Causing an issue for network package I maintain.
-- Jimmy
It does not seem that any of the new Tumbleweed snapshots have fixed the boost version issue in https://build.opensuse.org/package/show/network/dogecoin. I triggered a manual rebuild to be sure (assuming that would work).
It continues to build against boost 5.6 while 5.8 is provided in tumbleweed and thus the package cannot be installed due to unsatisfied dependencies.
As discussed with darix in #opensuse-buildservice last week it seems the network repository is using factory snapshots.
<repository name="openSUSE_Tumbleweed"> <path project="openSUSE:Factory" repository="snapshot"/> <arch>i586</arch> <arch>x86_64</arch> </repository>
From: https://build.opensuse.org/project/meta/network
Can anyone shed some light on this?
-- Jimmy
On Wed, May 27, 2015 at 5:21 PM, Jimmy Berry jimmy@boombatower.com wrote:
Causing an issue for network package I maintain.
-- Jimmy
On Tuesday 2015-06-02 07:19, Jimmy Berry wrote:
It does not seem that any of the new Tumbleweed snapshots have fixed the boost version issue in https://build.opensuse.org/package/show/network/dogecoin. I triggered a manual rebuild to be sure (assuming that would work).
It continues to build against boost 5.6 while 5.8 is provided in As discussed with darix in #opensuse-buildservice last week it seems the network repository is using factory snapshots.
Precisely that. And snapshot only contains 1.56. Not 5.6, and not 5.8.
So what's the solution? Presumably the snapshot needs an updated version of boost to match tumbleweed?
-- Jimmy
On Tue, Jun 2, 2015 at 2:43 AM, Jan Engelhardt jengelh@inai.de wrote:
On Tuesday 2015-06-02 07:19, Jimmy Berry wrote:
It does not seem that any of the new Tumbleweed snapshots have fixed the boost version issue in https://build.opensuse.org/package/show/network/dogecoin. I triggered a manual rebuild to be sure (assuming that would work).
It continues to build against boost 5.6 while 5.8 is provided in As discussed with darix in #opensuse-buildservice last week it seems the network repository is using factory snapshots.
Precisely that. And snapshot only contains 1.56. Not 5.6, and not 5.8.
To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On Tuesday 2015-06-02 19:09, Jimmy Berry wrote:
So what's the solution? Presumably the snapshot needs an updated version of boost to match tumbleweed?
Correct.
So...obviously I need that to happen. Several snapshots have come out and not resolved the problem.
Seems like we are beating around the bush. So what needs to happen?
-- Jimmy
On Wed, Jun 3, 2015 at 10:00 AM, Jan Engelhardt jengelh@inai.de wrote:
On Tuesday 2015-06-02 19:09, Jimmy Berry wrote:
So what's the solution? Presumably the snapshot needs an updated version of boost to match tumbleweed?
Correct.
On Wednesday 2015-06-03 23:20, Jimmy Berry wrote:
So...obviously I need that to happen. Several snapshots have come out and not resolved the problem.
Seems like we are beating around the bush. So what needs to happen?
Coolo needs to come out of vacation, I guess? TW snapshots are not the same as the "snapshot" repository of openSUSE:Factory. Or something. It was once explained in some mail how it works, but I forgot. Frankly, if it is not reasonably self-explanatory, it's still not good enough :-/
On 2015-06-03 23:30, Jan Engelhardt wrote:
TW snapshots are not the same as the "snapshot" repository of openSUSE:Factory.
Basically, factory is untested. TW is tested (automated tests).
On June 4, 2015 6:05:15 AM EDT, "Carlos E. R." robin.listas@telefonica.net wrote:
On 2015-06-03 23:30, Jan Engelhardt wrote:
TW snapshots are not the same as the "snapshot" repository of openSUSE:Factory.
Basically, factory is untested. TW is tested (automated tests).
Agreed. Factory snapshots are made every day. If and only if a factory snapshot passes aotoqa it is promoted to a tumbleweed snapshot.
Greg
Based on that explanation I am confused how this situation could have arose. As I understand it the following would have had to occur for boost to be updated in tumblweed.
- factory snapshot created - automated tests run against factory snapshot - tests passed and tumbleweed snapshot released
As such how could the network repository be compiling against factory snapshot that does not contain updated boost like tumbleweed?
<path project="openSUSE:Factory" repository="snapshot"/>
Either way, just looking to get this resolved. Also would it not make more sense for the tumbleweed repository in network to actually build against tumbleweed as seems to be the default configuration if creating a new project.
<repository name="openSUSE_Tumbleweed"> <path project="openSUSE:Tumbleweed" repository="standard"/> ... </repository>
-- Jimmy
On Thu, Jun 4, 2015 at 7:03 AM, greg.freemyer@gmail.com wrote:
On June 4, 2015 6:05:15 AM EDT, "Carlos E. R." robin.listas@telefonica.net wrote:
On 2015-06-03 23:30, Jan Engelhardt wrote:
TW snapshots are not the same as the "snapshot" repository of openSUSE:Factory.
Basically, factory is untested. TW is tested (automated tests).
Agreed. Factory snapshots are made every day. If and only if a factory snapshot passes aotoqa it is promoted to a tumbleweed snapshot.
Greg
Sent from my Android device with K-9 Mail. Please excuse my brevity.
To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org To contact the owner, e-mail: opensuse-factory+owner@opensuse.org
On June 4, 2015 8:45:19 PM EDT, Jimmy Berry jimmy@boombatower.com wrote:
Based on that explanation I am confused how this situation could have arose. As I understand it the following would have had to occur for boost to be updated in tumblweed.
- factory snapshot created
- automated tests run against factory snapshot
- tests passed and tumbleweed snapshot released
As such how could the network repository be compiling against factory snapshot that does not contain updated boost like tumbleweed?
<path project="openSUSE:Factory" repository="snapshot"/>
Either way, just looking to get this resolved. Also would it not make more sense for the tumbleweed repository in network to actually build against tumbleweed as seems to be the default configuration if creating a new project.
<repository name="openSUSE_Tumbleweed"> <path project="openSUSE:Tumbleweed" repository="standard"/> ... </repository>
Changing the "tumbleweed" repo in network to actually be the tumbleweed repo is the right thing to do.
The tumbleweed repo only came to exist 6 or 7 months ago and not everything has been redirected to the relatively new tumbleweed repo yet.
I don't have the credentials to do that.
I suggest you open a new thread asking for that to be done. If you're subscribed to -buildservice that might be a more productive list for that request than -factory.
Greg
On Thu, 2015-06-04 at 19:45 -0500, Jimmy Berry wrote:
Based on that explanation I am confused how this situation could have arose. As I understand it the following would have had to occur for boost to be updated in tumblweed.
- factory snapshot created
- automated tests run against factory snapshot
- tests passed and tumbleweed snapshot released
We have two things called 'snapshot' which is what the confusion is.
1) the 'snapshots' of Tumbleweed that are released to the public. That's once packages are checked into Factory, they pass openQA, then we release a snapshot 2) openSUSE:Factory/snapshot (a more static build target compared to openSUSE:Factory/standard). This is totally unrelated to the snapshots above (even though the name might imply). Based on various parameters (the main devel repos being idle, openSUSE:Factory build completed, the right machine/script running), this snapshot is being updated. At the moment, though, the last part is an issue (the machine running the script disappeared on us). As far as I know, this was running on coolo's machine, who happens to have gone on holiday (back next week). This is certainly a script we will be moving to a more central place asap.
As coolo will be back soon, we can expect this process to be recovering normal operation as well.
Not the help expected for now, but at least this gives some backgrounds what went wrong and we have a plan to correct the situation.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1
On 05.06.2015 10:17, Dimstar / Dominique Leuenberger wrote:
On Thu, 2015-06-04 at 19:45 -0500, Jimmy Berry wrote:
Based on that explanation I am confused how this situation could have arose. As I understand it the following would have had to occur for boost to be updated in tumblweed.
- factory snapshot created - automated tests run against factory
snapshot - tests passed and tumbleweed snapshot released
We have two things called 'snapshot' which is what the confusion is.
- the 'snapshots' of Tumbleweed that are released to the public.
That's once packages are checked into Factory, they pass openQA, then we release a snapshot 2) openSUSE:Factory/snapshot (a more static build target compared to openSUSE:Factory/standard). This is totally unrelated to the snapshots above (even though the name might imply). Based on various parameters (the main devel repos being idle, openSUSE:Factory build completed, the right machine/script running), this snapshot is being updated. At the moment, though, the last part is an issue (the machine running the script disappeared on us). As far as I know, this was running on coolo's machine, who happens to have gone on holiday (back next week). This is certainly a script we will be moving to a more central place asap.
The cron job most likely did run, but the conditions when to update the repository were most likely not met.
A (random) selection of 10 big and popular OBS repos has to be 100% finished so the script considers the OBS build farm to be idle enough to push the next round.
The problem is that sometimes a single package does not stop to build and as such halts the full process. I consider the whole snapshot repository an obsolete construct though - openSUSE:Tumbleweed should be what is published, but that unfortunately requires a feature in the OBS backend a bit more complex to implement ;(
Greetings, Stephan