-
Notifications
You must be signed in to change notification settings - Fork 148
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
DOL module vs. technical debt: generics #24082
DOL module vs. technical debt: generics #24082
Conversation
Note for myself - ejb_group_embedded passed locally twice. |
7a42966
to
08f0e6e
Compare
I will pause now, or slow down ... I will create another parallel PR for the TCK runner ... so I can ensure that this huge PR will not break anything. I committed part by part last week, because I could not find mistake I did ... now I found it. When I replaced Bad was that I could not find it so long, good is that it forced me to analyze the API (broken, leaking) deeply and add some notes how to clean it up so it would make sense. After finishing this PR DOL will not be a scary monster any more :-) - despite it doesn't look like that, the worst part is done. The more code changes will come yet, but they should be much safer. But again, TCK first. |
ab51861
to
b519bec
Compare
- I tried to narrow generics in this packages, which made visible some weird constructions, maybe from some unfinished former refactoring. - sometimes I inverted some conditions to make them more readable too, etc. Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
- unused and in PersistenceDescriptor was using heterogenous collection
Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
- recently we fixed typo in one name, so I created this. - the package; probably we should create some JDK adapter where we would collect JDK enhancements. Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
- fixed copy and paste, equals instead of == for string - reduced noise in logs - don't log twice, better set the exception cause. Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
- see issue eclipse-ee4j#24084 , now I just documented the problem.
- I forgot that I am in a cycle Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
- the interface wasn't an interface, just a container of constants. - all constants were used just on a single place Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
- it did not use any instance's methods nor attributes - writeDescriptor must accept just T as it is designed Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
- amended commit where I originally renamed these methods, but I put the original name back, so I am committing just the cleanup around here. Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
- they dont use anything of the object Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
- didn't use anything from the object Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
…wn object - original code was confusing to read, caused inconsistencies when I improved generics, and led to unwanted usages. Also initiated maps which later weren't used at all. - applied some conventions - removed copypasted javadoc of parent, public static methods first, don't set variables which you don't use any more... Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
- AuthContext renamed to AuthContextImpl - ClientAuthContext and ServerAuthContext have the new common parent, the AuthContext interface. More should be done later in own PR, now I just needed to know classes I am working with. - this module is really ugly, should be radically refactored or removed. - years ago I had to implement login module and realm - be careful when refactoring, protected fields are used by other implementations, so the refactoring will be a breaking change. This one shouldn't. Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
- private fields - bit optimized logging + stacktrace instead of toString - removed unused configSupport Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
…wn object Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
- fixes mistake from recent commit Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
b519bec
to
de13ab5
Compare
Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
- equals is not absolutely correct in the meaning of it's definitions, the desired mechanism is not equality, but if the method overrides another method. The order of processing of methods in the inheritance tree is given, that is why this work. After my refactoring it stopped working, because I tried to correctly implement equals (but it still wasn't symmetric). - it should be probably replaced by different mechanism. Signed-off-by: David Matějček <david.matejcek@omnifish.ee>
I'm yet passing TCK tests which failed yesterday, but it seems this should be the last fix. Refactoring will continue, but it would be good to have some safe point not just in the dev branch ...
REST passed locally twice, probably some issue with the Jenkins machine (cpu overload? 18 parallel docker machines) |
This PR is rather a "feature branch merge" than single fix, because all changes were done with a target to reduce the extreme tech debt of the DOL modules (deployment). Before the merge I would like to run all available tests to ensure that nothing broke, including TCK.