24: |
357 | # |
358 | # 30-Jan-1997 Jarkko Hietaniemi <jhi@iki.fi> |
359 | # |
360 | # * Fixing the note on -DDEBUGGING. |
361 | # |
362 | # * Note on -O5 -fast. |
363 | # |
364 | # |
8a019ce7 |
365 | # perl5.003_23: |
366 | # |
367 | # 26-Jan-1997 Jarkko Hietaniemi <jhi@iki.fi> |
368 | # |
369 | # * Notes on how to do both optimisation and debugging. |
370 | # |
371 | # |
372 | # 25-Jan-1997 Jarkko Hietaniemi <jhi@iki.fi> |
373 | # |
374 | # * Remove unneeded libraries from $libswanted: PW, bsd, c, dbm |
375 | # |
376 | # * Restructure the $lddlflags build. |
377 | # |
378 | # * $optimize based on which compiler we have. |
379 | # |
380 | # |
8e964347 |
381 | # perl5.003_22: |
382 | # |
383 | # 23-Jan-1997 Achim Bohnet <ach@rosat.mpe-garching.mpg.de> |
384 | # |
385 | # * Added comments 'how to create a debugging version of perl' |
386 | # |
387 | # * Fixed logic of this script to prevent stripping of shared |
388 | # objects by the loader (see ld man page for -s) is debugging |
389 | # is set via the -g switch. |
390 | # |
391 | # |
392 | # 21-Jan-1997 Achim Bohnet <ach@rosat.mpe-garching.mpg.de> |
393 | # |
394 | # * now 'dl' is always removed from libswanted. Not only if |
395 | # optimize is an empty string. |
396 | # |
397 | # |
398 | # 17-Jan-1997 Achim Bohnet <ach@rosat.mpe-garching.mpg.de> |
399 | # |
400 | # * Removed 'dl' from libswanted: When the FreePort binary |
401 | # translator for Sun binaries is installed Configure concludes |
402 | # that it should use libdl.x.yz.fpx.so :-( |
403 | # Because the dlopen, dlclose,... calls are in the |
404 | # C library it not necessary at all to check for the |
405 | # dl library. Therefore dl is removed from libswanted. |
406 | # |
407 | # |
408 | # 1-Jan-1997 Achim Bohnet <ach@rosat.mpe-garching.mpg.de> |
409 | # |
410 | # * Set -Olimit to 3200 because perl_yylex.c got too big |
411 | # for the optimizer. |
412 | # |