|
| 1 | +.. SPDX-License-Identifier: GPL-2.0 |
| 2 | +
|
| 3 | +===================================== |
| 4 | +Handling messy pull-request diffstats |
| 5 | +===================================== |
| 6 | + |
| 7 | +Subsystem maintainers routinely use ``git request-pull`` as part of the |
| 8 | +process of sending work upstream. Normally, the result includes a nice |
| 9 | +diffstat that shows which files will be touched and how much of each will |
| 10 | +be changed. Occasionally, though, a repository with a relatively |
| 11 | +complicated development history will yield a massive diffstat containing a |
| 12 | +great deal of unrelated work. The result looks ugly and obscures what the |
| 13 | +pull request is actually doing. This document describes what is happening |
| 14 | +and how to fix things up; it is derived from The Wisdom of Linus Torvalds, |
| 15 | +found in Linus1_ and Linus2_. |
| 16 | + |
| 17 | +.. _Linus1: https://lore.kernel.org/lkml/CAHk-=wg3wXH2JNxkQi+eLZkpuxqV+wPiHhw_Jf7ViH33Sw7PHA@mail.gmail.com/ |
| 18 | +.. _Linus2: https://lore.kernel.org/lkml/CAHk-=wgXbSa8yq8Dht8at+gxb_idnJ7X5qWZQWRBN4_CUPr=eQ@mail.gmail.com/ |
| 19 | + |
| 20 | +A Git development history proceeds as a series of commits. In a simplified |
| 21 | +manner, mainline kernel development looks like this:: |
| 22 | + |
| 23 | + ... vM --- vN-rc1 --- vN-rc2 --- vN-rc3 --- ... --- vN-rc7 --- vN |
| 24 | + |
| 25 | +If one wants to see what has changed between two points, a command like |
| 26 | +this will do the job:: |
| 27 | + |
| 28 | + $ git diff --stat --summary vN-rc2..vN-rc3 |
| 29 | + |
| 30 | +Here, there are two clear points in the history; Git will essentially |
| 31 | +"subtract" the beginning point from the end point and display the resulting |
| 32 | +differences. The requested operation is unambiguous and easy enough to |
| 33 | +understand. |
| 34 | + |
| 35 | +When a subsystem maintainer creates a branch and commits changes to it, the |
| 36 | +result in the simplest case is a history that looks like:: |
| 37 | + |
| 38 | + ... vM --- vN-rc1 --- vN-rc2 --- vN-rc3 --- ... --- vN-rc7 --- vN |
| 39 | + | |
| 40 | + +-- c1 --- c2 --- ... --- cN |
| 41 | + |
| 42 | +If that maintainer now uses ``git diff`` to see what has changed between |
| 43 | +the mainline branch (let's call it "linus") and cN, there are still two |
| 44 | +clear endpoints, and the result is as expected. So a pull request |
| 45 | +generated with ``git request-pull`` will also be as expected. But now |
| 46 | +consider a slightly more complex development history:: |
| 47 | + |
| 48 | + ... vM --- vN-rc1 --- vN-rc2 --- vN-rc3 --- ... --- vN-rc7 --- vN |
| 49 | + | | |
| 50 | + | +-- c1 --- c2 --- ... --- cN |
| 51 | + | / |
| 52 | + +-- x1 --- x2 --- x3 |
| 53 | + |
| 54 | +Our maintainer has created one branch at vN-rc1 and another at vN-rc2; the |
| 55 | +two were then subsequently merged into c2. Now a pull request generated |
| 56 | +for cN may end up being messy indeed, and developers often end up wondering |
| 57 | +why. |
| 58 | + |
| 59 | +What is happening here is that there are no longer two clear end points for |
| 60 | +the ``git diff`` operation to use. The development culminating in cN |
| 61 | +started in two different places; to generate the diffstat, ``git diff`` |
| 62 | +ends up having pick one of them and hoping for the best. If the diffstat |
| 63 | +starts at vN-rc1, it may end up including all of the changes between there |
| 64 | +and the second origin end point (vN-rc2), which is certainly not what our |
| 65 | +maintainer had in mind. With all of that extra junk in the diffstat, it |
| 66 | +may be impossible to tell what actually happened in the changes leading up |
| 67 | +to cN. |
| 68 | + |
| 69 | +Maintainers often try to resolve this problem by, for example, rebasing the |
| 70 | +branch or performing another merge with the linus branch, then recreating |
| 71 | +the pull request. This approach tends not to lead to joy at the receiving |
| 72 | +end of that pull request; rebasing and/or merging just before pushing |
| 73 | +upstream is a well-known way to get a grumpy response. |
| 74 | + |
| 75 | +So what is to be done? The best response when confronted with this |
| 76 | +situation is to indeed to do a merge with the branch you intend your work |
| 77 | +to be pulled into, but to do it privately, as if it were the source of |
| 78 | +shame. Create a new, throwaway branch and do the merge there:: |
| 79 | + |
| 80 | + ... vM --- vN-rc1 --- vN-rc2 --- vN-rc3 --- ... --- vN-rc7 --- vN |
| 81 | + | | | |
| 82 | + | +-- c1 --- c2 --- ... --- cN | |
| 83 | + | / | | |
| 84 | + +-- x1 --- x2 --- x3 +------------+-- TEMP |
| 85 | + |
| 86 | +The merge operation resolves all of the complications resulting from the |
| 87 | +multiple beginning points, yielding a coherent result that contains only |
| 88 | +the differences from the mainline branch. Now it will be possible to |
| 89 | +generate a diffstat with the desired information:: |
| 90 | + |
| 91 | + $ git diff -C --stat --summary linus..TEMP |
| 92 | + |
| 93 | +Save the output from this command, then simply delete the TEMP branch; |
| 94 | +definitely do not expose it to the outside world. Take the saved diffstat |
| 95 | +output and edit it into the messy pull request, yielding a result that |
| 96 | +shows what is really going on. That request can then be sent upstream. |
0 commit comments