This is the mail archive of the
cygwin-apps
mailing list for the Cygwin project.
Re: [HEADSUP] Dropping libopenssl098 from distro [gcc bug]
- From: Ken Brown <kbrown at cornell dot edu>
- To: cygwin-apps at cygwin dot com
- Cc: JonY <10walls at gmail dot com>
- Date: Mon, 02 Feb 2015 10:46:54 -0500
- Subject: Re: [HEADSUP] Dropping libopenssl098 from distro [gcc bug]
- Authentication-results: sourceware.org; auth=none
- References: <20150114141344 dot GG15791 at calimero dot vinschen dot de> <87ppahz3kt dot fsf at Rainer dot invalid> <20150114172404 dot GN15791 at calimero dot vinschen dot de> <87h9vtz254 dot fsf at Rainer dot invalid> <54B6DD5D dot 4060705 at cornell dot edu> <54C25113 dot 10200 at cornell dot edu> <54CA9721 dot 6090200 at cornell dot edu> <20150130094148 dot GO14265 at calimero dot vinschen dot de> <54CB865D dot 8040708 at cornell dot edu> <54CBB3FE dot 7030800 at cornell dot edu>
On 1/30/2015 11:40 AM, Ken Brown wrote:
On 1/30/2015 8:25 AM, Ken Brown wrote:
On 1/30/2015 4:41 AM, Corinna Vinschen wrote:
On Jan 29 15:25, Ken Brown wrote:
I'm attaching the patches that I applied (on top of Reini's patches) in
order to make the build succeed. I also had to use libdb4.5 instead of
libdb4.8.
Is that ok? I mean, shouldn't a package always try to use the latest
version? What's the problem you're observing ? Maybe Volker can help
here?
The clisp-2.48 code uses some macros and structures that are defined in the
db4.5 headers but have changed or disappeared in db4.8. I'll see if I can
figure out how to update the code for db4.8.
OK, that wasn't so hard. I'm attaching the patch that I applied, in case Reini
or Volker has a chance to look at it. All 76 tests for the berkeley-db module
passed.
I've now begun working on the 64-bit build. I found a few minor things I had to
change, and the build was pretty far along, when I apparently hit a gcc bug:
gcc -I/home/kbrown/src/cygclisp/clisp-2.48-5.x86_64/build/gllib -ggdb -O2 -pipe
-Wimplicit-function-declaration -W -Wswitch -Wcomment -Wpointer-arith -Wimplicit
-Wreturn-type -Wmissing-declarations -Wno-sign-compare -Wno-format-nonliteral -O
-falign-functions=4 -DUNICODE -I.
-I$(/home/kbrown/src/cygclisp/clisp-2.48-5.x86_64/build/clisp -K boot -B
/home/kbrown/src/cygclisp/clisp-2.48-5.x86_64/build -N
/home/kbrown/src/cygclisp/clisp-2.48-5.x86_64/build/locale -E UTF-8 -Epathname
1:1 -Emisc 1:1 -norc -q -b)/linkkit -c regexi.m.c -o regexi.o
regexi.c: In function ‘C_subr_regexp_regexp_compile’:
regexi.c:56:1: error: unrecognizable insn:
}
^
(insn 348 347 349 41 (set (reg/f:DI 259)
(plus:DI (symbol_ref:DI ("module__regexp__subr_tab") [flags 0x2]
<var_decl 0x6fffbfad360 module__regexp__subr_tab>)
(reg:DI 260))) regexi.c:54 -1
(expr_list:REG_EQUAL (const:DI (plus:DI (symbol_ref:DI
("module__regexp__subr_tab") [flags 0x2] <var_decl 0x6fffbfad360
module__regexp__subr_tab>)
(const_int 281474976710768 [0x1000000000070])))
(nil)))
regexi.c:56:1: internal compiler error: in extract_insn, at recog.c:2154
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.
This was initially with gcc-4.9.2-1, but the same thing happens with gcc-4.9.2-2
and gcc-4.8.3-2.
I'll poke around a little and see if I can find a problem in the source file,
but I may have to just report the gcc bug and put this aside until it's resolved.
Ken