Re: MAYLOSEDATA3 in VMS build [was: OpenSSL version 1.0.1g v. VMS]
Hi Richard, I could not find much more information about either, but on IA64 platform HP C V7.3-020 on OpenVMS IA64 V8.4 produces such warnings. Please check the following logs http://www.polarhome.com/openssl/BUILD100.LOG for 1.0.0m Please, advice how to solve this issue. Thank you, Z Quoting Richard Levitte : I just stumbled on problems with MAYLOSEDATA3 while looking at other stuff... and I wonder, what exactly is it? Sorry, I've been out of it for quite a while, and I haven't dug through all talks there may have been, so I may be missing something. Thing is, looking at the docs for the latest compiler (v7.3), there is no such warning message. There is MAYLOSEDATA and MAYLOSEDATA2, but no MAYLOSEDATA3. Is this a misunderstanding, a typo, or is there a secret newer HP CC that I don't know about? http://h71000.www7.hp.com/commercial/c/docs/decc_alpha_help_4.html#Ref504 In message <002401cf52e5$5791b8f0$06b52ad0$@com> on Tue, 8 Apr 2014 06:45:18 +0200, "Zoltan Arpadffy" said: zoli> Right approach Steven. zoli> zoli> Mea culpa :( zoli> I am terribly sorry proposing ignorance of MAYLOSEDATA3 while not testing on zoli> older compilers. zoli> zoli> Thank you. zoli> Regards, zoli> Z zoli> zoli> -Original Message- zoli> From: owner-openssl-...@openssl.org [mailto:owner-openssl-...@openssl.org] zoli> On Behalf Of Steven M. Schweda zoli> Sent: den 8 april 2014 05:12 zoli> To: openssl-dev@openssl.org zoli> Subject: Re: OpenSSL version 1.0.1g v. VMS zoli> zoli> From: "Dr. Stephen Henson" zoli> zoli> > Patch applied. Let me know of any problems. zoli> zoli>Thanks. Sadly, I forgot one refinement, which would keep the warning zoli> from the test out of the log. If I could talk you into smoothing over that zoli> blunder, then I'd be able to rest easy (for a while). zoli> zoli> --- ssl/ssl-lib.com;-1 2014-04-07 15:18:26 -0500 zoli> +++ ssl/ssl-lib.com2014-04-07 18:26:49 -0500 zoli> @@ -1024,6 +1024,8 @@ zoli> $ THEN zoli> $!Not all compiler versions support MAYLOSEDATA3. zoli> $ OPT_TEST = "MAYLOSEDATA3" zoli> +$ DEFINE /USER_MODE SYS$ERROR NL: zoli> +$ DEFINE /USER_MODE SYS$OUTPUT NL: zoli> $ 'CC' /NOCROSS_REFERENCE /NOLIST /NOOBJECT - zoli> /WARNINGS = DISABLE = ('OPT_TEST', EMPTYFILE) NL: zoli> $ IF ($SEVERITY) zoli> zoli>Sorry about the extra bother. zoli> zoli> zoli>For the record (no action required (or even requested)), the only zoli> remaining compiler warnings were: zoli> zoli> if (timeleft.tv_sec < 0) ^ zoli> %CC-I-QUESTCOMPARE, In this statement, the unsigned expression zoli> "timeleft.tv_sec" is being compared with a relational operator to a zoli> constant whose value is not greater than zero. This might not be what you zoli> intended. zoli> at line number 310 in file zoli> ALP$DKC100:[UTILITY.SOURCE.OPENSSL.openssl-1_0_1g.crypto.bio]bss_dgram.c;1 zoli> zoli>Around here, time_t tends to be unsigned. zoli> zoli> zoli> if (*outlen <= 0) zoli> ^ zoli> %CC-I-QUESTCOMPARE, In this statement, the unsigned expression "*outlen" zoli> is being compared with a relational operator to a constant whose value is zoli> not greater than zero. This might not be what you intended. zoli> at line number 180 in file zoli> ALP$DKC100:[UTILITY.SOURCE.OPENSSL.openssl-1_0_1g.engines.ccgost]gost94_keyx zoli> .c;1 zoli> zoli>Around here, size_t tends to be unsigned. zoli> zoli> zoli> if (size <= 0 || ((len = data[0])) != (size -1)) zoli> ^ %CC-I-QUESTCOMPARE, In this statement, the zoli> unsigned expression "size" is being compared with a relational operator to zoli> a constant whose value is not greater than zero. This might not be what zoli> you intended. zoli> at line number 1128 in file zoli> ALP$DKC100:[UTILITY.SOURCE.OPENSSL.openssl-1_0_1g.ssl]t1_lib.c;1 zoli> zoli>Here, "size" seems to be declared as unsigned: zoli> unsigned short size; zoli> making the "< 0" part of that test pointless everywhere. zoli> zoli>These are all Informational ("-I-") complaints, so they cause less zoli> trouble than a real warning ("-W-"). --- WebMail, polarhome.com __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager majord...@openssl.org
MAYLOSEDATA3 in VMS build [was: OpenSSL version 1.0.1g v. VMS]
I just stumbled on problems with MAYLOSEDATA3 while looking at other stuff... and I wonder, what exactly is it? Sorry, I've been out of it for quite a while, and I haven't dug through all talks there may have been, so I may be missing something. Thing is, looking at the docs for the latest compiler (v7.3), there is no such warning message. There is MAYLOSEDATA and MAYLOSEDATA2, but no MAYLOSEDATA3. Is this a misunderstanding, a typo, or is there a secret newer HP CC that I don't know about? http://h71000.www7.hp.com/commercial/c/docs/decc_alpha_help_4.html#Ref504 In message <002401cf52e5$5791b8f0$06b52ad0$@com> on Tue, 8 Apr 2014 06:45:18 +0200, "Zoltan Arpadffy" said: zoli> Right approach Steven. zoli> zoli> Mea culpa :( zoli> I am terribly sorry proposing ignorance of MAYLOSEDATA3 while not testing on zoli> older compilers. zoli> zoli> Thank you. zoli> Regards, zoli> Z zoli> zoli> -Original Message- zoli> From: owner-openssl-...@openssl.org [mailto:owner-openssl-...@openssl.org] zoli> On Behalf Of Steven M. Schweda zoli> Sent: den 8 april 2014 05:12 zoli> To: openssl-dev@openssl.org zoli> Subject: Re: OpenSSL version 1.0.1g v. VMS zoli> zoli> From: "Dr. Stephen Henson" zoli> zoli> > Patch applied. Let me know of any problems. zoli> zoli>Thanks. Sadly, I forgot one refinement, which would keep the warning zoli> from the test out of the log. If I could talk you into smoothing over that zoli> blunder, then I'd be able to rest easy (for a while). zoli> zoli> --- ssl/ssl-lib.com;-12014-04-07 15:18:26 -0500 zoli> +++ ssl/ssl-lib.com 2014-04-07 18:26:49 -0500 zoli> @@ -1024,6 +1024,8 @@ zoli> $ THEN zoli> $!Not all compiler versions support MAYLOSEDATA3. zoli> $ OPT_TEST = "MAYLOSEDATA3" zoli> +$ DEFINE /USER_MODE SYS$ERROR NL: zoli> +$ DEFINE /USER_MODE SYS$OUTPUT NL: zoli> $ 'CC' /NOCROSS_REFERENCE /NOLIST /NOOBJECT - zoli> /WARNINGS = DISABLE = ('OPT_TEST', EMPTYFILE) NL: zoli> $ IF ($SEVERITY) zoli> zoli>Sorry about the extra bother. zoli> zoli> zoli>For the record (no action required (or even requested)), the only zoli> remaining compiler warnings were: zoli> zoli> if (timeleft.tv_sec < 0) ^ zoli> %CC-I-QUESTCOMPARE, In this statement, the unsigned expression zoli> "timeleft.tv_sec" is being compared with a relational operator to a zoli> constant whose value is not greater than zero. This might not be what you zoli> intended. zoli> at line number 310 in file zoli> ALP$DKC100:[UTILITY.SOURCE.OPENSSL.openssl-1_0_1g.crypto.bio]bss_dgram.c;1 zoli> zoli>Around here, time_t tends to be unsigned. zoli> zoli> zoli> if (*outlen <= 0) zoli> ^ zoli> %CC-I-QUESTCOMPARE, In this statement, the unsigned expression "*outlen" zoli> is being compared with a relational operator to a constant whose value is zoli> not greater than zero. This might not be what you intended. zoli> at line number 180 in file zoli> ALP$DKC100:[UTILITY.SOURCE.OPENSSL.openssl-1_0_1g.engines.ccgost]gost94_keyx zoli> .c;1 zoli> zoli>Around here, size_t tends to be unsigned. zoli> zoli> zoli> if (size <= 0 || ((len = data[0])) != (size -1)) zoli> ^ %CC-I-QUESTCOMPARE, In this statement, the zoli> unsigned expression "size" is being compared with a relational operator to zoli> a constant whose value is not greater than zero. This might not be what zoli> you intended. zoli> at line number 1128 in file zoli> ALP$DKC100:[UTILITY.SOURCE.OPENSSL.openssl-1_0_1g.ssl]t1_lib.c;1 zoli> zoli>Here, "size" seems to be declared as unsigned: zoli> unsigned short size; zoli> making the "< 0" part of that test pointless everywhere. zoli> zoli>These are all Informational ("-I-") complaints, so they cause less zoli> trouble than a real warning ("-W-"). __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager majord...@openssl.org
RE: OpenSSL version 1.0.1g v. VMS
Right approach Steven. Mea culpa :( I am terribly sorry proposing ignorance of MAYLOSEDATA3 while not testing on older compilers. Thank you. Regards, Z -Original Message- From: owner-openssl-...@openssl.org [mailto:owner-openssl-...@openssl.org] On Behalf Of Steven M. Schweda Sent: den 8 april 2014 05:12 To: openssl-dev@openssl.org Subject: Re: OpenSSL version 1.0.1g v. VMS From: "Dr. Stephen Henson" > Patch applied. Let me know of any problems. Thanks. Sadly, I forgot one refinement, which would keep the warning from the test out of the log. If I could talk you into smoothing over that blunder, then I'd be able to rest easy (for a while). --- ssl/ssl-lib.com;-1 2014-04-07 15:18:26 -0500 +++ ssl/ssl-lib.com 2014-04-07 18:26:49 -0500 @@ -1024,6 +1024,8 @@ $ THEN $!Not all compiler versions support MAYLOSEDATA3. $ OPT_TEST = "MAYLOSEDATA3" +$ DEFINE /USER_MODE SYS$ERROR NL: +$ DEFINE /USER_MODE SYS$OUTPUT NL: $ 'CC' /NOCROSS_REFERENCE /NOLIST /NOOBJECT - /WARNINGS = DISABLE = ('OPT_TEST', EMPTYFILE) NL: $ IF ($SEVERITY) Sorry about the extra bother. For the record (no action required (or even requested)), the only remaining compiler warnings were: if (timeleft.tv_sec < 0) ^ %CC-I-QUESTCOMPARE, In this statement, the unsigned expression "timeleft.tv_sec" is being compared with a relational operator to a constant whose value is not greater than zero. This might not be what you intended. at line number 310 in file ALP$DKC100:[UTILITY.SOURCE.OPENSSL.openssl-1_0_1g.crypto.bio]bss_dgram.c;1 Around here, time_t tends to be unsigned. if (*outlen <= 0) ^ %CC-I-QUESTCOMPARE, In this statement, the unsigned expression "*outlen" is being compared with a relational operator to a constant whose value is not greater than zero. This might not be what you intended. at line number 180 in file ALP$DKC100:[UTILITY.SOURCE.OPENSSL.openssl-1_0_1g.engines.ccgost]gost94_keyx .c;1 Around here, size_t tends to be unsigned. if (size <= 0 || ((len = data[0])) != (size -1)) ^ %CC-I-QUESTCOMPARE, In this statement, the unsigned expression "size" is being compared with a relational operator to a constant whose value is not greater than zero. This might not be what you intended. at line number 1128 in file ALP$DKC100:[UTILITY.SOURCE.OPENSSL.openssl-1_0_1g.ssl]t1_lib.c;1 Here, "size" seems to be declared as unsigned: unsigned short size; making the "< 0" part of that test pointless everywhere. These are all Informational ("-I-") complaints, so they cause less trouble than a real warning ("-W-"). Steven M. Schweda sms@antinode-info 382 South Warwick Street(+1) 651-699-9818 Saint Paul MN 55105-2547 __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager majord...@openssl.org __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager majord...@openssl.org
Re: OpenSSL version 1.0.1g v. VMS
From: "Dr. Stephen Henson" > Patch applied. Let me know of any problems. Thanks. Sadly, I forgot one refinement, which would keep the warning from the test out of the log. If I could talk you into smoothing over that blunder, then I'd be able to rest easy (for a while). --- ssl/ssl-lib.com;-1 2014-04-07 15:18:26 -0500 +++ ssl/ssl-lib.com 2014-04-07 18:26:49 -0500 @@ -1024,6 +1024,8 @@ $ THEN $!Not all compiler versions support MAYLOSEDATA3. $ OPT_TEST = "MAYLOSEDATA3" +$ DEFINE /USER_MODE SYS$ERROR NL: +$ DEFINE /USER_MODE SYS$OUTPUT NL: $ 'CC' /NOCROSS_REFERENCE /NOLIST /NOOBJECT - /WARNINGS = DISABLE = ('OPT_TEST', EMPTYFILE) NL: $ IF ($SEVERITY) Sorry about the extra bother. For the record (no action required (or even requested)), the only remaining compiler warnings were: if (timeleft.tv_sec < 0) ^ %CC-I-QUESTCOMPARE, In this statement, the unsigned expression "timeleft.tv_sec" is being compared with a relational operator to a constant whose value is not greater than zero. This might not be what you intended. at line number 310 in file ALP$DKC100:[UTILITY.SOURCE.OPENSSL.openssl-1_0_1g.crypto.bio]bss_dgram.c;1 Around here, time_t tends to be unsigned. if (*outlen <= 0) ^ %CC-I-QUESTCOMPARE, In this statement, the unsigned expression "*outlen" is being compared with a relational operator to a constant whose value is not greater than zero. This might not be what you intended. at line number 180 in file ALP$DKC100:[UTILITY.SOURCE.OPENSSL.openssl-1_0_1g.engines.ccgost]gost94_keyx.c;1 Around here, size_t tends to be unsigned. if (size <= 0 || ((len = data[0])) != (size -1)) ^ %CC-I-QUESTCOMPARE, In this statement, the unsigned expression "size" is being compared with a relational operator to a constant whose value is not greater than zero. This might not be what you intended. at line number 1128 in file ALP$DKC100:[UTILITY.SOURCE.OPENSSL.openssl-1_0_1g.ssl]t1_lib.c;1 Here, "size" seems to be declared as unsigned: unsigned short size; making the "< 0" part of that test pointless everywhere. These are all Informational ("-I-") complaints, so they cause less trouble than a real warning ("-W-"). Steven M. Schweda sms@antinode-info 382 South Warwick Street(+1) 651-699-9818 Saint Paul MN 55105-2547 __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager majord...@openssl.org
Re: OpenSSL version 1.0.1g v. VMS
On Mon, Apr 07, 2014, Steven M. Schweda wrote: > >OpenSSL version 1.0.1g released > >A relatively recent change to the VMS builders causes trouble for > folks who lack the latest compiler. Using "HP C V7.3-009 on OpenVMS > Alpha V8.3" or "HP C V7.3-019 on OpenVMS IA64 V8.3-1H1", for example: > > [...] > s2_meth.c > > %CC-W-UNKMSGID, Unknown message id or group "MAYLOSEDATA3" is ignored. > %LIBRAR-W-COMCOD, compilation warnings in module S2_METH file > ALP$DKC100:[UTILITY.SOURCE.OPENSSL.openssl-1_0_1g.ALPHA.OBJ.SSL]s2_meth.OBJ;1 > [...] > %LINK-W-WRNERS, compilation warnings in module S2_METH file > ALP$DKC100:[UTILITY.SOURCE.OPENSSL.openssl-1_0_1g.ALPHA.EXE.SSL]SSL_LIBSSL32.OLB;1 > [...] > >I see about 48 of the %CC-W-UNKMSGID, 47 of the %LIBRAR-W-COMCOD, and > 157 of the %LINK-W-WRNERS. This is annoying. I propose a change to > suppress "MAYLOSEDATA3" warnings only if the compiler recognizes it: > Patch applied. Let me know of any problems. Steve. -- Dr Stephen N. Henson. OpenSSL project core developer. Commercial tech support now available see: http://www.openssl.org __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager majord...@openssl.org
OpenSSL version 1.0.1g v. VMS
>OpenSSL version 1.0.1g released A relatively recent change to the VMS builders causes trouble for folks who lack the latest compiler. Using "HP C V7.3-009 on OpenVMS Alpha V8.3" or "HP C V7.3-019 on OpenVMS IA64 V8.3-1H1", for example: [...] s2_meth.c %CC-W-UNKMSGID, Unknown message id or group "MAYLOSEDATA3" is ignored. %LIBRAR-W-COMCOD, compilation warnings in module S2_METH file ALP$DKC100:[UTILITY.SOURCE.OPENSSL.openssl-1_0_1g.ALPHA.OBJ.SSL]s2_meth.OBJ;1 [...] %LINK-W-WRNERS, compilation warnings in module S2_METH file ALP$DKC100:[UTILITY.SOURCE.OPENSSL.openssl-1_0_1g.ALPHA.EXE.SSL]SSL_LIBSSL32.OLB;1 [...] I see about 48 of the %CC-W-UNKMSGID, 47 of the %LIBRAR-W-COMCOD, and 157 of the %LINK-W-WRNERS. This is annoying. I propose a change to suppress "MAYLOSEDATA3" warnings only if the compiler recognizes it: --- ssl/ssl-lib.com_orig2014-03-17 11:14:20 -0500 +++ ssl/ssl-lib.com 2014-04-07 15:18:26 -0500 @@ -857,7 +857,7 @@ $ IF F$TYPE(USER_CCDEFS) .NES. "" THEN CCDEFS = CCDEFS + "," + USER_CCDEFS $ CCEXTRAFLAGS = "" $ IF F$TYPE(USER_CCFLAGS) .NES. "" THEN CCEXTRAFLAGS = USER_CCFLAGS -$ CCDISABLEWARNINGS = "MAYLOSEDATA3" !!! "LONGLONGTYPE,LONGLONGSUFX,FOUNDCR" +$ CCDISABLEWARNINGS = "" !!! "MAYLOSEDATA3" !!! "LONGLONGTYPE,LONGLONGSUFX,FOUNDCR" $ IF F$TYPE(USER_CCDISABLEWARNINGS) .NES. "" THEN - CCDISABLEWARNINGS = CCDISABLEWARNINGS + "," + USER_CCDISABLEWARNINGS $! @@ -1022,6 +1022,16 @@ $! $ IF COMPILER .EQS. "DECC" $ THEN +$!Not all compiler versions support MAYLOSEDATA3. +$ OPT_TEST = "MAYLOSEDATA3" +$ 'CC' /NOCROSS_REFERENCE /NOLIST /NOOBJECT - + /WARNINGS = DISABLE = ('OPT_TEST', EMPTYFILE) NL: +$ IF ($SEVERITY) +$ THEN +$ IF CCDISABLEWARNINGS .NES. "" THEN - + CCDISABLEWARNINGS = CCDISABLEWARNINGS+ "," +$ CCDISABLEWARNINGS = CCDISABLEWARNINGS+ OPT_TEST +$ ENDIF $ IF CCDISABLEWARNINGS .EQS. "" $ THEN $ CC4DISABLEWARNINGS = "DOLLARID" I don't have one, so I haven't actually tested this change using a compiler which accepts MAYLOSEDATA3, but it looks plausible (to me). 1.0.2-beta1 seems to need the same treatment in crypto/crypto-lib.com, as well. Steven M. Schweda sms@antinode-info 382 South Warwick Street(+1) 651-699-9818 Saint Paul MN 55105-2547 __ OpenSSL Project http://www.openssl.org Development Mailing List openssl-dev@openssl.org Automated List Manager majord...@openssl.org