Fix for marker verts bug #7809

lkilcher · 2017-01-12T18:07:50Z

This is a proposed solution for #7807 . I added the checks for array ndim and shape, so that the ValueError('Unrecognized marker style...)' error is raised if an incorrectly-shaped np.ndarray is specified as input.

codecov-io · 2017-01-12T20:14:17Z

Current coverage is 62.14% (diff: 100%)

Merging #7809 into master will increase coverage by 0.04%

@@             master      #7809   diff @@
==========================================
  Files           174        175     +1   
  Lines         56050      56134    +84   
  Methods           0          0          
  Messages          0          0          
  Branches          0          0          
==========================================
+ Hits          34805      34883    +78   
- Misses        21245      21251     +6   
  Partials          0          0

Powered by Codecov. Last update d7879bc...91191ce

NelleV · 2017-01-13T00:33:01Z

What does BF stand for?

dopplershift · 2017-01-13T00:44:00Z

Bugfix?

lkilcher · 2017-01-13T01:16:22Z

Yes. Bug fix.

NelleV · 2017-01-13T17:51:31Z

Can you add a regression test?

The patch looks fine to me elsewise.

phobson

Could you write a test that demonstrates that this fix works properly?

NelleV · 2017-01-28T19:25:13Z

@lkilcher I've create a pull request onto your branch with some tests and an additional fix.

NelleV

It looks good to me. 👍

Thanks @lkilcher !

phobson

Just a minor correction to your comments

phobson · 2017-01-29T18:22:33Z

lib/matplotlib/tests/test_marker.py

+def test_markers_invalid():
+    marker_style = markers.MarkerStyle()
+    mrk_array = np.array([[-0.5, 0,  1, 2, 3]])
+    # Checking this doesn't fail.


checking that this does fail

Fixed. 23ce141

lkilcher · 2017-01-29T18:24:02Z

Thank you @NelleV , especially for the test!

Also, am I correct in my understanding of how Travis-CI (and other checks) work: they run the changes in the PR merged with HEAD of master (wherever that is at the time the PR is modified/updated)?

NelleV · 2017-01-29T18:39:53Z

No, travis-ci runs the test as they are on your branch. It doesn't merge into master, but once the pull request gets merged into master, it reruns the tests there.

Just a side note; HEAD is a tag that references the branch or the commit you are in. There's only one HEAD (so $HEAD of master" doesn't really mean anything in git jargon, but I think you are referring to matplotlib/matplotlib's master branch :) )

lkilcher · 2017-01-29T18:47:23Z

Thanks for the clarification. I suppose that's why the contributing doc suggests rebasing onto upstream master (is that the right jargon?) when finalizing a commit? Should I do that?

tacaswell · 2017-01-29T18:57:26Z

@NelleV I am pretty sure that travis runs against a commit which is the merge of the current branch into master (this is why fixing CI on master and re-starting PRs propagates the fix).

@lkilcher The rebase on master should only be if there are conflicts (in contrast to merging master into your branch). Can you point me at where in the docs suggest that you should rebase if there are no conflicts?

NelleV · 2017-01-29T19:07:34Z

@tacaswell really? I don't see the merge happening in the travis log

lkilcher · 2017-01-29T19:21:59Z

I think it was here, but now that I look at that it's fairly clear that this only needs to be done if there are conflicts. Perhaps it was an old version of that page, but more likely I'm just imagining things. Sorry!

tacaswell · 2017-01-29T19:23:18Z

The head it pulls from GH has the merge already done. The way the decide if the branch can be merged is by merging it and they expose that SHA out under as a specific ref. Copy the fetch logic from travis:

13:46 $ git fetch matplotlib +refs/pull/7976/merge
remote: Counting objects: 19, done.
remote: Compressing objects: 100% (19/19), done.
remote: Total 19 (delta 6), reused 0 (delta 0), pack-reused 0
Unpacking objects: 100% (19/19), done.
From git://github.com/matplotlib/matplotlib
 * branch                refs/pull/7976/merge -> FETCH_HEAD
(dd36) ✔ ~/source/matplotlib [fix_hatch_color L|✔] 
14:18 $ gitk
(dd36) ✔ ~/source/matplotlib [fix_hatch_color L|✔] 
14:18 $ gitk --all
(dd36) ✔ ~/source/matplotlib [fix_hatch_color L|✔] 
14:18 $ git checkout FETCH_HEAD
Note: checking out 'FETCH_HEAD'.

You are in 'detached HEAD' state. You can look around, make experimental
changes and commit them, and you can discard any commits you make in this
state without impacting any branches by performing another checkout.

If you want to create a new branch to retain commits you create, you may
do so (now or later) by using -b with the checkout command again. Example:

  git checkout -b <new-branch-name>

HEAD is now at 2d8f488a8... Merge f6b4a32e6fb4d4687b40da3a81b0eefaad57a995 into 9b6f7c1aa36460648d9485caa61a258ebe852b78
(dd36) ✔ ~/source/matplotlib [:2d8f488a8|✔] 
14:19 $ git log --graph
*   commit 2d8f488a8ebaf2c3a27eef8968ad016f743c8400
|\  Merge: 9b6f7c1aa f6b4a32e6
| | Author: Thomas A Caswell <[email protected]>
| | Date:   Sun Jan 29 18:50:01 2017 +0000
| | 
| |     Merge f6b4a32e6fb4d4687b40da3a81b0eefaad57a995 into 9b6f7c1aa36460648d9485caa61a258ebe852b78
| | 
| * commit f6b4a32e6fb4d4687b40da3a81b0eefaad57a995
| | Author: Thomas A Caswell <[email protected]>
| | Date:   Sat Jan 28 18:27:08 2017 -0500
| | 
| |     FIX: Restore hatch color tracking edge color
| | 
* | commit 9b6f7c1aa36460648d9485caa61a258ebe852b78
|/  Author: Nelle Varoquaux <[email protected]>
|   Date:   Sun Jan 29 10:40:50 2017 -0800
|   
|       Merge pull request #7961 from jkseppan/issue7937
|       
|       Compute glyph widths similarly in Type 42 as in Type 3
| 
* commit 5a895daebb47b960a04be2a4619aa421b1839782
| Author: Nelle Varoquaux <[email protected]>
| Date:   Mon Jan 23 23:40:48 2017 -0800
| 
|     Merge pull request #7752 from newville/master
|     
|     bugfix for wx backend: release mouse on loss of focus and before trying to recapture
|

This does leave some questions about exactly when that ref gets updated (and when the web-hooks get triggered).

NelleV · 2017-01-29T20:53:15Z

@lkilcher It seems that you'll have to rebase anyways… There is a conflict with master.

lkilcher · 2017-01-30T02:29:07Z

OK. I rebased and the tests passed, so I think this is good to go. :)

dstansby

Looks good, thanks @lkilcher

Fix for marker verts bug

dstansby · 2017-01-31T15:04:45Z

Backported to v2.0.x via. 98a9347

[TAC edited to fix sha link]

QuLogic · 2017-01-31T22:28:23Z

The test shouldn't have been backported as is; it requires pytest which is not one of the dependencies of the 2.0.x line.

tacaswell added this to the 2.0.1 (next bug fix release) milestone Jan 13, 2017

phobson added the status: needs revision label Jan 24, 2017

phobson requested changes Jan 24, 2017

View reviewed changes

NelleV self-assigned this Jan 28, 2017

NelleV changed the title ~~[BF] Fix for marker verts bug~~ [MRG+1] Fix for marker verts bug Jan 29, 2017

NelleV approved these changes Jan 29, 2017

View reviewed changes

phobson requested changes Jan 29, 2017

View reviewed changes

NelleV removed the status: needs revision label Jan 29, 2017

dstansby added the status: needs revision label Jan 29, 2017

NelleV removed the status: needs revision label Jan 29, 2017

lkilcher and others added 3 commits January 29, 2017 18:50

Fix for marker verts bug

4b7ad54

TST Added a test to check valid and invalid markers (#2)

41671e4

Correct test comment.

019d651

lkilcher force-pushed the marker_verts_bug branch from 23ce141 to 019d651 Compare January 30, 2017 01:50

phobson approved these changes Jan 30, 2017

View reviewed changes

NelleV changed the title ~~[MRG+1] Fix for marker verts bug~~ [MRG+2] Fix for marker verts bug Jan 30, 2017

dstansby changed the title ~~[MRG+2] Fix for marker verts bug~~ Fix for marker verts bug Jan 30, 2017

dstansby approved these changes Jan 30, 2017

View reviewed changes

dstansby merged commit 70831bd into matplotlib:master Jan 30, 2017

dstansby mentioned this pull request Jan 30, 2017

ValueError when specifying marker vertices #7807

Closed

dstansby self-assigned this Jan 30, 2017

dstansby added a commit that referenced this pull request Jan 31, 2017

Merge pull request #7809 from lkilcher/marker_verts_bug

98a9347

Fix for marker verts bug

dstansby removed their assignment Jan 31, 2017

dstansby mentioned this pull request Jan 31, 2017

Fix broken backported test #8004

Merged

Uh oh!

Fix for marker verts bug #7809

Fix for marker verts bug #7809

Uh oh!

Conversation

lkilcher commented Jan 12, 2017

Uh oh!

codecov-io commented Jan 12, 2017 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Current coverage is 62.14% (diff: 100%)

Uh oh!

NelleV commented Jan 13, 2017

Uh oh!

dopplershift commented Jan 13, 2017

Uh oh!

lkilcher commented Jan 13, 2017

Uh oh!

NelleV commented Jan 13, 2017

Uh oh!

phobson left a comment

Choose a reason for hiding this comment

Uh oh!

NelleV commented Jan 28, 2017

Uh oh!

NelleV left a comment

Choose a reason for hiding this comment

Uh oh!

phobson left a comment

Choose a reason for hiding this comment

Uh oh!

phobson Jan 29, 2017

Choose a reason for hiding this comment

Uh oh!

lkilcher Jan 29, 2017

Choose a reason for hiding this comment

Uh oh!

lkilcher commented Jan 29, 2017

Uh oh!

NelleV commented Jan 29, 2017

Uh oh!

lkilcher commented Jan 29, 2017

Uh oh!

tacaswell commented Jan 29, 2017

Uh oh!

NelleV commented Jan 29, 2017

Uh oh!

lkilcher commented Jan 29, 2017

Uh oh!

tacaswell commented Jan 29, 2017

Uh oh!

NelleV commented Jan 29, 2017

Uh oh!

lkilcher commented Jan 30, 2017

Uh oh!

dstansby left a comment

Choose a reason for hiding this comment

Uh oh!

dstansby commented Jan 31, 2017 • edited by tacaswell Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

QuLogic commented Jan 31, 2017

Uh oh!

Uh oh!

codecov-io commented Jan 12, 2017 •

edited

Loading

dstansby commented Jan 31, 2017 •

edited by tacaswell

Loading