Today I've taken another close look at branches and gerrit, and have written down my conclusions here:
http://osmocom.org/projects/cellular-infrastructure/wiki/Gerrit#Submitting-B...
In summary:
* Branch re-submission needs the workaround of tweaking the first commit. It would be nice to fix gerrit to not need this, but for the time being the effort of cosmetically changing the first commit log is acceptable.
* It *is* possible to have private branches in the gerrit repos and still be able to submit them to master, with the proper project config.
I have thus switched all our gerrit projects' configs to rebase-if-necessary with the flags as described on the wiki page. (Except for osmo-pcap, where we still disallow content merges.)
I will now go back to pushing my "private" developments to the public repository (gerrit) under the neels/ or sysmocom/ namespaces.
~Neels
On 17 Jun 2016, at 19:47, Neels Hofmeyr nhofmeyr@sysmocom.de wrote:
Hi!
- Branch re-submission needs the workaround of tweaking the first commit.
It would be nice to fix gerrit to not need this, but for the time being the effort of cosmetically changing the first commit log is acceptable.
did you file a bug report for it? It looks like there is also a better fix for the rebase if necessary change. So maybe it is fixing that issue too?
https://bugs.chromium.org/p/gerrit/issues/detail?id=2167&colspec=ID%20Ty...
https://gerrit-review.googlesource.com/#/c/54550/
I will now go back to pushing my "private" developments to the public repository (gerrit) under the neels/ or sysmocom/ namespaces.
cool. thank you for your investigation.
holger
On Fri, Jun 17, 2016 at 08:18:30PM +0200, Holger Freyther wrote:
- Branch re-submission needs the workaround of tweaking the first commit.
It would be nice to fix gerrit to not need this, but for the time being the effort of cosmetically changing the first commit log is acceptable.
did you file a bug report for it?
I coined it a feature request... https://bugs.chromium.org/p/gerrit/issues/detail?id=4210
It looks like there is also a better fix for the rebase if necessary change. So maybe it is fixing that issue too?
The fix in question was already in our gerrit version.
I found our issue is entirely different. I have no idea why your patch helped... I suspect we had some confusion with the "Create a new change for every commit not in the target branch" config item.
But the NullPointer you fixed has since been fixed by someone else on stable-2.12 :) so that was an actual bug.
I commented in abundant detail...
https://bugs.chromium.org/p/gerrit/issues/detail?id=4158
~Neels
On 20 Jun 2016, at 20:21, Neels Hofmeyr nhofmeyr@sysmocom.de wrote:
Hi!
I found our issue is entirely different. I have no idea why your patch helped... I suspect we had some confusion with the "Create a new change for every commit not in the target branch" config item.
But the NullPointer you fixed has since been fixed by someone else on stable-2.12 :) so that was an actual bug.
I have two fixes, one to rebase if necessary and the other the NPE.
I commented in abundant detail...
On Mon, Jun 20, 2016 at 08:22:42PM +0200, Holger Freyther wrote:
I have two fixes, one to rebase if necessary and the other the NPE.
As I said, someone else fixed the NPE slightly differently since: https://gerrit.googlesource.com/gerrit/+/929f7bb7a9c566210d5e1ba54ecc423aa27...
I couldn't reproduce the rebase-if-necessary fix, the problem as explained in the issue remained: https://bugs.chromium.org/p/gerrit/issues/detail?id=4158
~Neels
-
Holger Freyther -
Neels Hofmeyr