[wip] Job id prefix (jid) is now using travis numbers, shoudl work for PRs … #279
Conversation
ssbarnea
force-pushed
the
feature/travis
branch
4 times, most recently
from
3a19ec4
to
9004766
Compare
…too. Change-Id: I17bb2e4ab0651edb6aad7b1cad0c9d65c7ebfb63 Signed-off-by: Sorin Sbarnea <ssbarnea@redhat.com>
ssbarnea
force-pushed
the
feature/travis
branch
from
9004766
to
769e4bd
Compare
| Worklog, IssueLink, IssueLinkType, IssueType, Priority, Version, Role, Resolution, SecurityLevel, Status, User, \ | ||
| CustomFieldOption, RemoteLink | ||
| from jira.resources import * # NOQA | ||
| import xml.etree.ElementTree as etree |
There was a problem hiding this comment.
Please use defusedxml to access ElementTree instead of directly using it.
ssbarnea
force-pushed
the
feature/travis
branch
10 times, most recently
from
59d8ffc
to
ccf1291
Compare
Change-Id: I4c069a777986452ea275417f7526dd24291f41a4 Signed-off-by: Sorin Sbarnea <ssbarnea@redhat.com>
ssbarnea
force-pushed
the
feature/travis
branch
from
ccf1291
to
2dcbc4e
Compare
| getpass.getuser().upper()))[0:6] + \ | ||
| str(sys.version_info[0]) + \ | ||
| str(sys.version_info[1]) | ||
| """ `jid` is important for avoiding concurency problems when |
There was a problem hiding this comment.
Looks like the comments above no longer reflect what the code is doing, which is not helpful :-(
| tests outside Travis | ||
| * Travis is using "Travis" username | ||
|
|
||
| https://docs.travis-ci.com/user/environment-variables/ |
There was a problem hiding this comment.
I'm confused. That page says not to depend on the USER being travis...
| if user == 'TRAVIS' and 'TRAVIS_JOB_NUMBER' in os.environ: | ||
| # please note that user underline (_) is not suppored by | ||
| # jira even if is documented as supported. | ||
| self.jid = 'T' + hashify(user + os.environ['TRAVIS_JOB_NUMBER']) |
There was a problem hiding this comment.
I guess we don't need to bring travis user into it if we're namespacing the jid with T for travis and Z for others?
Also, TRAVIS_JOB_NUMBER is going to be something like 531.3 - why do we even need to hash this to 8 digits? Can't we just replace the '.' and have a much more human-readable project key?
| identifier = user + \ | ||
| chr(ord('A') + sys.version_info[0]) + \ | ||
| chr(ord('A') + sys.version_info[1]) | ||
| self.jid = 'Z' + hashify(identifier) |
There was a problem hiding this comment.
Similarly, do we really need to hash this? With the old ZUSERNAME35 the only way we'll clash is if two users that start with the same 8 letters are running tests simultaneously (on the same version of python). Do we have any evidence of this being a problem?
And even if it was, it would never affect Travis because of the Z prefix, which is the important part...
Job id prefix (jid) is now using travis numbers, should work for PRs …
…too.
Change-Id: I17bb2e4ab0651edb6aad7b1cad0c9d65c7ebfb63
Signed-off-by: Sorin Sbarnea ssbarnea@redhat.com