sdboyer committed one patch some little time ago.
I just notice some fixes to do, so I create this, and also because that commit do not refer to an issue.
What is pending:
- There is a new vc_op_id field on the {versioncontrol_item_revisions} table(coming from the original {versioncontrol_operation_items} table), but the VersioncontrolOperation class do not have a vc_op_id data member.
- The operation class is not handling correctly the pass of the recently-created vc_op_id at VersioncontrolOperation::insert().
- The commit mentioned also add logic to versioncontrol_update_6303(), but it only wants to add a new hook_update_N() function, so we need to remove that code(IIRC I talked to Sam about this item but neither him nor me committed a fix for it :-p).
I realized this trying to fix #965930: Get the log parser finished. Really.
Preparing the patch
| Comment | File | Size | Author |
|---|---|---|---|
| #7 | 0001-bug-972058-follow-up-Complete-removing-of-versioncon.patch | 14.88 KB | marvil07 |
| #1 | 0001-bug-972058-Fixed-Complete-removing-of-versioncontrol.patch | 2.19 KB | marvil07 |
Comments
Comment #1
marvil07 commentedThe patch :-)
To test this it try to import a repo with git backend, then see that vc_op_id is always 0 at {versioncontrol_item_revisions}, and after the patch it have the right value.
@sdboyer: if you can review the patch it would be great.
Comment #2
marvil07 commentedMmm, we need a bigger patch.
Comment #3
sdboyer commentedGod I hate trying to keep everything up to date with a patch-based workflow. Thanks for catching this, clearly I neglected to commit the second half of what I have locally.
I have to nitpick on your coding standards, though:
should be
One space. I've noticed this on most of the stuff you add, so I dunno if it's what your IDE does, or what, but...
In any case, yep, committed.
Comment #4
marvil07 commentedThanks for committing the change with the formatting issues fixed(it was an ancient inherit from dia2code :-p, my vim does not do anything special), but this still needs work as the grep output mentioned in #2.
Comment #5
sdboyer commentedTackling the VersioncontrolOperation stuff from the grep output in #2 via this issue is not the right approach - it's legacy-ish code, which is being refactored as part of other issues, and we'd end up keeping everything open if we were waiting for all of that. The views stuff I fixed in a separate commit already, and I've been intentionally leaving those broken functions in versioncontrol.module more or less as a reminder that we need to address them. They're going to need to be addressed as part of some further refactoring of how items work.
Comment #6
marvil07 commentedchanging the status as I will be providing a patch for this excluding views as it seems to be fixed somewhere
Comment #7
marvil07 commentedHere the patch, it would be great to receive a review here.
Comment #8
sdboyer commentedGood to go. But if
versioncontrol_fetch_item_commit_operations()gets removed, then we need another issue to be opened for recreating it elsewhere. We might decide we don't want and close it later, and that'd be fine, but if the function is gone from the code then we need an issue as a reminder.Comment #9
sdboyer commentedCommitted this, so that I could incorporate it into the changes in #879858: Unify entity C(R)UD . Follow-up posted at #973922: Reintroduce versioncontrol_fetch_item_commit_operations() behavior.
Comment #10
marvil07 commentedOk