Travis failures on NMatrix
10 views
Skip to first unread message
John Woods
Oct 12, 2016, 6:01:09 PM10/12/16
to sciru...@googlegroups.com
Anyone able to provide any insight on this issue?
It's currently blocking at least two pull requests and possibly a third. Both the errors seem to relate only to Ruby 2.0.
John
Isuru Fernando
Oct 13, 2016, 12:21:45 AM10/13/16
to sciru...@googlegroups.com
This seems to be because Travis is using 2.0.0-p648-clang instead of 2.0.0-p648. You'll have to talk with travis-ci to resolve this.
Failure you are getting with clang is legit. You just need to fix it. This seems to be because of an old commit, 10a87624e7d849a1ab13cf4ed00ecef1dcf95605 which assumed that `VALUE` is not equal to `uint64_t`, but it is not for 64-bit systems.
g++ ignores this redefinition and treats it as an error only if the method is called. clang++ on the other hand treats it as an error even if the method is not called. Therefore NMatrix is not compilable via clang on Linux. This problem is easy to fix though, see commit and travis-build.
Thanks,
Isuru Fernando
Failure you are getting with clang is legit. You just need to fix it. This seems to be because of an old commit, 10a87624e7d849a1ab13cf4ed00ecef1dcf95605 which assumed that `VALUE` is not equal to `uint64_t`, but it is not for 64-bit systems.
g++ ignores this redefinition and treats it as an error only if the method is called. clang++ on the other hand treats it as an error even if the method is not called. Therefore NMatrix is not compilable via clang on Linux. This problem is easy to fix though, see commit and travis-build.
Thanks,
Isuru Fernando
--
You received this message because you are subscribed to the Google Groups "SciRuby Development" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sciruby-dev+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
Kenta Murata
Oct 18, 2016, 9:43:14 PM10/18/16
to sciru...@googlegroups.com
All supports for Ruby 2.0.0 had ended on 24 Feb 2016.
We can drop 2.0.0 from supported Ruby versions, and move 2.0.0 to allow_failure section.
How do you think about dropping 2.0.0 support, John?
2016年10月13日(木) 13:21 Isuru Fernando <isu...@gmail.com>:
To unsubscribe from this group and stop receiving emails from it, send an email to sciruby-dev...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
--
You received this message because you are subscribed to the Google Groups "SciRuby Development" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sciruby-dev...@googlegroups.com.
John Woods
Oct 19, 2016, 12:40:57 PM10/19/16
to sciru...@googlegroups.com
That's a good point. I think I'll probably merge Isuru's change and then we can deprecate Ruby 2.0.0 support as well.
Isuru Fernando
Oct 20, 2016, 9:41:57 AM10/20/16
to sciru...@googlegroups.com
Why is it that when dropping support for a ruby version, it is moved to allow_failure section? This seems like a waste of travis resources. You can just drop the builds and don't test at all.
Isuru Fernando
Isuru Fernando
On Wed, Oct 19, 2016 at 10:10 PM, John Woods <john.o...@gmail.com> wrote:
That's a good point. I think I'll probably merge Isuru's change and then we can deprecate Ruby 2.0.0 support as well.
On Tue, Oct 18, 2016 at 8:43 PM Kenta Murata <mr...@mrkn.jp> wrote:
All supports for Ruby 2.0.0 had ended on 24 Feb 2016.We can drop 2.0.0 from supported Ruby versions, and move 2.0.0 to allow_failure section.How do you think about dropping 2.0.0 support, John?2016年10月13日(木) 13:21 Isuru Fernando <isu...@gmail.com>:
This seems to be because Travis is using 2.0.0-p648-clang instead of 2.0.0-p648. You'll have to talk with travis-ci to resolve this.
Failure you are getting with clang is legit. You just need to fix it. This seems to be because of an old commit, 10a87624e7d849a1ab13cf4ed00ecef1dcf95605 which assumed that `VALUE` is not equal to `uint64_t`, but it is not for 64-bit systems.
g++ ignores this redefinition and treats it as an error only if the method is called. clang++ on the other hand treats it as an error even if the method is not called. Therefore NMatrix is not compilable via clang on Linux. This problem is easy to fix though, see commit and travis-build.
Thanks,
Isuru Fernando
On Thu, Oct 13, 2016 at 3:30 AM, John Woods <john.o...@gmail.com> wrote:
Anyone able to provide any insight on this issue?It's currently blocking at least two pull requests and possibly a third. Both the errors seem to relate only to Ruby 2.0.John
--
You received this message because you are subscribed to the Google Groups "SciRuby Development" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sciruby-dev+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
--
You received this message because you are subscribed to the Google Groups "SciRuby Development" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sciruby-dev+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
--
You received this message because you are subscribed to the Google Groups "SciRuby Development" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sciruby-dev+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
--
You received this message because you are subscribed to the Google Groups "SciRuby Development" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sciruby-dev+unsubscribe@googlegroups.com.
John Woods
Oct 20, 2016, 10:47:14 AM10/20/16
to sciru...@googlegroups.com
If Travis were working more consistently, it would help us to know when we actually introduce a commit that breaks compatibility. We then would make an active choice to drop that Ruby version from allowable failures.
So, as an example, I think only Ruby 2.0.0 should be among the allowable failures at this point.
But it probably doesn't matter much.
So, as an example, I think only Ruby 2.0.0 should be among the allowable failures at this point.
But it probably doesn't matter much.
On Thu, Oct 20, 2016 at 8:41 AM Isuru Fernando <isu...@gmail.com> wrote:
Why is it that when dropping support for a ruby version, it is moved to allow_failure section? This seems like a waste of travis resources. You can just drop the builds and don't test at all.
Isuru Fernando
On Wed, Oct 19, 2016 at 10:10 PM, John Woods <john.o...@gmail.com> wrote:
That's a good point. I think I'll probably merge Isuru's change and then we can deprecate Ruby 2.0.0 support as well.
On Tue, Oct 18, 2016 at 8:43 PM Kenta Murata <mr...@mrkn.jp> wrote:
All supports for Ruby 2.0.0 had ended on 24 Feb 2016.We can drop 2.0.0 from supported Ruby versions, and move 2.0.0 to allow_failure section.How do you think about dropping 2.0.0 support, John?2016年10月13日(木) 13:21 Isuru Fernando <isu...@gmail.com>:
This seems to be because Travis is using 2.0.0-p648-clang instead of 2.0.0-p648. You'll have to talk with travis-ci to resolve this.
Failure you are getting with clang is legit. You just need to fix it. This seems to be because of an old commit, 10a87624e7d849a1ab13cf4ed00ecef1dcf95605 which assumed that `VALUE` is not equal to `uint64_t`, but it is not for 64-bit systems.
g++ ignores this redefinition and treats it as an error only if the method is called. clang++ on the other hand treats it as an error even if the method is not called. Therefore NMatrix is not compilable via clang on Linux. This problem is easy to fix though, see commit and travis-build.
Thanks,
Isuru Fernando
On Thu, Oct 13, 2016 at 3:30 AM, John Woods <john.o...@gmail.com> wrote:
Anyone able to provide any insight on this issue?It's currently blocking at least two pull requests and possibly a third. Both the errors seem to relate only to Ruby 2.0.John
--
You received this message because you are subscribed to the Google Groups "SciRuby Development" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sciruby-dev...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
--
You received this message because you are subscribed to the Google Groups "SciRuby Development" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sciruby-dev...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
--
You received this message because you are subscribed to the Google Groups "SciRuby Development" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sciruby-dev...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
--
You received this message because you are subscribed to the Google Groups "SciRuby Development" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sciruby-dev...@googlegroups.com.
--
You received this message because you are subscribed to the Google Groups "SciRuby Development" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sciruby-dev...@googlegroups.com.
Reply all
Reply to author
Forward
0 new messages