| Summary: | check for real signed-off-by names | ||
|---|---|---|---|
| Product: | [Yocto Project Subprojects] Patchwork/Patchtest | Reporter: | Leonardo Sandoval Gonzalez <leonardo.sandoval.gonzalez> |
| Component: | Patchtest | Assignee: | Trevor Gamblin <tgamblin> |
| Status: | RESOLVED WORKSFORME | QA Contact: | |
| Severity: | enhancement | ||
| Priority: | Medium | CC: | bluelightning, randy.macleod, richard.purdie, ross.burton, tgamblin, tim.orling |
| Version: | 2.4 | ||
| Target Milestone: | Future | ||
| Hardware: | x86 | ||
| OS: | Multiple | ||
| Whiteboard: | |||
| OS type for building Yocto: | --- | Type of Regression: | --- |
| Verified: | Documentation change: | No (bug/feature does not impact docs) | |
|
Description
Leonardo Sandoval Gonzalez
2017-05-17 14:25:33 UTC
Moving to M4 waiting for an execution issue to be solved (bitbake/tinfoil related). For this particular enhacement, there is no need to wait for the tinfoil changes so there is no blocker. Moving to 2.5 M1, I'll work on all enhancements on such milestone. Due to time constraints and the importance of this enhancement (this check is not as important as other ones, no need to rush into a solution for now), moving it to Future. the problem here is to define what is a 'correct' name. So in my opinion, this is an invalid check. Sorry for the noise. When we opened this the original discussion was around people accidentally submitting patches from the AUH with "Upgrade Helper" as the committer. I have also see people sending patches with just "one" name - usually the first part of their email address - which is not ideal. I'd still like to see this called out by patchtest (and we of course would have the option of just ignoring the message and merging the patch anyway, if for example the submitter does only have one name), so I'm reopening this. (In reply to comment #6) > When we opened this the original discussion was around people accidentally > submitting patches from the AUH with "Upgrade Helper" as the committer. I for this particular case, we have already one check http://git.yoctoproject.org/cgit/cgit.cgi/patchtest-oe/tree/tests/test_mbox_author.py > have also see people sending patches with just "one" name - usually the > first part of their email address - which is not ideal. ok, names with a single string could be bad but I wont bet this is universal. > > I'd still like to see this called out by patchtest (and we of course would > have the option of just ignoring the message and merging the patch anyway, > if for example the submitter does only have one name), so I'm reopening this. I'd be very careful about trying to detect "real names". Aside from the fact that someone could literally have a full legal name that is just one word, we really only care that the name is "identifiable". So detecting Auto Upgrade Helper is good, but automatically determining that a name is not "real" is not. For more context, see https://github.com/torvalds/linux/commit/d4563201f33a022fc0353033d9dfeb1606a88330. I'm thinking we should check for known tools (like AUH, which this does already) and just let humans decide if eg S-o-b: Wookey <wookey@wookware.org> is a "real" name or not (hint: that's his legal full name). Patchtest has a basic name/email validity check that includes ensuring the submitter is not AUH. I think this is sufficient for the purposes of this bug (especially given how old it is), but since I didn't actually change that myself I'm just marking this as WORKSFORME. If we run into name issues in the future then we can certainly revisit this. |