[1.34.0] [thread] [graph] finish markup sync
I've checked in the attached patches to: 1. Merge changes to required toolset list from RC to HEAD 2. Merge a couple of new test markups fron HEAD to RC. (2) applies to graph and thread libraries. I've verified that the mentioned tests exist on RC; if the library authors believe those changes should not be on RC, just say. I've also moved merged_to_RC_1_34_0 tag on .xml and .xsd files. - Volodya
Vladimir Prus wrote:
(2) applies to graph and thread libraries. I've verified that the mentioned tests exist on RC; if the library authors believe those changes should not be on RC, just say. + <mark-expected-failures> + <test name="test_barrier_lib"/> + <toolset name="msvc-8.0"/> + <note author="Aleksey Gurtovoy" refid="6"/> + </mark-expected-failures> + + <mark-expected-failures> + <test name="test_thread"/> + <toolset name="msvc-7.1"/> + <note author="Aleksey Gurtovoy" refid="6"/> + </mark-expected-failures>
<mark-expected-failures reason="?"> <test name="*_lib"/>
The markup for thread were correct on RC_. Don't know what these on HEAD are. For Boost.Thread we currently do have the unfortunate situation that the "real" development is not on HEAD but on thread_rewrite branch. (I know this is odd, but this is how things are at the moment.) This branch also does not cover the entire tree, but only the thread subdirectories. We (Anthony and me) are planning to move the thread_rewrite to SVN trunk after the release to not introduce even more ambiguities before the 1.34 release. David told me that the plan after 1.34 release is to make the current RC_1_34_0 branch the HEAD branch on upcoming SVN. The HEAD branch for Boost.Thread currently is kind of "orphaned" and given the above I cannot see much sense maintaining it. In case of differences between HEAD and RC_1_34_0 currently it is safer to leave RC_1_34_0 as is and forward differences to HEAD. Thank you, Roland
Roland Schwarz wrote:
Vladimir Prus wrote:
(2) applies to graph and thread libraries. I've verified that the mentioned tests exist on RC; if the library authors believe those changes should not be on RC, just say. + <mark-expected-failures> + <test name="test_barrier_lib"/> + <toolset name="msvc-8.0"/> + <note author="Aleksey Gurtovoy" refid="6"/> + </mark-expected-failures> + + <mark-expected-failures> + <test name="test_thread"/> + <toolset name="msvc-7.1"/> + <note author="Aleksey Gurtovoy" refid="6"/> + </mark-expected-failures>
<mark-expected-failures reason="?"> <test name="*_lib"/>
The markup for thread were correct on RC_. Don't know what these on HEAD are.
For Boost.Thread we currently do have the unfortunate situation that the "real" development is not on HEAD but on thread_rewrite branch. (I know this is odd, but this is how things are at the moment.) This branch also does not cover the entire tree, but only the thread subdirectories. We (Anthony and me) are planning to move the thread_rewrite to SVN trunk after the release to not introduce even more ambiguities before the 1.34 release.
David told me that the plan after 1.34 release is to make the current RC_1_34_0 branch the HEAD branch on upcoming SVN.
It's the first time I hear this; was this ever discussed?
The HEAD branch for Boost.Thread currently is kind of "orphaned" and given the above I cannot see much sense maintaining it.
In case of differences between HEAD and RC_1_34_0 currently it is safer to leave RC_1_34_0 as is and forward differences to HEAD.
Hmm, at this moment, failures markup on HEAD and RC_1_34_0 are exactly the same. Do you suggest the above markup should be removed from RC_1_34_0, or what? - Volodya
Vladimir Prus <ghost@cs.msu.su> writes:
I've checked in the attached patches to:
1. Merge changes to required toolset list from RC to HEAD 2. Merge a couple of new test markups fron HEAD to RC.
(2) applies to graph and thread libraries. I've verified that the mentioned tests exist on RC; if the library authors believe those changes should not be on RC, just say.
+ <mark-expected-failures> + <test name="test_barrier_lib"/> + <toolset name="msvc-8.0"/> + <note author="Aleksey Gurtovoy" refid="6"/> + </mark-expected-failures> + + <mark-expected-failures> + <test name="test_thread"/> + <toolset name="msvc-7.1"/> + <note author="Aleksey Gurtovoy" refid="6"/> + </mark-expected-failures>
I don't think there should be any expected failures for the thread library on msvc. Anthony -- Anthony Williams Just Software Solutions Ltd - http://www.justsoftwaresolutions.co.uk Registered in England, Company Number 5478976. Registered Office: 15 Carrallack Mews, St Just, Cornwall, TR19 7UL
Anthony Williams wrote:
Vladimir Prus <ghost@cs.msu.su> writes:
I've checked in the attached patches to:
1. Merge changes to required toolset list from RC to HEAD 2. Merge a couple of new test markups fron HEAD to RC.
(2) applies to graph and thread libraries. I've verified that the mentioned tests exist on RC; if the library authors believe those changes should not be on RC, just say.
+ <mark-expected-failures> + <test name="test_barrier_lib"/> + <toolset name="msvc-8.0"/> + <note author="Aleksey Gurtovoy" refid="6"/> + </mark-expected-failures> + + <mark-expected-failures> + <test name="test_thread"/> + <toolset name="msvc-7.1"/> + <note author="Aleksey Gurtovoy" refid="6"/> + </mark-expected-failures>
I don't think there should be any expected failures for the thread library on msvc.
Why there's one on HEAD? Should the markup be removed on HEAD and RC, just HEAD, just RC, or something else? Thanks, Volodya
Vladimir Prus <ghost@cs.msu.su> writes:
Anthony Williams wrote:
Vladimir Prus <ghost@cs.msu.su> writes:
I've checked in the attached patches to:
1. Merge changes to required toolset list from RC to HEAD 2. Merge a couple of new test markups fron HEAD to RC.
(2) applies to graph and thread libraries. I've verified that the mentioned tests exist on RC; if the library authors believe those changes should not be on RC, just say.
+ <mark-expected-failures> + <test name="test_barrier_lib"/> + <toolset name="msvc-8.0"/> + <note author="Aleksey Gurtovoy" refid="6"/> + </mark-expected-failures> + + <mark-expected-failures> + <test name="test_thread"/> + <toolset name="msvc-7.1"/> + <note author="Aleksey Gurtovoy" refid="6"/> + </mark-expected-failures>
I don't think there should be any expected failures for the thread library on msvc.
Why there's one on HEAD? Should the markup be removed on HEAD and RC, just HEAD, just RC, or something else?
The markup should be removed on RC. As Roland explained, the new work on the thread lib is happening on a branch (it started as a major rewrite, and includes some breaking changes), so HEAD is pretty much orphaned. It's probably best to keep HEAD in sync with the RC branch for now. Anthony -- Anthony Williams Just Software Solutions Ltd - http://www.justsoftwaresolutions.co.uk Registered in England, Company Number 5478976. Registered Office: 15 Carrallack Mews, St Just, Cornwall, TR19 7UL
Anthony Williams wrote:
Vladimir Prus <ghost@cs.msu.su> writes:
Anthony Williams wrote:
Vladimir Prus <ghost@cs.msu.su> writes:
I've checked in the attached patches to:
1. Merge changes to required toolset list from RC to HEAD 2. Merge a couple of new test markups fron HEAD to RC.
(2) applies to graph and thread libraries. I've verified that the mentioned tests exist on RC; if the library authors believe those changes should not be on RC, just say.
+ <mark-expected-failures> + <test name="test_barrier_lib"/> + <toolset name="msvc-8.0"/> + <note author="Aleksey Gurtovoy" refid="6"/> + </mark-expected-failures> + + <mark-expected-failures> + <test name="test_thread"/> + <toolset name="msvc-7.1"/> + <note author="Aleksey Gurtovoy" refid="6"/> + </mark-expected-failures>
I don't think there should be any expected failures for the thread library on msvc.
Why there's one on HEAD? Should the markup be removed on HEAD and RC, just HEAD, just RC, or something else?
The markup should be removed on RC. As Roland explained, the new work on the thread lib is happening on a branch (it started as a major rewrite, and includes some breaking changes), so HEAD is pretty much orphaned. It's probably best to keep HEAD in sync with the RC branch for now.
I'm sorry, I'm still lost. Above you say: It's probably best to keep HEAD in sync with the RC branch for now. If I remove markup from RC branch, the HEAD will no longer be in sync with RC. Is that what you want? - Volodya
Vladimir Prus <ghost@cs.msu.su> writes:
Anthony Williams wrote:
Vladimir Prus <ghost@cs.msu.su> writes:
Anthony Williams wrote:
Vladimir Prus <ghost@cs.msu.su> writes:
I've checked in the attached patches to:
1. Merge changes to required toolset list from RC to HEAD 2. Merge a couple of new test markups fron HEAD to RC.
(2) applies to graph and thread libraries. I've verified that the mentioned tests exist on RC; if the library authors believe those changes should not be on RC, just say.
+ <mark-expected-failures> + <test name="test_barrier_lib"/> + <toolset name="msvc-8.0"/> + <note author="Aleksey Gurtovoy" refid="6"/> + </mark-expected-failures> + + <mark-expected-failures> + <test name="test_thread"/> + <toolset name="msvc-7.1"/> + <note author="Aleksey Gurtovoy" refid="6"/> + </mark-expected-failures>
I don't think there should be any expected failures for the thread library on msvc.
Why there's one on HEAD? Should the markup be removed on HEAD and RC, just HEAD, just RC, or something else?
The markup should be removed on RC. As Roland explained, the new work on the thread lib is happening on a branch (it started as a major rewrite, and includes some breaking changes), so HEAD is pretty much orphaned. It's probably best to keep HEAD in sync with the RC branch for now.
I'm sorry, I'm still lost. Above you say:
It's probably best to keep HEAD in sync with the RC branch for now.
If I remove markup from RC branch, the HEAD will no longer be in sync with RC. Is that what you want?
Sorry for not being clear. The above was a round-about way of saying "remove the markup from RC and HEAD". Anthony -- Anthony Williams Just Software Solutions Ltd - http://www.justsoftwaresolutions.co.uk Registered in England, Company Number 5478976. Registered Office: 15 Carrallack Mews, St Just, Cornwall, TR19 7UL
Anthony Williams wrote:
The markup should be removed on RC. As Roland explained, the new work on the thread lib is happening on a branch (it started as a major rewrite, and includes some breaking changes), so HEAD is pretty much orphaned. It's probably best to keep HEAD in sync with the RC branch for now.
I'm sorry, I'm still lost. Above you say:
It's probably best to keep HEAD in sync with the RC branch for now.
If I remove markup from RC branch, the HEAD will no longer be in sync with RC. Is that what you want?
Sorry for not being clear. The above was a round-about way of saying "remove the markup from RC and HEAD".
Ok, now that we're on the same page, I've removed those two markup elements everywhere. Thanks, Volodya
Vladimir Prus wrote:
If I remove markup from RC branch, the HEAD will no longer be in sync with RC. Is that what you want?
Basically yes. But please don't sync both of them now. There is still some documentation in HEAD which I will need later on. Sorry for this mess, but I think there is not much to do for Boost.Thread as we do not have errors on RC_1_34_0. So I think the best is not to touch RC_1_34_0 thread related things for the momenet. Ok? Just revert the last change you have posted. Roland
Vladimir Prus wrote:
Why there's one on HEAD? Should the markup be removed on HEAD and RC, just HEAD, just RC, or something else?
Didn't you just merge the markups from HEAD to RC_1_34_0? I was refering to those. As Anthony said: we are currently not expecting anything to fail for thread on msvc. Roland
participants (3)
-
Anthony Williams -
Roland Schwarz -
Vladimir Prus