You can view Azure Pipelines logs in the browser if you click the name
of the job (like here if you click "doctests"
https://dev.azure.com/SymPy/SymPy/_build/results?buildId=1396).

However, that reminds me about another annoyance of Azure, which is
that you can't link to a specific script or job log. You can only link
to the log for the whole build. Actually I discovered just now that if
you click a line number on a log it offers to create a link to that
line, but I tested it and it doesn't seem to actually work (it creates
a huge url like
https://dev.azure.com/SymPy/SymPy/_build/results?buildId=1396&view=logs&jobId=60d3e802-b22a-5dd9-4fed-c37ef5e6d2bc&taskId=f72b1324-e456-5562-0912-687e803699f2&lineStart=33&lineEnd=34&colStart=1&colEnd=1
that doesn't actually open the log at that point).

On Travis you just click a line number and it anchors to that line,
making a simple link like
https://travis-ci.org/sympy/sympy/jobs/485626173#L1696. And when you
download the raw log you get a plain text file
https://api.travis-ci.org/v3/job/485626173/log.txt.

Aaron Meurer

On Mon, Jan 28, 2019 at 1:27 PM Kalevi Suominen <jks...@gmail.com> wrote:
>
> I have found Travis logs quite useful as the browser can show them. Azure logs
> come in zip files which are rather inconvenient to deal with. I don't think 
> that
> dropping Azure would be a loss for me.
>
> Kalevi Suominen
>
> On Monday, January 28, 2019 at 10:01:51 PM UTC+2, Aaron Meurer wrote:
>>
>> So Azure Pipelines has been running now for about 3 1/2 months. I
>> would like to get feedback on how people like it.
>>
>> For myself, it seems that Travis has fixed most of the failing builds
>> issues that it was having. I've also noticed that Travis actually
>> finishes faster than Azure, possibly because of the higher concurrent
>> build limit.
>>
>> Frankly, I'm not too happy with Azure. The web interface is about as
>> complicated as they could possibly make it. It took me half an hour
>> just to find something as simple as the setting to disable email
>> notifications. The YAML spec is also much more complicated, and
>> despite being so, has some serious limitations, such as environment
>> variables not being shared across "scripts". The most annoying thing
>> by far is that people with push access don't have the ability to
>> restart failed builds. I think I as an admin can give people access,
>> but 1) this is annoying to do, and 2) I really can't even figure out
>> where I should do that (I can't stress just how complicated and
>> unintuitive their web interface is).
>>
>> Travis for its part is extremely simple, both in its web interface and
>> YAML spec. It's also very widely used, meaning most issues you would
>> encounter with it you can quickly find a workaround by Googling. For
>> instance, there is a workaround required to get Python 3.7 working,
>> but a Google search for "Python 3.7 Travis" turns up
>> https://github.com/travis-ci/travis-ci/issues/9815, which has the
>> workaround in the first comment. Contrast that with "Azure pipelines
>> pypy", which only turns up my issue requesting official support, with
>> no workarounds 
>> (https://github.com/Microsoft/azure-pipelines-tasks/issues/8514).
>>
>> So for my part, I would like to stop using it. I'm adding Python 3.7
>> support to Travis here, which was the only thing that was only on
>> Azure https://github.com/sympy/sympy/pull/15867 (yes, this will add
>> more Travis builds, but we should be able to drop 3.4 support after
>> this next release).
>>
>> But I would also like to hear others' opinions on it. Have you noticed
>> that Azure is any better than Travis in some way? I don't follow every
>> PR, so I don't notice every issue that comes up with Travis or Azure.
>>
>> Aaron Meurer
>>
>> On Fri, Nov 9, 2018 at 12:43 PM Aaron Meurer <asmeu...@gmail.com> wrote:
>> >
>> > Azure has fewer concurrent builds than Travis (10 instead of 15), but
>> > it also has longer builds 60 instead of 50, meaning we can split the
>> > tests 2 ways instead of 4. I think Azure also boots up faster and
>> > possibly has faster machines, though I haven't tested it.
>> >
>> > Aaron Meurer
>> > On Thu, Nov 8, 2018 at 9:06 PM Sidhant Nagpal <sidhantnagpa...@gmail.com> 
>> > wrote:
>> > >
>> > > Right, the test time is likely to be affected further on the account of 
>> > > lesser concurrent builds on Azure.
>> > >
>> > > --
>> > > You received this message because you are subscribed to the Google 
>> > > Groups "sympy" group.
>> > > To unsubscribe from this group and stop receiving emails from it, send 
>> > > an email to sympy+unsubscr...@googlegroups.com.
>> > > To post to this group, send email to sympy@googlegroups.com.
>> > > Visit this group at https://groups.google.com/group/sympy.
>> > > To view this discussion on the web visit 
>> > > https://groups.google.com/d/msgid/sympy/08646799-88cc-4375-af17-58fd6682fa9d%40googlegroups.com.
>> > > For more options, visit https://groups.google.com/d/optout.
>
> --
> You received this message because you are subscribed to the Google Groups 
> "sympy" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to sympy+unsubscr...@googlegroups.com.
> To post to this group, send email to sympy@googlegroups.com.
> Visit this group at https://groups.google.com/group/sympy.
> To view this discussion on the web visit 
> https://groups.google.com/d/msgid/sympy/bcf320e4-3dae-4165-83db-c61ee1f91a4b%40googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.

-- 
You received this message because you are subscribed to the Google Groups 
"sympy" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to sympy+unsubscr...@googlegroups.com.
To post to this group, send email to sympy@googlegroups.com.
Visit this group at https://groups.google.com/group/sympy.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sympy/CAKgW%3D6L8J3jtwuaCpw_M%2BDGU7j9QMGkM4NLc-qj%3DJu1nAjDkOA%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

Reply via email to