Thread: [matplotlib-devel] github pull requests: odd listings of hundreds of commits

I suspect that the series of May 6 commits ending with 
a50874b711983cba505ecdb2801c4996eccf3812 has tangled the history in such 
a way that some (but not all) older pull requests on github like this one

https://github.com/matplotlib/matplotlib/pull/1

are showing hundreds of commits and diffs.

Eric

Darned if I know what I did differently that time.  (I'm sure I hit 
git's "misunderstanding" feature again).

It only seems to have affected 2 or 3 of the pull requests -- I suspect 
rebasing those branches off of the current master would fix the problem, 
but maybe there's an easier way.

Mike

On 05/17/2011 03:16 PM, Eric Firing wrote:
> I suspect that the series of May 6 commits ending with
> a50874b711983cba505ecdb2801c4996eccf3812 has tangled the history in such
> a way that some (but not all) older pull requests on github like this one
>
> https://github.com/matplotlib/matplotlib/pull/1
>
> are showing hundreds of commits and diffs.
>
> Eric
>
> ------------------------------------------------------------------------------
> Achieve unprecedented app performance and reliability
> What every C/C++ and Fortran developer should know.
> Learn how Intel has extended the reach of its next-generation tools
> to help boost performance applications - inlcuding clusters.
> http://p.sf.net/sfu/intel-dev2devmay
> _______________________________________________
> Matplotlib-devel mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel
>    


-- 
Michael Droettboom
Science Software Branch
Space Telescope Science Institute
Baltimore, Maryland, USA

On 05/17/2011 10:14 AM, Michael Droettboom wrote:
> Darned if I know what I did differently that time.  (I'm sure I hit
> git's "misunderstanding" feature again).
>
> It only seems to have affected 2 or 3 of the pull requests -- I suspect
> rebasing those branches off of the current master would fix the problem,
> but maybe there's an easier way.

Mike,

First, regarding the pull requests, I think that in some cases it is not 
a problem because they are being closed (rejected or superseded), and in 
any case, it is easy to handle them locally and then ensure a clean 
changeset to be pushed up to the main repo.  So they look terrible in 
the github web view, but no real damage is done to the information in 
the branches they represent when one pulls those branches and works with 
them locally.

Second, I think the problems came from one or two exceptions to 
recommended git practice (this is coming from a git non-expert, so take 
with a grain of salt; git experts, please chime in and correct me as 
needed), which appear to be of two types: one is pulling from "origin" 
when you intended to pull from "upstream":

* | |   commit 14406a68c039dc6578730f23955bf34d34008a08
|\ \ \  Merge: fdf5fae 132f967
| | | | Author: Michael Droettboom <md...@gm...>
| | | | Date:   Fri May 6 15:25:01 2011 -0400
| | | |
| | | |     Merge remote branch 'origin/v1.0.x'

(this is the one that deleted scads of files, which you replaced in the 
subsequent commit)

and the other is merging from origin or upstream into a feature branch:

| | * |   commit 668ef6d826eff5eb4aac6251e1ddd9ac713616e5
| | |\ \  Merge: de18d9a 53f8139
| | |/ /  Author: Michael Droettboom <md...@gm...>
| |/| |   Date:   Thu May 12 11:43:48 2011 -0400
| | | |
| | | |       Merge remote branch 'upstream/v1.0.x' into gtk_crash

My understanding is that:

1) merges should always be from a feature branch into v1.0.x or master 
(and from v1.0.x into master, not the reverse); this can be done by 
updating local copies of v1.0.x etc from upstream, merging from the 
feature branch, and then pushing to upstream.

2) If a feature branch is local only (no one else has it) then it can be 
rebased so that merging it into v1.0.x etc. does not involve a merge 
commit; but if the feature branch has been published on github, then it 
should not be rebased.

3) We don't have to always push sets of changes from an original pull 
request to upstream; they can be consolidated using any of a variety of 
methods to form a new local feature branch with the same net effect but 
fewer commits (maybe only one), and then that can be merged (which 
should be fast-forward, no merge commit) and pushed to upstream.  This 
takes a little more work than simply accepting (merging) a pull request 
as-is, but in many cases it may be worth it because it can yield a 
cleaner history.  Similarly, if someone is developing a feature branch 
on github, and the net effect is correct but the branch has intermediate 
commits that distract from the net result, then a good practice would be 
for that person to consolidate the changes into a new feature branch 
with a cleaner history, close the pull request on the old one and open a 
new request for the polished branch.

Regarding your 14406a commit: if I understand correctly, a better 
recovery method would have been to use

git reset --hard HEAD^

to simply back up the branch pointer by one.

The bad commit would still be in your local tree, but it would be 
un-named (that is, it would have no branch name) and would not go 
anywhere else.

Eric


>
> Mike
>
> On 05/17/2011 03:16 PM, Eric Firing wrote:
>> I suspect that the series of May 6 commits ending with
>> a50874b711983cba505ecdb2801c4996eccf3812 has tangled the history in such
>> a way that some (but not all) older pull requests on github like this one
>>
>> https://github.com/matplotlib/matplotlib/pull/1
>>
>> are showing hundreds of commits and diffs.
>>
>> Eric

On Tue, 17 May 2011 16:14:53 -0400, Michael Droettboom wrote:
> Darned if I know what I did differently that time.  (I'm sure I hit
> git's "misunderstanding" feature again).
> 
> It only seems to have affected 2 or 3 of the pull requests -- I suspect
> rebasing those branches off of the current master would fix the problem,
> but maybe there's an easier way.

Looks like some Github quirk:

    git log master..scottza/fix-autofmt_xdate-twinx

shows only four commits.

The merges done after 2011-02-19 seem to have connected a chunk of
previously unconnected 0_91_maint history to master.

On 18/05/2011 5:14 AM, Eric Firing wrote:
> 3) We don't have to always push sets of changes from an original pull
> request to upstream; they can be consolidated using any of a variety of
> methods to form a new local feature branch with the same net effect but
> fewer commits (maybe only one), and then that can be merged (which
> should be fast-forward, no merge commit) and pushed to upstream.  This
> takes a little more work than simply accepting (merging) a pull request
> as-is, but in many cases it may be worth it because it can yield a
> cleaner history.  Similarly, if someone is developing a feature branch
> on github, and the net effect is correct but the branch has intermediate
> commits that distract from the net result, then a good practice would be
> for that person to consolidate the changes into a new feature branch
> with a cleaner history, close the pull request on the old one and open a
> new request for the polished branch.
>
I believe this script more or less automates the process: 
https://github.com/jeresig/pulley

Gerald.

On Tue, May 17, 2011 at 9:07 PM, Gerald Storer <gd...@mr...> wrote:

> On 18/05/2011 5:14 AM, Eric Firing wrote:
> > 3) We don't have to always push sets of changes from an original pull
> > request to upstream; they can be consolidated using any of a variety of
> > methods to form a new local feature branch with the same net effect but
> > fewer commits (maybe only one), and then that can be merged (which
> > should be fast-forward, no merge commit) and pushed to upstream.  This
> > takes a little more work than simply accepting (merging) a pull request
> > as-is, but in many cases it may be worth it because it can yield a
> > cleaner history.  Similarly, if someone is developing a feature branch
> > on github, and the net effect is correct but the branch has intermediate
> > commits that distract from the net result, then a good practice would be
> > for that person to consolidate the changes into a new feature branch
> > with a cleaner history, close the pull request on the old one and open a
> > new request for the polished branch.
> >
> I believe this script more or less automates the process:
> https://github.com/jeresig/pulley
>
> Gerald.
>
>
Don't know if this was a mistake or not, but I see that commit e7f1e83 (the
one to fix a clipping issue when a patch's line width is 1 but there is no
color) seems to have been merged back into itself... somehow... in commit
0c886b8.  I have seen things like this before, and I never quite understood
how they happen.  Plus, do we want to get that patch merged down to master?

I think we definitely need to see what sort of controls we can put in place
to prevent mix-ups in the future.  One thing I did like about SVN was that
it was next to impossible to change the history.  Meanwhile, with git, it
becomes possible.  Is there some way we can disallow forced pushes, maybe?
Just a thought...

Ben Root

On Wed, 18 May 2011 13:47:04 -0500, Benjamin Root wrote:
[clip]
> Don't know if this was a mistake or not, but I see that commit e7f1e83
> (the one to fix a clipping issue when a patch's line width is 1 but
> there is no color) seems to have been merged back into itself...
> somehow... in commit 0c886b8.  I have seen things like this before, and
> I never quite understood how they happen.

"git merge --no-ff" most likely, as it always creates a separate
merge commit, even if the branch in question has only a single commit.

[clip]
> I think we definitely need to see what sort of controls we can put in
> place to prevent mix-ups in the future.  One thing I did like about SVN
> was that it was next to impossible to change the history. Meanwhile,
> with git, it becomes possible. 
> Is there some way we can disallow forced pushes, maybe? Just a thought...

Disabling forced pushes will not help against what occurred here,
at least as far I understand, because the "additional" commits in the
pull requests come from valid merges done after the pull request branched.
The only difference to SVN here is that Git actually tracks the merge
history...

... and, apparently, the Github UI (but not Git itself) gets confused
in some cases as to what's new in a branch.

-- 
Pauli Virtanen

On 05/18/2011 08:47 AM, Benjamin Root wrote:
>
>
> On Tue, May 17, 2011 at 9:07 PM, Gerald Storer <gd...@mr...
> <mailto:gd...@mr...>> wrote:
>
>     On 18/05/2011 5:14 AM, Eric Firing wrote:
>      > 3) We don't have to always push sets of changes from an original pull
>      > request to upstream; they can be consolidated using any of a
>     variety of
>      > methods to form a new local feature branch with the same net
>     effect but
>      > fewer commits (maybe only one), and then that can be merged (which
>      > should be fast-forward, no merge commit) and pushed to upstream.
>       This
>      > takes a little more work than simply accepting (merging) a pull
>     request
>      > as-is, but in many cases it may be worth it because it can yield a
>      > cleaner history.  Similarly, if someone is developing a feature
>     branch
>      > on github, and the net effect is correct but the branch has
>     intermediate
>      > commits that distract from the net result, then a good practice
>     would be
>      > for that person to consolidate the changes into a new feature branch
>      > with a cleaner history, close the pull request on the old one and
>     open a
>      > new request for the polished branch.
>      >
>     I believe this script more or less automates the process:
>     https://github.com/jeresig/pulley
>
>     Gerald.
>
>
> Don't know if this was a mistake or not, but I see that commit e7f1e83
> (the one to fix a clipping issue when a patch's line width is 1 but
> there is no color) seems to have been merged back into itself...
> somehow... in commit 0c886b8.  I have seen things like this before, and
> I never quite understood how they happen.  Plus, do we want to get that
> patch merged down to master?

Yes, it needs to get merged to master; but no, I don't think anything 
was "merged back into itself"; instead, it is just a case where a 
fast-forward with no merge commit would have left a simpler history--one 
commit instead of two.

Merging from v1.0.x to master doesn't have to be done after every change 
to v1.0.x, but it shouldn't be left undone for very long.  A few days 
ago I wasted a chunk of time thrashing around on master because of a bug 
that had been fixed on 1.0.x but was not yet merged into master--and I 
had not thought to check their relative states.  The bug had actually 
been introduced to master via a recent change merged from 1.0.x.

We are experiencing some bumps on the git/github learning curve, but not 
nearly enough to make me pine for svn.

>
> I think we definitely need to see what sort of controls we can put in
> place to prevent mix-ups in the future.  One thing I did like about SVN
> was that it was next to impossible to change the history.  Meanwhile,
> with git, it becomes possible.  Is there some way we can disallow forced
> pushes, maybe?  Just a thought...
>

I suspect the anomalies have not resulted from forced pushes, but from 
local pulls and merges followed by innocuous pushes.  So the key is 
understanding how to ensure one's local branches have the desired 
history before pushing to github.  (And making sure one is pushing from 
the correct source to the correct destination.  Trying first with 
--dry-run can help.)

Eric

> Ben Root
>
>
>
> ------------------------------------------------------------------------------
> What Every C/C++ and Fortran developer Should Know!
> Read this article and learn how Intel has extended the reach of its
> next-generation tools to help Windows* and Linux* C/C++ and Fortran
> developers boost performance applications - including clusters.
> http://p.sf.net/sfu/intel-dev2devmay
>
>
>
> _______________________________________________
> Matplotlib-devel mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/matplotlib-devel

On Wed, May 18, 2011 at 3:22 PM, Eric Firing <ef...@ha...> wrote:
> On 05/18/2011 08:47 AM, Benjamin Root wrote:
>>
>>
>> On Tue, May 17, 2011 at 9:07 PM, Gerald Storer <gd...@mr...
>> <mailto:gd...@mr...>> wrote:
>>
>>     On 18/05/2011 5:14 AM, Eric Firing wrote:
>>      > 3) We don't have to always push sets of changes from an original pull
>>      > request to upstream; they can be consolidated using any of a
>>     variety of
>>      > methods to form a new local feature branch with the same net
>>     effect but
>>      > fewer commits (maybe only one), and then that can be merged (which
>>      > should be fast-forward, no merge commit) and pushed to upstream.
>>       This
>>      > takes a little more work than simply accepting (merging) a pull
>>     request
>>      > as-is, but in many cases it may be worth it because it can yield a
>>      > cleaner history.  Similarly, if someone is developing a feature
>>     branch
>>      > on github, and the net effect is correct but the branch has
>>     intermediate
>>      > commits that distract from the net result, then a good practice
>>     would be
>>      > for that person to consolidate the changes into a new feature branch
>>      > with a cleaner history, close the pull request on the old one and
>>     open a
>>      > new request for the polished branch.
>>      >
>>     I believe this script more or less automates the process:
>>     https://github.com/jeresig/pulley
>>
>>     Gerald.
>>
>>
>> Don't know if this was a mistake or not, but I see that commit e7f1e83
>> (the one to fix a clipping issue when a patch's line width is 1 but
>> there is no color) seems to have been merged back into itself...
>> somehow... in commit 0c886b8.  I have seen things like this before, and
>> I never quite understood how they happen.  Plus, do we want to get that
>> patch merged down to master?
>
> Yes, it needs to get merged to master; but no, I don't think anything
> was "merged back into itself"; instead, it is just a case where a
> fast-forward with no merge commit would have left a simpler history--one
> commit instead of two.
>
> Merging from v1.0.x to master doesn't have to be done after every change
> to v1.0.x, but it shouldn't be left undone for very long.  A few days
> ago I wasted a chunk of time thrashing around on master because of a bug
> that had been fixed on 1.0.x but was not yet merged into master--and I
> had not thought to check their relative states.  The bug had actually
> been introduced to master via a recent change merged from 1.0.x.
>
> We are experiencing some bumps on the git/github learning curve, but not
> nearly enough to make me pine for svn.
>
>>
>> I think we definitely need to see what sort of controls we can put in
>> place to prevent mix-ups in the future.  One thing I did like about SVN
>> was that it was next to impossible to change the history.  Meanwhile,
>> with git, it becomes possible.  Is there some way we can disallow forced
>> pushes, maybe?  Just a thought...
>>
>
> I suspect the anomalies have not resulted from forced pushes, but from
> local pulls and merges followed by innocuous pushes.  So the key is
> understanding how to ensure one's local branches have the desired
> history before pushing to github.  (And making sure one is pushing from
> the correct source to the correct destination.  Trying first with
> --dry-run can help.)

Before pushing, I also recommend inspecting the history graph, either
with "gitk --all" or "git log --oneline --graph --all". I try to
remember to make sure the history graph looks the way I expect it
should before I push anywhere.

On 05/18/2011 11:45 AM, Darren Dale wrote:

>>
>> I suspect the anomalies have not resulted from forced pushes, but from
>> local pulls and merges followed by innocuous pushes.  So the key is
>> understanding how to ensure one's local branches have the desired
>> history before pushing to github.  (And making sure one is pushing from
>> the correct source to the correct destination.  Trying first with
>> --dry-run can help.)
>
> Before pushing, I also recommend inspecting the history graph, either
> with "gitk --all" or "git log --oneline --graph --all". I try to
> remember to make sure the history graph looks the way I expect it
> should before I push anywhere.

QGit is another alternative.

Eric