Re: [openssl-dev] [openssl.org #4255] OpenSSL-1.1.0-pre2 failures using MinGW-W64
On 28/05/16 16:49, sav...@ukr.net wrote: > > > --- Исходное сообщение --- > От кого: "Matt Caswell" > Дата: 27 мая 2016, 18:03:50 > > > 2. Results for some tests using MSVC there are: > > > > skipped: TLSProxy isn't usable on MSWin32 > > > > Since PR: > > Fix intermittent windows failures in TLSProxy tests > > > https://github.com/openssl/openssl/commit/cb2e10f257a464c6b475b321dd9e4769df84dbf6 > > > > already pushed to master, does it make sense to enable that tests for > > MSVC builds? I can open such issue on GitHub, if needed. > > No, I already tried it. There are yet more issues which only seem to > affect the MSVC build that I haven't tracked down yet. > > > Just tested OpenSSL build using Mingw-w64, and all test passes without > looping. But some of test failed during run. > > Could you tell, is this known/expected behavior and is there any issue > related to it? No - it is not expected. On my machine all tests are passing consistently. Which tests are failing? Is it consistently always failing in the same place? Please raise this as a new issue. Matt -- openssl-dev mailing list To unsubscribe: https://mta.openssl.org/mailman/listinfo/openssl-dev
Re: [openssl-dev] [openssl.org #4255] OpenSSL-1.1.0-pre2 failures using MinGW-W64
On 27/05/16 15:58, sav...@ukr.net wrote: > > > --- Исходное сообщение --- > От кого: "Matt Caswell via RT" > Дата: 27 мая 2016, 17:45:28 > > The "make test" hang issue on mingw should now be resolved in the head of > master. Unfortunately there is now a completely different issue preventing > compilation for mingw :-( That is nothing to do with this ticket though so > closing. > > Hi, > > Thank you for fixing that issue. I'll test it when Mingw-w64 builds > would be restored. > > Got few questions, if you don't mind: > > 1. Which issue or PR I should look to figure out, whether Mingw builds > are available again. There isn't an existing ticket for it but Andy has a commit entitled "crypto/o_str.c: strerror_s is provided by specific compiler run-time," which is currently in internal review that fixes the issue. > > 2. Results for some tests using MSVC there are: > > skipped: TLSProxy isn't usable on MSWin32 > > Since PR: > Fix intermittent windows failures in TLSProxy tests > https://github.com/openssl/openssl/commit/cb2e10f257a464c6b475b321dd9e4769df84dbf6 > > already pushed to master, does it make sense to enable that tests for > MSVC builds? I can open such issue on GitHub, if needed. No, I already tried it. There are yet more issues which only seem to affect the MSVC build that I haven't tracked down yet. Matt -- openssl-dev mailing list To unsubscribe: https://mta.openssl.org/mailman/listinfo/openssl-dev
[openssl-dev] [openssl.org #4255] OpenSSL-1.1.0-pre2 failures using MinGW-W64
On Tue May 10 12:36:40 2016, matt wrote: > Re-opening. OP reports there are still issues with "make test" hanging. The "make test" hang issue on mingw should now be resolved in the head of master. Unfortunately there is now a completely different issue preventing compilation for mingw :-( That is nothing to do with this ticket though so closing. Matt -- Ticket here: http://rt.openssl.org/Ticket/Display.html?id=4255 Please log in as guest with password guest if prompted -- openssl-dev mailing list To unsubscribe: https://mta.openssl.org/mailman/listinfo/openssl-dev
[openssl-dev] [openssl.org #4255] OpenSSL-1.1.0-pre2 failures using MinGW-W64
On Tue May 10 14:57:43 2016, sav...@ukr.net wrote: > > Hi, Everyone > > Updates sources from master branch and made OpenSSL build and test > using command: > (./Configure --prefix="$PWD/../GCC64r" -std=gnu11 -mtune=native > -march=native threads no-deprecated shared mingw64 >_Configure.log && > make -j > $NUMBER_OF_PROCESSORS >_make.log && make -j$NUMBER_OF_PROCESSORS > install >_install.log) 2>_stderr.log && HARNESS_VERBOSE=yes make test > >_test.log 2>&1 > > Logs added in attachment. > > On some stage of OpenSSL testsuite run, process > C:\libOPENSSL-1.1.0-dev\src\apps\openssl.exe s_server -no_comp -rev > -engine ossltest -accept 4443 -cert ../apps/server.pem -naccept 1 > -cipher AES128-SHA > stop respond over 5 minutes and was terminated manually. > > Same repeated a bit later with processes > C:\libOPENSSL-1.1.0-dev\src\apps\openssl.exe s_client -engine ossltest > -connect 127.0.0.1:4453 -4 > C:\libOPENSSL-1.1.0-dev\src\apps\openssl.exe s_client -engine ossltest > -connect [::1]:4453 -6 > After their termination OpenSSL testsuite stop execution. > > Could you tell, whether this errors reproduced on other Windows > systems? (mean not using cross-compilation or Wine) Ah, I see that it's the tests based on TLSProxy that are troublesome... we have seen this before, and I can't remember if this was resolved. Matt, I think this is more your department -- Richard Levitte levi...@openssl.org -- Ticket here: http://rt.openssl.org/Ticket/Display.html?id=4255 Please log in as guest with password guest if prompted -- openssl-dev mailing list To unsubscribe: https://mta.openssl.org/mailman/listinfo/openssl-dev
[openssl-dev] [openssl.org #4255] OpenSSL-1.1.0-pre2 failures using MinGW-W64
On Tue May 10 12:36:40 2016, matt wrote: > Re-opening. OP reports there are still issues with "make test" hanging. I would like OP to submit a 'make test' log, preferably verbose: HARNESS_VERBOSE=yes make test -- Richard Levitte levi...@openssl.org -- Ticket here: http://rt.openssl.org/Ticket/Display.html?id=4255 Please log in as guest with password guest if prompted -- openssl-dev mailing list To unsubscribe: https://mta.openssl.org/mailman/listinfo/openssl-dev
[openssl-dev] [openssl.org #4255] OpenSSL-1.1.0-pre2 failures using MinGW-W64
Re-opening. OP reports there are still issues with "make test" hanging. Matt -- Ticket here: http://rt.openssl.org/Ticket/Display.html?id=4255 Please log in as guest with password guest if prompted -- openssl-dev mailing list To unsubscribe: https://mta.openssl.org/mailman/listinfo/openssl-dev
[openssl-dev] [openssl.org #4255] OpenSSL-1.1.0-pre2 failures using MinGW-W64
This doesn't seem to be a problem with OpenSSL so closing. Matt -- Ticket here: http://rt.openssl.org/Ticket/Display.html?id=4255 Please log in as guest with password guest if prompted -- openssl-dev mailing list To unsubscribe: https://mta.openssl.org/mailman/listinfo/openssl-dev
[openssl-dev] [openssl.org #4255] OpenSSL-1.1.0-pre2 failures using MinGW-W64
Hi, Got suggestion from Viktor Szakats (https://github.com/vszakats) concerning OpenSSL build errors using MinGW-W64 with -std=c11 parameter: == [snip] The error details indicate that some source code is not compliant with the C11 standard and therefore fail to compile in this mode. It doesn't seem to be gcc/mingw/Windows/64-bit specific: In file included from md4_locl.h:87:0, from md4_dgst.c:62: md4_dgst.c: In function 'md4_block_data_order': ../md32_common.h:167:33: warning: implicit declaration of function 'asm' [-Wimplicit-function-declaration] asm ( \ ^ md4_locl.h:105:11: note: in expansion of macro 'ROTATE' Try with -std=gnu11 instead to leave some optional C extensions enabled, like inline asm support. Current OpenSSL code seems to require it. [snip] I wouldn't call it an error per se. Fully standard compliant modes are rarely used in real-life projects, and I assume this was not a goal when creating this piece of code. Speaking of this specific error, IMO it's also generally cleaner not to mix languages inside the same source file. I can't see the full scope of this in case of OpenSSL, but it probably be useful to report this and suggest to move assembly code into separate .s files. Such practice is already used in quite a few places, so this exception may well be an oversight or heritage waiting to be updated. == Regards, Alexander - http://rt.openssl.org/Ticket/Display.html?id=4255 Please log in as guest with password guest if prompted ___ openssl-dev mailing list To unsubscribe: https://mta.openssl.org/mailman/listinfo/openssl-dev
Re: [openssl-dev] [openssl.org #4255] OpenSSL-1.1.0-pre2 failures using MinGW-W64
In message on Tue, 19 Jan 2016 14:22:31 +, "sav...@ukr.net via RT" said: > 1. If add parameter '-std=C11' (or '-std=C89') to 'cflags' variable in > 'mingw' and 'mingw64' sections of file 'OpenSSL/Configurations/10-main.conf', > and built OpenSSL with commands: Do you know that you can do that on the configuration command line as well? Like this: ./Configure ming2 -std=C11 rt> got error: rt> === rt> : recipe for target 'md4_dgst.o]' failed rt> make[2]: *** [md4_dgst.o] Error 1 rt> === rt> rt> Is it a flaw, or standard behavior? (e rror inherent to OpenSSL-1.0.2e too) That means you need to look above, at the output from the compile line of md4_dgst.o. Those line that you show is are from make, basically complaining that the compiler (or whatever program that reports an error) was complaining. rt> 2. Command 'make test' finishes with errors for + configurations (logs added in attachment). While shared builds freeze on some step and should be terminated manually. rt> Is it a flaw, or standard behavior on Windows? You can find more logs in test/*.log, one file for each test. -- Richard Levitte levi...@openssl.org OpenSSL Project http://www.openssl.org/~levitte/ ___ openssl-dev mailing list To unsubscribe: https://mta.openssl.org/mailman/listinfo/openssl-dev