[GNC-dev] Test failure

Sun Jan 27 10:17:06 EST 2019

You can do it that way or you can do

git clone https://github.com/christopherlam/gnucash.git -b maint-chartjs-budget-barchart/

or 

git clone https://github.com/christopherlam/gnucash.git
git checkout  -b maint-chartjs-budget-barchart/

or the way we do for testing pull requests; this is a bit fiddly because of the possible merge commit and conflicts, but it doesn't permanently add a remote:

git checkout -b maint-chartjs-budget-barchart maint
git pull https://github.com/christopherlam/gnucash.git maint-chartjs-budget-barchart/

You can get a fast-forward if you go to https://github.com/christopherlam/gnucash/tree/maint-chartjs-budget-barchart, click the "20,557 Commits", and scroll down to find the first commit by someone else, in this case it's f9f714c. Use that instead of maint as the last argument to git checkout.

You can't clone a GitHub tree URL, that's not a repository. 

Regards,
John Ralls

> On Jan 26, 2019, at 11:20 PM, Christopher Lam <christopher.lck at gmail.com> wrote:
> 
> Ok my bad
> 
> May be safer to reuse your own gnucash repository
> 
> * cd gnucash
> * git remote add chris https://github.com/christopherlam/gnucash.git
> * git fetch --all
> * git checkout chris/maint-chartjs-budget-barchart
> * dpkg etc
> 
> and to revert to main branch:
> 
> * git checkout maint
> 
> 
> 
> On 27/1/19 3:04 pm, Stephen M. Butler wrote:
>> On 1/26/19 10:54 PM, Christopher Lam wrote:
>>> Stephen please spend some CPU burning time on my chartJS branch which
>>> needs beta testing?
>>> 
>>> $git clone
>>> https://github.com/christopherlam/gnucash/tree/maint-chartjs-budget-barchart
>> 
>> Thought I'd start this while headed toward some z's.  Too tired to
>> figure out what happened.
>> 
>>  git clone
>> https://github.com/christopherlam/gnucash/tree/maint-chartjs-budget-barchart
>> Cloning into 'maint-chartjs-budget-barchart'...
>> fatal: repository
>> 'https://github.com/christopherlam/gnucash/tree/maint-chartjs-budget-barchart/'
>> not found
>> 
>> 
>> --Steve
>> 
>>> 
>>> On 27/1/19 2:52 pm, Stephen M. Butler wrote:
>>>> On 1/26/19 10:29 PM, John Ralls wrote:
>>>>> Stephen,
>>>>> 
>>>>> By all means learn about git bisect, it's an extremely useful
>>>>> troubleshooting tool. But don't exercise yourself too much about
>>>>> this particular failure, Chris and I know exactly what commit caused
>>>>> it and that the underlying cause is a bug in ICU (International
>>>>> Components for Unicode, http://site.icu-project.org/) versions 57-62.
>>>> I figured the devs were way ahead of me.  And I'm burning the CPU this
>>>> evening finding that I can't compile 3.4-21 thru (I'm guessing) 3.4-38.
>>>> I'm up to 3.4-29 at this point and already know that -37 and -38 don't
>>>> compile on my box.  But -39, -40, and -43 do.  Don't know about -41 and
>>>> -42 (and probably won't check them).
>>>> 
>>>> Busy day tomorrow so I doubt the cpu will even get warmed up!  It's
>>>> pushing 11 pm so I'll call it quits on this exercise.
>>>> 
>>>> Thanks for the details.
>>>> 
>>>> --Steve
>>>> 
>>>>> I had thought that the problem was confined to 1969, but tonight got
>>>>> https://bugs.gnucash.org/show_bug.cgi?id=797067 which makes me think
>>>>> that it's every year, so more testing needed.
>>>>> 
>>>>> Regards,
>>>>> John Ralls
>>>>> 
>>>>> 
>>>>>> On Jan 26, 2019, at 5:55 PM, Stephen M. Butler <kg7je at arrl.net> wrote:
>>>>>> 
>>>>>> On 1/26/19 3:13 AM, Christopher Lam wrote:
>>>>>>> Another experiment. Try prefix your build with "TZ=UTC " on current
>>>>>>> maint.
>>>>>>> 
>>>>>>> Eg if using ninja, it'll be "TZ=UTC ninja check"
>>>>>> TZ=UTC dpkg- -rfakeroot -b -uc
>>>>>> 
>>>>>> ....
>>>>>> 
>>>>>> End testing: Jan 27 01:47 UTC
>>>>>> + exit 2
>>>>>> debian/rules:65: recipe for target 'override_dh_auto_test' failed
>>>>>> make[1]: *** [override_dh_auto_test] Error 2
>>>>>> make[1]: Leaving directory '/home/steve/Projects/GnuCash/gnucash'
>>>>>> debian/rules:24: recipe for target 'build' failed
>>>>>> make: *** [build] Error 2
>>>>>> dpkg-buildpackage: error: debian/rules build subprocess returned exit
>>>>>> status 2
>>>>>> 
>>>>>> Same error -- Test 109 that fails.
>>>>>> 
>>>>>> 
>>>>>> This is with the patches for 3.4.  I verified before Sabbath that 3.4
>>>>>> builds and passes the tests.  Tried your test first thing this evening
>>>>>> after git checkout maint and git pull.  I am at 3.4-43-g84d1c3645
>>>>>> which
>>>>>> builds but doesn't pass the tests.  So, I could build turning off the
>>>>>> test step.
>>>>>> 
>>>>>> However, Friday I was learning more about git bisect to see if I could
>>>>>> figure out which patch caused the tests to start failing.  However, a
>>>>>> couple of items got in the way:  1--time;  2--some commits don't
>>>>>> build.
>>>>>> That messed me up a bit.  So, I may start over again with the bisect
>>>>>> (just for my own edification) and see what happens.  Maybe this time
>>>>>> I'll do a git describe each time and take notes so I'll be a little
>>>>>> more
>>>>>> cognizant of what I'm learning.  <<did that make sense>>
>>>>>> 
>>>>>> 
>>>>>> --Steve (in the dark)
>>>>>> 
>>>>>>> On Fri., 25 Jan. 2019, 11:52 Stephen M. Butler <kg7je at arrl.net
>>>>>>> <mailto:kg7je at arrl.net> wrote:
>>>>>>> 
>>>>>>>     On 1/24/19 7:36 PM, Christopher Lam wrote:
>>>>>>>> Revert is a git terminology, not a gnucash one. Welcome to version
>>>>>>>> control.
>>>>>>>     That's the manual to which I referred.  It meant something
>>>>>>> else to my
>>>>>>>     mind when I first saw your post.  But, after further thought, my
>>>>>>>     previous thought was faulty.
>>>>>>> 
>>>>>>> 
>>>>>>>     My first boss after college would draw a picture to describe the
>>>>>>>     difference between a bug and a feature.
>>>>>>> 
>>>>>>>     I can't draw so a word will have to do (at less than a thousand
>>>>>>>     words)!
>>>>>>> 
>>>>>>>     BUG:  draw stick figure.
>>>>>>> 
>>>>>>> 
>>>>>>>     FEATURE:  Add pants, tie, gloves, shoes and a top hat.
>>>>>>> 
>>>>>>> 
>>>>>>>     MORAL:  A feature is a bug all dressed up and ready to go out.
>>>>>>> 
>>>>>>>> Please remember this revert is reverting a buggy code with a
>>>>>>>     previous
>>>>>>>> buggy code, so, cannot be consisted safe to package; best wait until
>>>>>>>> the clever devs can find a proper fix for gnc-date.
>>>>>>>> 
>>>>>>>> On Fri., 25 Jan. 2019, 05:46 Stephen M. Butler <kg7je at arrl.net
>>>>>>>     <mailto:kg7je at arrl.net>
>>>>>>>> <mailto:kg7je at arrl.net <mailto:kg7je at arrl.net>> wrote:
>>>>>>>> 
>>>>>>>>       OH!  Light bulb on!
>>>>>>>>       OK.  I understand what you said earlier.  I'll put the patches
>>>>>>>>       back into
>>>>>>>>       the packaging.
>>>>>>>> 
>>>>>>>>       Thanks for clearing that up.
>>>>>>>> 
>>>>>>>>       --Steve
>>>>>>>> 
>>>>>>>>       PS Looking in manual for 'revert' command.
> _______________________________________________
> gnucash-devel mailing list
> gnucash-devel at gnucash.org
> https://lists.gnucash.org/mailman/listinfo/gnucash-devel