Re: [openssl-project] Milestones and the 1.1.1 release

2018-06-26 Thread Matt Caswell


On 26/06/18 20:43, Salz, Rich wrote:
> That's interesting.  Would we put a bugfix in 1.1.0, not put the fix in 1.1.1 
> until our first "a" release?
> 
> Or are you saying that if it's in 1.1.0, then we don't have to fix it until 
> after 1.1.1 comes out?  That seems justifiable to me.

The latter.

I mean it doesn't *prevent* us from fixing something that's in both
1.1.0 and 1.1.1 - but our focus should be on fixing issues that are
newly introduced in 1.1.1.

Matt

> 
> On 6/26/18, 3:32 PM, "Matt Caswell"  wrote:
> 
> 
> 
> On 26/06/18 18:18, Salz, Rich wrote:
> > So are you saying look at the 73 open issues at 
> https://github.com/openssl/openssl/milestone/9 and re-evaluate them?
> 
> Exactly. My guess is that a significant proportion of them also apply to
> 1.1.0 and therefore should not hold up the 1.1.1 release. At the moment
> though it is impossible to tell which are the high priority issues we
> should be focussing on.
> 
> Matt
> 
> 
> > 
> > 
> > 
> > On 6/26/18, 11:56 AM, "Matt Caswell"  wrote:
> > 
> > I'm thinking that we should maybe re-asses the current milestones 
> in github.
> > 
> > We currently use the following milestones:
> > 
> > Assessed - Anything against this milestone isn't relevant to the 
> 1.1.1
> > release (e.g. 1.0.2 specific issue)
> > 
> > 1.1.1 - This is relevant to the 1.1.1 release but may not be 
> specific to
> > it (e.g. an issue that affects both 1.1.1 and 1.1.0)
> > 
> > Post 1.1.1 - Feature request to be looked at once 1.1.1 is released
> > 
> > 
> > I think we should re-asses everything currently against the 1.1.1
> > milestone so that anything which isn't specific to that release gets
> > moved to the "Assessed" milestone.
> > 
> > At the moment its difficult to see the "wood for the trees" between
> > issues which are newly introduced and those that are long standing. 
> In
> > terms of getting the 1.1.1 release out the door we should focus on 
> the
> > former.
> > 
> > Thoughts?
> > 
> > Matt
> > 
> > ___
> > openssl-project mailing list
> > openssl-project@openssl.org
> > https://mta.openssl.org/mailman/listinfo/openssl-project
> > 
> > 
> > ___
> > openssl-project mailing list
> > openssl-project@openssl.org
> > https://mta.openssl.org/mailman/listinfo/openssl-project
> > 
> ___
> openssl-project mailing list
> openssl-project@openssl.org
> https://mta.openssl.org/mailman/listinfo/openssl-project
> 
> ___
> openssl-project mailing list
> openssl-project@openssl.org
> https://mta.openssl.org/mailman/listinfo/openssl-project
> 
___
openssl-project mailing list
openssl-project@openssl.org
https://mta.openssl.org/mailman/listinfo/openssl-project

Re: [openssl-project] Milestones and the 1.1.1 release

2018-06-26 Thread Benjamin Kaduk
On Tue, Jun 26, 2018 at 07:43:45PM +, Salz, Rich wrote:
> That's interesting.  Would we put a bugfix in 1.1.0, not put the fix in 1.1.1 
> until our first "a" release?
> 
> Or are you saying that if it's in 1.1.0, then we don't have to fix it until 
> after 1.1.1 comes out?  That seems justifiable to me.

I assume the latter -- we feel obligated to fix regressions from 1.1.0 to
1.1.1 before finalizing 1.1.1, but bugs that are present in 1.1.0 can be
present in the 1.1.1 initial release (to be fixed in 1.1.1a and 1.1.0next).
(This is what I do for OpenAFS.)

-Ben
___
openssl-project mailing list
openssl-project@openssl.org
https://mta.openssl.org/mailman/listinfo/openssl-project


Re: [openssl-project] Milestones and the 1.1.1 release

2018-06-26 Thread Matt Caswell


On 26/06/18 18:18, Salz, Rich wrote:
> So are you saying look at the 73 open issues at 
> https://github.com/openssl/openssl/milestone/9 and re-evaluate them?

Exactly. My guess is that a significant proportion of them also apply to
1.1.0 and therefore should not hold up the 1.1.1 release. At the moment
though it is impossible to tell which are the high priority issues we
should be focussing on.

Matt


> 
> 
> 
> On 6/26/18, 11:56 AM, "Matt Caswell"  wrote:
> 
> I'm thinking that we should maybe re-asses the current milestones in 
> github.
> 
> We currently use the following milestones:
> 
> Assessed - Anything against this milestone isn't relevant to the 1.1.1
> release (e.g. 1.0.2 specific issue)
> 
> 1.1.1 - This is relevant to the 1.1.1 release but may not be specific to
> it (e.g. an issue that affects both 1.1.1 and 1.1.0)
> 
> Post 1.1.1 - Feature request to be looked at once 1.1.1 is released
> 
> 
> I think we should re-asses everything currently against the 1.1.1
> milestone so that anything which isn't specific to that release gets
> moved to the "Assessed" milestone.
> 
> At the moment its difficult to see the "wood for the trees" between
> issues which are newly introduced and those that are long standing. In
> terms of getting the 1.1.1 release out the door we should focus on the
> former.
> 
> Thoughts?
> 
> Matt
> 
> ___
> openssl-project mailing list
> openssl-project@openssl.org
> https://mta.openssl.org/mailman/listinfo/openssl-project
> 
> 
> ___
> openssl-project mailing list
> openssl-project@openssl.org
> https://mta.openssl.org/mailman/listinfo/openssl-project
> 
___
openssl-project mailing list
openssl-project@openssl.org
https://mta.openssl.org/mailman/listinfo/openssl-project

[openssl-project] FW: [openssl-commits] Build failed in Jenkins: master_noec #574

2018-06-26 Thread Salz, Rich
FYI

On 6/26/18, 2:45 PM, "Barry Fussell (bfussell)"  wrote:

The evp_test is failing intermittently because there is an attempt to
malloc zero bytes when running the new test that came in with
this commit.


https://bitbucket-eng-rtp1.cisco.com/bitbucket/projects/TS/repos/ciscossl/commits/7b3e775a6a78650bbd3e8e19a5aa12981880402b#test/evptests.txt


static int pderive_test_run(struct evp_test *t)
{
struct pkey_data *kdata = t->data;
unsigned char *out = NULL;
size_t out_len;
const char *err = "INTERNAL_ERROR";

out_len = kdata->output_len;
out = OPENSSL_malloc(out_len);  <- out is zero because there is no 
SharedSecret
if (!out) {

-Original Message-
From: osslsan...@gmail.com [mailto:osslsan...@gmail.com] 
Sent: Tuesday, June 19, 2018 11:28 AM
To: Barry Fussell (bfussell) ; 
openssl-comm...@openssl.org
Subject: Build failed in Jenkins: master_noec #574

See 


Changes:

[Matthias.St.Pierre] Fix & update documentation about RAND_priv_bytes()

[Matthias.St.Pierre] Improve the output of `make doc-nits`

--
[...truncated 505.72 KB...]
rm -f test/v3ext
${LDCMD:-gcc} -pthread -m64 -Wa,--noexecstack -Wall -O3 -L.   \
-o test/v3ext test/v3ext.o \
 test/libtestutil.a -lcrypto -ldl -pthread gcc  -Iinclude 
-pthread -m64 -Wa,--noexecstack -Wall -O3 -DNDEBUG  -MMD -MF 
test/v3nametest.d.tmp -MT test/v3nametest.o -c -o test/v3nametest.o 
test/v3nametest.c rm -f test/v3nametest
${LDCMD:-gcc} -pthread -m64 -Wa,--noexecstack -Wall -O3 -L.   \
-o test/v3nametest test/v3nametest.o \
 test/libtestutil.a -lcrypto -ldl -pthread gcc  -Iinclude 
-pthread -m64 -Wa,--noexecstack -Wall -O3 -DNDEBUG  -MMD -MF 
test/verify_extra_test.d.tmp -MT test/verify_extra_test.o -c -o 
test/verify_extra_test.o test/verify_extra_test.c rm -f test/verify_extra_test
${LDCMD:-gcc} -pthread -m64 -Wa,--noexecstack -Wall -O3 -L.   \
-o test/verify_extra_test test/verify_extra_test.o \
 test/libtestutil.a -lcrypto -ldl -pthread gcc  -Iinclude 
-pthread -m64 -Wa,--noexecstack -Wall -O3 -DNDEBUG  -MMD -MF 
test/versions.d.tmp -MT test/versions.o -c -o test/versions.o test/versions.c 
rm -f test/versions
${LDCMD:-gcc} -pthread -m64 -Wa,--noexecstack -Wall -O3 -L.   \
-o test/versions test/versions.o \
 -lcrypto -ldl -pthread
gcc  -Iinclude -pthread -m64 -Wa,--noexecstack -Wall -O3 -DNDEBUG  -MMD -MF 
test/wpackettest.d.tmp -MT test/wpackettest.o -c -o test/wpackettest.o 
test/wpackettest.c rm -f test/wpackettest
${LDCMD:-gcc} -pthread -m64 -Wa,--noexecstack -Wall -O3 -L.   \
-o test/wpackettest test/wpackettest.o \
 libssl.a test/libtestutil.a -lcrypto -ldl -pthread gcc  
-Iinclude -pthread -m64 -Wa,--noexecstack -Wall -O3 -DNDEBUG  -MMD -MF 
test/x509_check_cert_pkey_test.d.tmp -MT test/x509_check_cert_pkey_test.o -c -o 
test/x509_check_cert_pkey_test.o test/x509_check_cert_pkey_test.c rm -f 
test/x509_check_cert_pkey_test
${LDCMD:-gcc} -pthread -m64 -Wa,--noexecstack -Wall -O3 -L.   \
-o test/x509_check_cert_pkey_test 
test/x509_check_cert_pkey_test.o \
 test/libtestutil.a -lcrypto -ldl -pthread gcc  -Iinclude 
-pthread -m64 -Wa,--noexecstack -Wall -O3 -DNDEBUG  -MMD -MF 
test/x509_dup_cert_test.d.tmp -MT test/x509_dup_cert_test.o -c -o 
test/x509_dup_cert_test.o test/x509_dup_cert_test.c rm -f 
test/x509_dup_cert_test
${LDCMD:-gcc} -pthread -m64 -Wa,--noexecstack -Wall -O3 -L.   \
-o test/x509_dup_cert_test test/x509_dup_cert_test.o \
 test/libtestutil.a -lcrypto -ldl -pthread gcc  -I. -Iinclude 
-pthread -m64 -Wa,--noexecstack -Wall -O3 -DNDEBUG  -MMD -MF 
test/x509_internal_test.d.tmp -MT test/x509_internal_test.o -c -o 
test/x509_internal_test.o test/x509_internal_test.c rm -f 
test/x509_internal_test
${LDCMD:-gcc} -pthread -m64 -Wa,--noexecstack -Wall -O3 -L.   \
-o test/x509_internal_test test/x509_internal_test.o \
 test/libtestutil.a libcrypto.a -ldl -pthread gcc  -Iinclude 
-pthread -m64 -Wa,--noexecstack -Wall -O3 -DNDEBUG  -MMD -MF 
test/x509_time_test.d.tmp -MT test/x509_time_test.o -c -o test/x509_time_test.o 
test/x509_time_test.c rm -f test/x509_time_test
${LDCMD:-gcc} -pthread -m64 -Wa,--noexecstack -Wall -O3 -L.   \
-o test/x509_time_test test/x509_time_test.o \
 test/libtestutil.a -lcrypto -ldl -pthread gcc  -Iinclude 
-pthread -m64 -Wa,--noexecstack -Wall -O3 -DNDEBUG  -MMD -MF test/x509aux.d.tmp 
-MT test/x509aux.o -c -o test/x509aux.o test/x509aux.c rm -f test/x509aux

Re: [openssl-project] Milestones and the 1.1.1 release

2018-06-26 Thread Benjamin Kaduk
On Tue, Jun 26, 2018 at 04:56:26PM +0100, Matt Caswell wrote:
> I'm thinking that we should maybe re-asses the current milestones in github.
> 
> We currently use the following milestones:
> 
> Assessed - Anything against this milestone isn't relevant to the 1.1.1
> release (e.g. 1.0.2 specific issue)
> 
> 1.1.1 - This is relevant to the 1.1.1 release but may not be specific to
> it (e.g. an issue that affects both 1.1.1 and 1.1.0)
> 
> Post 1.1.1 - Feature request to be looked at once 1.1.1 is released
> 
> 
> I think we should re-asses everything currently against the 1.1.1
> milestone so that anything which isn't specific to that release gets
> moved to the "Assessed" milestone.
> 
> At the moment its difficult to see the "wood for the trees" between
> issues which are newly introduced and those that are long standing. In
> terms of getting the 1.1.1 release out the door we should focus on the
> former.
> 
> Thoughts?

If the choice is between your proposal and the current state, your proposal
seems better.  I don't want to start a bikeshed, so I'll try to leave the
discussion on the immediate topic which relates to getting 1.1.1 out the
door and not add in other things I'd like to see changed (but do not really
seem to be blocking 1.1.1).

-Ben
___
openssl-project mailing list
openssl-project@openssl.org
https://mta.openssl.org/mailman/listinfo/openssl-project


[openssl-project] Milestones and the 1.1.1 release

2018-06-26 Thread Matt Caswell
I'm thinking that we should maybe re-asses the current milestones in github.

We currently use the following milestones:

Assessed - Anything against this milestone isn't relevant to the 1.1.1
release (e.g. 1.0.2 specific issue)

1.1.1 - This is relevant to the 1.1.1 release but may not be specific to
it (e.g. an issue that affects both 1.1.1 and 1.1.0)

Post 1.1.1 - Feature request to be looked at once 1.1.1 is released


I think we should re-asses everything currently against the 1.1.1
milestone so that anything which isn't specific to that release gets
moved to the "Assessed" milestone.

At the moment its difficult to see the "wood for the trees" between
issues which are newly introduced and those that are long standing. In
terms of getting the 1.1.1 release out the door we should focus on the
former.

Thoughts?

Matt

___
openssl-project mailing list
openssl-project@openssl.org
https://mta.openssl.org/mailman/listinfo/openssl-project