Bug#765156: flashproxy is marked for autoremoval from testing
Hi, Holger Levsen wrote (05 Nov 2014 14:40:06 GMT) : 1. bumping the standards version now is often perceived as unwanted noise by those reviewing the changes to decide whether to let it enter jessie. leave it now, but next time please only include non-cosmetic changes _if_ you add non ron-RC fixes at all. Definitely. See the freeze [policy] and the various messages pointing to it, sent by the release team to d-d-a. [policy] https://release.debian.org/jessie/freeze_policy.html Cheers, -- intrigeri -- To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
Bug#765156: flashproxy is marked for autoremoval from testing
Hi, On Dienstag, 4. November 2014, Ximin Luo wrote: OK, I've uploaded a release candidate to mentors, same address as above. I took a look, and there are two or three problems, though the fix itself is mostly fine I think :-) 1. bumping the standards version now is often perceived as unwanted noise by those reviewing the changes to decide whether to let it enter jessie. leave it now, but next time please only include non-cosmetic changes _if_ you add non ron-RC fixes at all. 2. your fix for #765156 looks good to me, I just wonder whether in the following it really should only be 20 and not 40 or 100... are you sure that safe enough now and in 5 years? ++for i in xrange(0, 20): (I think so, as hw gets faster but... maybe 40 is still better as there could be even slower hw??) 3. there are lot of changes in debian/rules between 1.7-1 and 1.7-2 and there's no mentioning of those in debian/changelog at all. Is there a bug# for the problem they are fixing? Get well soon! thanks, I'm on it..! :-) cheers, Holger signature.asc Description: This is a digitally signed message part.
Bug#765156: flashproxy is marked for autoremoval from testing
On 05/11/14 14:40, Holger Levsen wrote: Hi, On Dienstag, 4. November 2014, Ximin Luo wrote: OK, I've uploaded a release candidate to mentors, same address as above. I took a look, and there are two or three problems, though the fix itself is mostly fine I think :-) 1. bumping the standards version now is often perceived as unwanted noise by those reviewing the changes to decide whether to let it enter jessie. leave it now, but next time please only include non-cosmetic changes _if_ you add non ron-RC fixes at all. OK. 2. your fix for #765156 looks good to me, I just wonder whether in the following it really should only be 20 and not 40 or 100... are you sure that safe enough now and in 5 years? ++for i in xrange(0, 20): (I think so, as hw gets faster but... maybe 40 is still better as there could be even slower hw??) The previous sleep time was 0.1s, and now it will sleep for a maximum of 1s (10 times the previous), so I think this should be OK. 3. there are lot of changes in debian/rules between 1.7-1 and 1.7-2 and there's no mentioning of those in debian/changelog at all. Is there a bug# for the problem they are fixing? I just noticed that I wasn't running some of the tests, that's all it is (and some variable renames for consistency). So I figured this is too trivial to put in debian/changelog, and there is no effect on the binary packages. X -- GPG: 4096R/1318EFAC5FBBDBCE git://github.com/infinity0/pubkeys.git signature.asc Description: OpenPGP digital signature
Bug#765156: flashproxy is marked for autoremoval from testing
Hi, On Mittwoch, 5. November 2014, Ximin Luo wrote: 1. bumping the standards version now is often perceived as unwanted noise by those reviewing the changes to decide whether to let it enter jessie. leave it now, but next time please only include non-cosmetic changes _if_ you add non ron-RC fixes at all. OK. :-) 2. [...] The previous sleep time was 0.1s, and now it will sleep for a maximum of 1s (10 times the previous), so I think this should be OK. ok, cool, thanks. 3. there are lot of changes in debian/rules between 1.7-1 and 1.7-2 and there's no mentioning of those in debian/changelog at all. Is there a bug# for the problem they are fixing? I just noticed that I wasn't running some of the tests, that's all it is (and some variable renames for consistency). So I figured this is too trivial to put in debian/changelog, and there is no effect on the binary packages. please describe this in debian/changelog, re-upload to mentors and I'll happily upload to sid! basically any changes should be somehow mentioned in debian/changelog and at this time such changes certainly..! Thanks! cheers, Holger signature.asc Description: This is a digitally signed message part.
Bug#765156: flashproxy is marked for autoremoval from testing
On 05/11/14 17:01, Holger Levsen wrote: please describe this in debian/changelog, re-upload to mentors and I'll happily upload to sid! basically any changes should be somehow mentioned in debian/changelog and at this time such changes certainly..! Thanks! Done! https://mentors.debian.net/package/flashproxy X -- GPG: 4096R/1318EFAC5FBBDBCE git://github.com/infinity0/pubkeys.git signature.asc Description: OpenPGP digital signature
Bug#765156: flashproxy is marked for autoremoval from testing
Hi, On Dienstag, 4. November 2014, Ximin Luo wrote: OK, so it turns out you can do this in automake easily with a flag. I've added it to debian/rules. cool Candidate for 1.7-2: https://mentors.debian.net/package/flashproxy Let me know if it works, and I'll upload a proper release. I wont check today or probably tomorrow anyway, I'm sick with fever atm, so why don't you make it a release anyway and once I'm better I will look at probably just upload. That also saves us another roundtrip... cheers, Holger signature.asc Description: This is a digitally signed message part.
Bug#765156: flashproxy is marked for autoremoval from testing
On 04/11/14 11:54, Holger Levsen wrote: Candidate for 1.7-2: https://mentors.debian.net/package/flashproxy Let me know if it works, and I'll upload a proper release. I wont check today or probably tomorrow anyway, I'm sick with fever atm, so why don't you make it a release anyway and once I'm better I will look at probably just upload. That also saves us another roundtrip... OK, I've uploaded a release candidate to mentors, same address as above. Get well soon! X -- GPG: 4096R/1318EFAC5FBBDBCE git://github.com/infinity0/pubkeys.git signature.asc Description: OpenPGP digital signature
Bug#765156: flashproxy is marked for autoremoval from testing
On 02/11/14 21:33, Ximin Luo wrote: On 01/11/14 18:41, Holger Levsen wrote: well, you are the maintainer you can make it output the test.log if there are failures... It's pretty tricky to do this and I don't want to commit a hack that bypasses automake. I've filed a feature request upstream to automake: http://debbugs.gnu.org/cgi/pkgreport.cgi?pkg=automake Hopefully it will show up soon, I called in feature: output test failures to stdout. OK, so it turns out you can do this in automake easily with a flag. I've added it to debian/rules. OTOH ideally, the CI system would let me view these test logs as well. It's a bit unhelpful to give partial information when there's a problem. I didn't choose to use these systems, and there is no Debian policy that states packages should output test logs to stdout; whoever set these systems up should be working to make sure it's as useful for as many packages as possible. Apparently there is a way to get jenkins to keep artefacts: http://stackoverflow.com/questions/9280619/how-can-i-configure-jenkins-to-show-my-logs-in-the-workspace The pattern suggested, **/*.log seems pretty reasonable and should cover most test logs, without picking up large build products like object files etc. it's in ./facilitator/test-suite.log and attached. :) I'd be happy to rebuild again if you can give me something to try :) Thanks, that was very useful. I think I have a fix - try this? https://gitweb.torproject.org/debian/flashproxy.git/commitdiff/fedcd28195f03fdc9bb456b7e16d9c14b3648478 Candidate for 1.7-2: https://mentors.debian.net/package/flashproxy Let me know if it works, and I'll upload a proper release. X -- GPG: 4096R/1318EFAC5FBBDBCE git://github.com/infinity0/pubkeys.git signature.asc Description: OpenPGP digital signature
Bug#765156: flashproxy is marked for autoremoval from testing
On 01/11/14 18:41, Holger Levsen wrote: well, you are the maintainer you can make it output the test.log if there are failures... It's pretty tricky to do this and I don't want to commit a hack that bypasses automake. I've filed a feature request upstream to automake: http://debbugs.gnu.org/cgi/pkgreport.cgi?pkg=automake Hopefully it will show up soon, I called in feature: output test failures to stdout. OTOH ideally, the CI system would let me view these test logs as well. It's a bit unhelpful to give partial information when there's a problem. I didn't choose to use these systems, and there is no Debian policy that states packages should output test logs to stdout; whoever set these systems up should be working to make sure it's as useful for as many packages as possible. Apparently there is a way to get jenkins to keep artefacts: http://stackoverflow.com/questions/9280619/how-can-i-configure-jenkins-to-show-my-logs-in-the-workspace The pattern suggested, **/*.log seems pretty reasonable and should cover most test logs, without picking up large build products like object files etc. it's in ./facilitator/test-suite.log and attached. :) I'd be happy to rebuild again if you can give me something to try :) Thanks, that was very useful. I think I have a fix - try this? https://gitweb.torproject.org/debian/flashproxy.git/commitdiff/fedcd28195f03fdc9bb456b7e16d9c14b3648478 X -- GPG: 4096R/1318EFAC5FBBDBCE git://github.com/infinity0/pubkeys.git signature.asc Description: OpenPGP digital signature
Bug#765156: flashproxy is marked for autoremoval from testing
On 28/10/14 11:19, Ximin Luo wrote: On 28/10/14 10:58, Holger Levsen wrote: Hi, On Dienstag, 28. Oktober 2014, Holger Levsen wrote: FWIW, https://jenkins.debian.net/userContent/rb-pkg/flashproxy.html shows that it build fine there on 2014-10-13 19:14 on amd64 too. I've scheduled another build there (should be visible in 15min roughly) to see if this is still the case... https://jenkins.debian.net/userContent/rb-pkg/flashproxy.html indeed shows the build failure now (but obviously not the test.log) cheers, Holger ACK, I will sort it out this weekend. I've so far been unable to reproduce this build failure. I've tried doing it through cowbuilder, disconnecting from the internet, setting `net.ipv6.bindv6only = 1` and `sudo sysctl -p`, and in every case the package builds fine. Can you, or anyone else, reproduce it locally? If not, I will have to request access to a build porter machine to try to reproduce it. In the meantime I will probably downgrade the severity of this bug so it doesn't get autoremoved. X -- GPG: 4096R/1318EFAC5FBBDBCE git://github.com/infinity0/pubkeys.git signature.asc Description: OpenPGP digital signature
Bug#765156: flashproxy is marked for autoremoval from testing
Hi, On Samstag, 1. November 2014, Ximin Luo wrote: I've so far been unable to reproduce this build failure. I've tried doing it through cowbuilder, disconnecting from the internet, setting `net.ipv6.bindv6only = 1` and `sudo sysctl -p`, and in every case the package builds fine. you should send this to the bug. Can you, or anyone else, reproduce it locally? If not, I will have to request access to a build porter machine to try to reproduce it. it fails to build here too: https://jenkins.debian.net/userContent/rb-pkg/flashproxy.html In the meantime I will probably downgrade the severity of this bug so it doesn't get autoremoved. before I saw this failure on jenkins.d.n I was going to say sounds reasonable, now I'm not so sure anymore ;-) OTOH, fails to build sometimes could be justified as just an important bug esp as it should be fixed no matter the severity :) cheers, Holger signature.asc Description: This is a digitally signed message part.
Bug#765156: flashproxy is marked for autoremoval from testing
On 01/11/14 18:00, Holger Levsen wrote: Hi, On Samstag, 1. November 2014, Ximin Luo wrote: I've so far been unable to reproduce this build failure. I've tried doing it through cowbuilder, disconnecting from the internet, setting `net.ipv6.bindv6only = 1` and `sudo sysctl -p`, and in every case the package builds fine. you should send this to the bug. It is sent to the bug. Can you, or anyone else, reproduce it locally? If not, I will have to request access to a build porter machine to try to reproduce it. it fails to build here too: https://jenkins.debian.net/userContent/rb-pkg/flashproxy.html Yes, but it gives me no actionable information. Great, my build is failing, I can't even see the test output or the build environment. :/ In the meantime I will probably downgrade the severity of this bug so it doesn't get autoremoved. before I saw this failure on jenkins.d.n I was going to say sounds reasonable, now I'm not so sure anymore ;-) OTOH, fails to build sometimes could be justified as just an important bug esp as it should be fixed no matter the severity :) It's a slap in the face for my package to be auto-removed, without me even having a chance to fix things. I can't reproduce the bug locally and the build machines don't even output test-suite.log, which is a standard autotools test output file. Can you try building it locally? X -- GPG: 4096R/1318EFAC5FBBDBCE git://github.com/infinity0/pubkeys.git signature.asc Description: OpenPGP digital signature
Bug#765156: flashproxy is marked for autoremoval from testing
Hi, On Samstag, 1. November 2014, Ximin Luo wrote: you should send this to the bug. It is sent to the bug. right, here we're talking... Yes, but it gives me no actionable information. Great, my build is failing, I can't even see the test output or the build environment. :/ well, you are the maintainer you can make it output the test.log if there are failures... It's a slap in the face for my package to be auto-removed, without me even having a chance to fix things. I can't reproduce the bug locally and the build machines don't even output test-suite.log, which is a standard autotools test output file. Can you try building it locally? I did, but no log :( here's how I build: 504 apt-get source flashproxy 506 time ionice -c 3 nice sudo DEB_BUILD_OPTIONS=-j15 -b pbuilder --build flashproxy_1.7-1.dsc f.log 21 and then I build it manually... and it failed and I saw this: See ./test-suite.log which is wrong. it's not there. it's in ./facilitator/test-suite.log and attached. :) I'd be happy to rebuild again if you can give me something to try :) cheers, Holger == flashproxy-facilitator 1.7: ./test-suite.log == # TOTAL: 1 # PASS: 0 # SKIP: 0 # XFAIL: 0 # FAIL: 1 # XPASS: 0 # ERROR: 0 .. contents:: :depth: 2 FAIL: fp-facilitator-test = ...E.E... == ERROR: test_af_v4_v6 (__main__.FacilitatorProcTest) Test that IPv4 proxies do not get IPv6 clients. -- Traceback (most recent call last): File ./fp-facilitator-test.py, line 257, in test_af_v4_v6 fac.put_reg(FACILITATOR_ADDR, self.IPV6_CLIENT_ADDR, CLIENT_TP) File /srv/reproducible-results/f/flashproxy-1.7/flashproxy/fac.py, line 147, in put_reg f = fac_socket(facilitator_addr) File /srv/reproducible-results/f/flashproxy-1.7/flashproxy/fac.py, line 132, in fac_socket return socket.create_connection(facilitator_addr, 1.0).makefile() File /usr/lib/python2.7/socket.py, line 571, in create_connection raise err error: [Errno 111] Connection refused == ERROR: test_af_v6_v6 (__main__.FacilitatorProcTest) Test that IPv6 proxies can get IPv6 clients. -- Traceback (most recent call last): File ./fp-facilitator-test.py, line 269, in test_af_v6_v6 fac.put_reg(FACILITATOR_ADDR, self.IPV4_CLIENT_ADDR, CLIENT_TP) File /srv/reproducible-results/f/flashproxy-1.7/flashproxy/fac.py, line 147, in put_reg f = fac_socket(facilitator_addr) File /srv/reproducible-results/f/flashproxy-1.7/flashproxy/fac.py, line 132, in fac_socket return socket.create_connection(facilitator_addr, 1.0).makefile() File /usr/lib/python2.7/socket.py, line 571, in create_connection raise err error: [Errno 111] Connection refused -- Ran 17 tests in 2.204s FAILED (errors=2) signature.asc Description: This is a digitally signed message part.
Bug#765156: flashproxy is marked for autoremoval from testing
Hi, FWIW, https://jenkins.debian.net/userContent/rb-pkg/flashproxy.html shows that it build fine there on 2014-10-13 19:14 on amd64 too. I've scheduled another build there (should be visible in 15min roughly) to see if this is still the case... cheers, Holger signature.asc Description: This is a digitally signed message part.
Bug#765156: flashproxy is marked for autoremoval from testing
Hi, On Dienstag, 28. Oktober 2014, Holger Levsen wrote: FWIW, https://jenkins.debian.net/userContent/rb-pkg/flashproxy.html shows that it build fine there on 2014-10-13 19:14 on amd64 too. I've scheduled another build there (should be visible in 15min roughly) to see if this is still the case... https://jenkins.debian.net/userContent/rb-pkg/flashproxy.html indeed shows the build failure now (but obviously not the test.log) cheers, Holger signature.asc Description: This is a digitally signed message part.
Bug#765156: flashproxy is marked for autoremoval from testing
On 28/10/14 10:58, Holger Levsen wrote: Hi, On Dienstag, 28. Oktober 2014, Holger Levsen wrote: FWIW, https://jenkins.debian.net/userContent/rb-pkg/flashproxy.html shows that it build fine there on 2014-10-13 19:14 on amd64 too. I've scheduled another build there (should be visible in 15min roughly) to see if this is still the case... https://jenkins.debian.net/userContent/rb-pkg/flashproxy.html indeed shows the build failure now (but obviously not the test.log) cheers, Holger ACK, I will sort it out this weekend. X -- GPG: 4096R/1318EFAC5FBBDBCE git://github.com/infinity0/pubkeys.git signature.asc Description: OpenPGP digital signature
