Re: [openssl.org #1391] OpenSSL 0.9.8c build fails on old Linux
Hello gentlemen, On Sunday, September 17, 2006 at 0:04:24 +0200, Alain Guibert wrote: during OpenSSL 0.9.8c build, make fails: | bss_dgram.c: In function `dgram_ctrl': | bss_dgram.c:345: `SO_RCVTIMEO' undeclared (first use this function) | bss_dgram.c:355: `SO_SNDTIMEO' undeclared (first use this function) This is now nicely fixed in the OpenSSL 0.9.8i release. You can close this ticket. Much thanks to all the team! Alain. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [openssl.org #1667] 0.9.8-stable sha512 segfaults
Hello Andy, and thank you for your reply. On Tuesday, April 22, 2008 at 16:05:44 +0200, Andy Polyakov wrote: Could you collect disassembler output for offending code? I.e. run 'gdb test/sha512t core' (or rather util/shlib_wrap.sh gdb test/sha512t as you have build for shared), run 'where', note address, run 'disassemble' and advance to just noted address. Submit *complete* disassemble output from function start to offending instruction. Thank you for the step-by-step procedure, as I'm not familiar at all with gdb nor assembler. Disassembly of sha512_block_data_order() rebuilt non-shared -O3 comes attached as sha512_block_data_order.asm.gz. Drop optimization level... -O3 and -O2 segfault. -O1 and -O0 do work and test passes. Test attached patch. | $ test/sha512t | Testing SHA-512 ... passed. | Testing SHA-384 ... passed. Excellent! Your patch works at any optimization level. Much thanks to you, Andy. Alain. sha512_block_data_order.asm.gz Description: application/gunzip
[openssl.org #1667] 0.9.8-stable sha512 segfaults
Hello, During make test of latest CVS revision OpenSSL_0_9_8-stable, the sha512 test fails on a segmentation fault in sha512_block_data_order(): | ../util/shlib_wrap.sh ./sha512t | make[1]: *** [test_sha] Segmentation fault (core dumped) | make[1]: Leaving directory `/tmp/openssl-0.9.8-cvs/test' | make: *** [tests] Error 2 Note OpenSSL release 0.9.8g works OK. Output of make report comes attached as testlog.gz Alain. testlog.gz Description: application/gunzip
Re: [openssl.org #1452] OpenSSL 0.9.7l make install syntax error
Hello Andy, On Saturday, October 13, 2007 at 13:13:22 +0200, Andy Polyakov via RT wrote: this was eventually fixed. thanks for report. a. This syntax error problem is still present in latest 0.9.7m release, but is indeed fixed in the CVS 0.9.7 branch. And it never impacted the 0.9.8 branch nor HEAD. Much thanks to you! Alain. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [openssl.org #1392] OpenSSL 0.9.7k build fails missing poll.h
Hello, On Sunday, September 17, 2006 at 0:04:48 +0200, Alain Guibert via RT wrote: | rand_unix.c:128: poll.h: No such file or directory This problem touches OpenSSL 0.9.7[kl], 0.9.8[cd], and latest CVS HEAD. Here poll.h doesn't exist, but there is a sys/poll.h. And once rand_unix.c modified to use it, OpenSSL builds properly. Alain. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [openssl.org #1301] OpenSSL 0.9.7i shared build fails
Hello, On Friday, March 31, 2006 at 14:00:13 +0200, Alain Guibert via RT wrote: On my old Linux box (Intel Debian bo: kernel 2.0.40, gcc 2.7.2.1, libc 5.4.33) during OpenSSL 0.9.7i shared lib build, make fails: | [...snip hundreds of undefined references...] This problem touches all versions since 0.9.7h until latest CVS HEAD. Previous versions until 0.9.7g built correct shared libs. The difference between g and h is that in Makefile.org a -lc option was removed from shared build commands. For some reason I don't understand, only when building OpenSSL, my linker doesn't seem to catch the libc without an explicit -lc. Hence the tons of undefined refs. I added -lc to either the EX_LIBS or SHARED_LDFLAGS fields in the Configure table for linux-elf, and the errors are gone: Shared OpenSSL (all versions) now build OK on this old system. Alain. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [openssl.org #1048] OpenSSL 0.9.7g failing make test
Hello, On Wednesday, April 20, 2005 at 18:49:53 +0200, Alain Guibert wrote: during OpenSSL 0.9.7g build, make test fails: | ../util/shlib_wrap.sh ./destest | ./destest: can't load library '../util/../libcrypto.so.0.9.7 ../util/../libssl.so.0.9.7' These symptoms are the same with all OpenSSL versions since 0.9.7g, including all 0.9.8*, and latest CVS HEAD, for shared builds. The problem seems to be in util/shlib_wrap.sh, where LD_PRELOAD is set to both the just built libcrypto and libssl, separated by a space (for all platforms except BSD): | case $SYSNAME in | *BSD) LD_PRELOAD=$LIBCRYPTOSO:$LIBSSLSO ;; # *BSD | *) LD_PRELOAD=$LIBCRYPTOSO $LIBSSLSO ;; # SunOS, Linux, ELF HP-UX | esac However for some reason my old Linux system doesn't allow space as separator in LD_PRELOAD. It wants a colon, just like BSD. And then make test of a shared OpenSSL becomes succesfull. I suppose ld.so is responsible for this LD_PRELOAD space or colon behaviour: Here it's version 1.8.10. Alain. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
[openssl.org #1452] OpenSSL 0.9.7l make install syntax error
Hello, When doing make install in OpenSSL 0.9.7l, there is a syntax error, some apps are not installed, but Makefile still continues: | # make install | [...] | installing apps... | make[1]: Entering directory `/tmp/openssl-0.9.7l/apps' | /bin/sh: -c: line 1: syntax error near unexpected token `;' | /bin/sh: -c: line 1: `for i in openssl; do (echo installing $i; cp $i /usr/local/bin/$i.new; chmod 755 /usr/local/bin/$i.new; mv -f /usr/local/bin/$i.new /usr/local/bin/$i; ) done;' | make[1]: *** [install] Error 2 | make[1]: Leaving directory `/tmp/openssl-0.9.7l/apps' | installing test... | make[1]: Entering directory `/tmp/openssl-0.9.7l/test' | make[1]: Nothing to be done for `install'. | make[1]: Leaving directory `/tmp/openssl-0.9.7l/test' | installing tools... | make[1]: Entering directory `/tmp/openssl-0.9.7l/tools' | make[1]: Leaving directory `/tmp/openssl-0.9.7l/tools' | installing libcrypto.a | installing libssl.a | cp openssl.pc /tmp/ossl/lib/pkgconfig | chmod 644 /tmp/ossl/lib/pkgconfig/openssl.pc Syntax error seems to be a misplaced semicolon in shell syntax inside apps/Makefile. Schematically the bad: | for i in list; do (commands) done; Instead of the correct: | for i in list; do (commands); done The attached patch fixes that. Problem seen in 0.9.7[hijkl], absent in 0.9.7g and before. Problem absent in 0.9.8*, but there is still a useless semicolon after the done. Alain. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
[openssl.org #1391] OpenSSL 0.9.8c build fails on old Linux
Hello, On my old Linux box (Intel Pentium 200 MMX, Debian bo, kernel 2.0.40, gcc 2.7.2.1, libc 5.4.33) during OpenSSL 0.9.8c build, make fails: | $ ./config no-sse2 no-hw; make | [...] | gcc -I.. -I../.. -I../../include -DOPENSSL_THREADS -D_REENTRANT -DDSO_DLFCN -DHAVE_DLFCN_H -DL_ENDIAN -DTERMIO -O3 -fomit-frame-pointer -Wall -DOPENSSL_BN_ASM_PART_WORDS -DSHA1_ASM -DMD5_ASM -DRMD160_ASM -DAES_ASM -c bss_dgram.c -o bss_dgram.o | bss_dgram.c: In function `dgram_ctrl': | bss_dgram.c:345: `SO_RCVTIMEO' undeclared (first use this function) | bss_dgram.c:345: (Each undeclared identifier is reported only once | bss_dgram.c:345: for each function it appears in.) | bss_dgram.c:355: `SO_SNDTIMEO' undeclared (first use this function) | make[2]: *** [bss_dgram.o] Error 1 | make[2]: Leaving directory `/tmp/openssl-0.9.8c/crypto/bio' | make[1]: *** [subdirs] Error 1 | make[1]: Leaving directory `/tmp/openssl-0.9.8c/crypto' | make: *** [build_crypto] Error 1 Alain. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
[openssl.org #1392] OpenSSL 0.9.7k build fails missing poll.h
On my old Linux box (Intel Pentium 200 MMX, Debian bo, kernel 2.0.40, gcc 2.7.2.1, libc 5.4.33) during OpenSSL 0.9.7k build, make fails: | $ ./config; make | gcc -I.. -I../.. -I../../include -DOPENSSL_THREADS -D_REENTRANT -DDSO_DLFCN -DHAVE_DLFCN_H -DOPENSSL_NO_KRB5 -DL_ENDIAN -DTERMIO -O3 -fomit-frame-pointer -m486 -Wall -DSHA1_ASM -DMD5_ASM -DRMD160_ASM -c rand_unix.c -o rand_unix.o | rand_unix.c:128: poll.h: No such file or directory | make[2]: *** [rand_unix.o] Error 1 | make[2]: Leaving directory `/tmp/openssl-0.9.7k/crypto/rand' | make[1]: *** [subdirs] Error 1 | make[1]: Leaving directory `/tmp/openssl-0.9.7k/crypto' | make: *** [sub_all] Error 1 Alain. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
[openssl.org #1303] OpenSSL 0.9.8a build fails
Hello, On my old Linux box (Intel Pentium 200 MMX, Debian bo, kernel 2.0.40, gcc 2.7.2.1, libc 5.4.33) during OpenSSL 0.9.8a build, make fails: | $ ./config; make | [...] | gcc -I. -I.. -I../include -DOPENSSL_THREADS -D_REENTRANT -DDSO_DLFCN -DHAVE_DLFCN_H -DL_ENDIAN -DTERMIO -O3 -fomit-frame-pointer -Wall -DOPENSSL_BN_ASM_PART_WORDS -DOPENSSL_IA32_SSE2 -DSHA1_ASM -DMD5_ASM -DRMD160_ASM -DAES_ASM -c -o x86cpuid-elf.o x86cpuid-elf.s | x86cpuid-elf.s: Assembler messages: | x86cpuid-elf.s:144: Error: bad register name ('%xmm0') | x86cpuid-elf.s:145: Error: bad register name ('%xmm1') | x86cpuid-elf.s:146: Error: bad register name ('%xmm2') | x86cpuid-elf.s:147: Error: bad register name ('%xmm3') | x86cpuid-elf.s:148: Error: bad register name ('%xmm4') | x86cpuid-elf.s:149: Error: bad register name ('%xmm5') | x86cpuid-elf.s:150: Error: bad register name ('%xmm6') | x86cpuid-elf.s:151: Error: bad register name ('%xmm7') | make[1]: *** [x86cpuid-elf.o] Error 1 | make[1]: Leaving directory `/tmp/openssl-0.9.8a/crypto' | make: *** [build_crypto] Error 1 Output of make report comes attached as testlog1.gz. Then I tried with no-sse2 option, also failing later: | $ ./config no-sse2; make | [...] | gcc -I.. -I../.. -I../../include -DOPENSSL_THREADS -D_REENTRANT -DDSO_DLFCN -DHAVE_DLFCN_H -DL_ENDIAN -DTERMIO -O3 -fomit-frame-pointer -Wall -DOPENSSL_BN_ASM_PART_WORDS -DSHA1_ASM -DMD5_ASM -DRMD160_ASM -DAES_ASM -c eng_padlock.c -o eng_padlock.o | eng_padlock.c: In function `padlock_available': | eng_padlock.c:346: output operand constraint contains `+' | eng_padlock.c:353: output operand constraint contains `+' | eng_padlock.c:360: output operand constraint contains `+' | eng_padlock.c: In function `padlock_bswapl': | eng_padlock.c:378: output operand constraint contains `+' | eng_padlock.c: In function `padlock_verify_context': | eng_padlock.c:417: output operand constraint contains `+' | make[2]: *** [eng_padlock.o] Error 1 | make[2]: Leaving directory `/tmp/openssl-0.9.8a/crypto/engine' | make[1]: *** [subdirs] Error 1 | make[1]: Leaving directory `/tmp/openssl-0.9.8a/crypto' | make: *** [build_crypto] Error 1 Output of make report comes attached as testlog.no-sse2.gz While at it: The file INSTALL says that zlib-dynamic is the default choice, but I seem to get no-zlib-dynamic by default. Alain. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
[openssl.org #1301] OpenSSL 0.9.7i shared build fails
Hello, On my old Linux box (Intel Debian bo: kernel 2.0.40, gcc 2.7.2.1, libc 5.4.33) during OpenSSL 0.9.7i shared lib build, make fails: | $ ./config shared ; make | [...snip hundreds of undefined references...] | ./libcrypto.a(ui_openssl.o): In function `close_console': | ui_openssl.o(.text+0x5a5): undefined reference to `_IO_stdin_' | ui_openssl.o(.text+0x5ad): undefined reference to `fclose' | ui_openssl.o(.text+0x5bc): undefined reference to `_IO_stderr_' | ui_openssl.o(.text+0x5c4): undefined reference to `fclose' | ./libcrypto.a(ui_openssl.o): In function `pushsig': | ui_openssl.o(.text+0x60c): undefined reference to `memset' | ui_openssl.o(.text+0x646): undefined reference to `sigaction' | ui_openssl.o(.text+0x65b): undefined reference to `signal' | ./libcrypto.a(ui_openssl.o): In function `popsig': | ui_openssl.o(.text+0x6a2): undefined reference to `sigaction' | make[3]: *** [do_gnu-shared] Error 1 | make[3]: Leaving directory `/tmp/openssl-0.9.7i' | make[2]: *** [libssl.so.0.9.7] Error 2 | make[2]: Leaving directory `/tmp/openssl-0.9.7i' | make[1]: *** [shared] Error 2 | make[1]: Leaving directory `/tmp/openssl-0.9.7i/ssl' | make: *** [sub_all] Error 1 Output of make report comes attached. Note non-shared build succeeds and make test then passes OK. BTW: How to change reporter's email address of old reports on Request Tracker? I want to change reporter of #201 and #1048 to my present email, but failed. Alain. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
[openssl.org #1048] OpenSSL 0.9.7g failing make test
Hello, On my old Linux box (Intel Debian bo: kernel 2.0.40, gcc 2.7.2.1, libc 5.4.33) during OpenSSL 0.9.7g build, make test fails: | Doing certs | /tmp/openssl-0.9.7g/util/../apps/openssl: can't load library '/tmp/openssl-0.9.7g/util/../libcrypto.so.0.9.7 /tmp/openssl-0.9.7g/util/../libssl.so.0.9.7' | eng1.pem = .0 | /tmp/openssl-0.9.7g/util/../apps/openssl: can't load library '/tmp/openssl-0.9.7g/util/../libcrypto.so.0.9.7 /tmp/openssl-0.9.7g/util/../libssl.so.0.9.7' | WARNING: Skipping duplicate certificate eng2.pem | /tmp/openssl-0.9.7g/util/../apps/openssl: can't load library '/tmp/openssl-0.9.7g/util/../libcrypto.so.0.9.7 /tmp/openssl-0.9.7g/util/../libssl.so.0.9.7' | WARNING: Skipping duplicate certificate eng3.pem [...snip...] | touch rehash.time | testing... | make[1]: Entering directory `/tmp/openssl-0.9.7g/test' | make[2]: Entering directory `/tmp/openssl-0.9.7g' | making all in apps... | make[3]: Entering directory `/tmp/openssl-0.9.7g/apps' | make[3]: Nothing to be done for `all'. | make[3]: Leaving directory `/tmp/openssl-0.9.7g/apps' | make[2]: Leaving directory `/tmp/openssl-0.9.7g' | ../util/shlib_wrap.sh ./destest | ./destest: can't load library '../util/../libcrypto.so.0.9.7 ../util/../libssl.so.0.9.7' | make[1]: *** [test_des] Error 15 | make[1]: Leaving directory `/tmp/openssl-0.9.7g/test' | make: *** [tests] Error 2 No problem with 0.9.7e. Output of make report comes attached. Alain. __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager [EMAIL PROTECTED]
Re: [openssl.org #201] OpenSSL 0.9.6e failing make test
On Monday, December 2, 2002 at 10:55:26 AM +0100, Alain Guibert via RT wrote: Tried OpenSSL 0.9.7 beta4 and it seems to work. Both make test and make install passes. Great! Thanks. I'll confirm with 0.9.7 when it's released, but please feel free to close the bug. Confirmation for the just released 0.9.6h (not engine): Everything is OK, test and install pass. Thanks again! :-) Side note: I sent on 1 Dec 2002 at 21:42:47 a mail (uninteresting: basically a question I answered myself in next mail) that got lost. Not received back from Request Tracker, no delivery bounce, and not reached URL:http://www.aet.TU-Cottbus.DE/rt2/Ticket/Display.html?id=201. It was id [EMAIL PROTECTED]. Bye!Alain. __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [openssl.org #201] OpenSSL 0.9.6e failing make test
Hello again Lutz, On Friday, November 29, 2002 at 1:20:25 PM +0100, Lutz Jänicke via RT wrote: On Fri, Nov 29, 2002 at 12:59:29PM +0100, Alain Guibert via RT wrote: - with 0.9.6g: installing man3/SSL_write.3 /usr/bin/pod2man: Invalid man page - 1st pod line is not NAME in SSL_write.pod make: *** [install_docs] Error 29 Hmm. Seems your version of pod2man has some problems, that were not detected. util/pod2mantest has been improved since the release of 0.9.6g. The version of pod2man.pl shipping with OpenSSL should definitely work. Tried OpenSSL 0.9.7 beta4 and it seems to work. Both make test and make install passes. Great! Thanks. I'll confirm with 0.9.7 when it's released, but please feel free to close the bug. Installing manpages gives: | /usr/bin/pod2man: Invalid man page - 1st pod line is not NAME in pod2mantest.pod | pod2man does not work properly ('BasicTest' failed). Looking for another pod2man ... | /usr/bin/pod2man: Invalid man page - 1st pod line is not NAME in pod2mantest.pod | /usr/bin/pod2man does not work properly ('BasicTest' failed). Looking for another |pod2man ... | No working pod2man found. Consider installing a new version. | As a workaround, we'll use a bundled old copy of pod2man.pl. | installing man1/CA.pl.1 | installing man1/asn1parse.1 [and so on] There is just an alert at some point: | installing man3/engine.3 | ../../util/pod2man.pl: Unrecognized pod directive in paragraph 34 of engine.pod: |head3 | ../../util/pod2man.pl: Unrecognized pod directive in paragraph 39 of engine.pod: |head3 | ../../util/pod2man.pl: Unrecognized pod directive in paragraph 59 of engine.pod: |head3 | ../../util/pod2man.pl: Unrecognized pod directive in paragraph 62 of engine.pod: |head3 | ../../util/pod2man.pl: Unrecognized pod directive in paragraph 69 of engine.pod: |head3 | ../../util/pod2man.pl: Unrecognized pod directive in paragraph 73 of engine.pod: |head3 But the installation continues, and man engine is there after. Bye!Alain. __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [openssl.org #201] OpenSSL 0.9.6e failing make test
Hello Lutz, On Thursday, November 28, 2002 at 8:37:36 PM +0100, Lutz Jaenicke via RT wrote: [jaenicke - Fri Nov 15 22:52:05 2002]: The test_req script actually performs a ../apps/openssl req -config ../apps/openssl.cnf testreq.pem \ -inform p -outform d f.d Can you run this manually? It fails with unknown option testreq.pem message. But if I replace argument testreq.pem by an input redirection testreq.pem it succeeds with a 326 bytes binary output f.d file. It succeeds today, but was perhaps failing before: Situation has changed. Any updates? Well, I'm sorry, don't know why: make test works now. With all versions previously failing: 0.9.6d, e and g. And the exact same build commands as before. So it obviously was something broken on my system... The only important change I made was cleanly installing a working 0.9.6c some weeks ago. Appart of that, I didn't add nor upgrade anything in build tools (gcc, make and friends), nor any system libraries (well no: I upgraded libiconv 1.8, Ncurses 5.3, and libbz2 1.0.2). Note that with this install of 0.9.6c I successfully experimented playing with CA.pl to create a CA then a server certificate. This was failing before, with test failing 0.9.6d, with errors I didn't note (I thought I'd make later a complete report but lacked time), unless I removed nearly everything from openssl.cnf as nsComment, subjectKeyIdentifier, authorityKeyIdentifier, basicConstraints, keyUsage, etc... From my poor memory, errors were something as unknown object type. Today it's working, with test succeeding 0.9.6d. I wanted to test with 0.9.6e/g, but if they now pass make and make test, they fail make install (?!?) while installing manpages: - with 0.9.6e: [snip many same errors] installing man1/verify.1 /usr/bin/pod2man: Invalid man page - 1st pod line is not NAME in pod2mantest.pod /usr/bin/pod2man: Invalid man page - 1st pod line is not NAME in verify.pod installing man1/version.1 /usr/bin/pod2man: Invalid man page - 1st pod line is not NAME in pod2mantest.pod /usr/bin/pod2man: Invalid man page - 1st pod line is not NAME in version.pod installing man1/x509.1 /usr/bin/pod2man: Invalid man page - 1st pod line is not NAME in pod2mantest.pod /usr/bin/pod2man: Invalid man page - 1st pod line is not NAME in x509.pod make: *** [install_docs] Error 29 - with 0.9.6g: [snip many same errors] installing man3/SSL_state_string.3 /usr/bin/pod2man: Invalid man page - 1st pod line is not NAME in SSL_state_string.pod installing man3/SSL_want.3 /usr/bin/pod2man: Invalid man page - 1st pod line is not NAME in SSL_want.pod installing man3/SSL_write.3 /usr/bin/pod2man: Invalid man page - 1st pod line is not NAME in SSL_write.pod make: *** [install_docs] Error 29 In both cases, I end without openssl installed. Summary: - 0.9.6c always tested and installed OK. - before 0.9.6d/e/g tested BAD. - now 0.9.6d tests and installs OK. - now 0.9.6e/g test OK but install BAD. So I now upgraded my system to 0.9.6d, the latest working version. Hope this makes sense... Thanks again, bye! Alain. __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]
Re: [openssl.org #201] OpenSSL 0.9.6e failing make test
Hello Richard, On Thursday, August 15, 2002 at 12:36:52 PM +0200, Richard Levitte via RT wrote: On my old Linux box (Intel Debian bo: kernel 2.0.39, gcc 2.7.2.1, libc 5.4.33) during OpenSSL 0.9.6e build a test in make test fails: | testing req conversions | p - d | make[1]: *** [test_req] Error 1 | make: *** [tests] Error 2 The only thing I can think of is that your disk is full Well no, don't think so: I have more than 80 Mo free on /tmp when I start make test. or test/testreq.pem is corrupted in some way. It's coming from a fresh tar.gz just uncompressed witout problems. BTW I get the same error with 0.9.6g, but not with older 0.9.6c. So same error with d, e, and g. Alain. __ OpenSSL Project http://www.openssl.org Development Mailing List [EMAIL PROTECTED] Automated List Manager [EMAIL PROTECTED]