Re: PgJDBC: code reformat - Mailing list pgsql-jdbc
From | Markus KARG |
---|---|
Subject | Re: PgJDBC: code reformat |
Date | |
Msg-id | 064101d140b9$8cf3f260$a6dbd720$@eu Whole thread Raw |
In response to | Re: PgJDBC: code reformat (Vladimir Sitnikov <sitnikov.vladimir@gmail.com>) |
Responses |
Re: PgJDBC: code reformat
|
List | pgsql-jdbc |
I actually did NOT vote for "braces on new line". In fact I hate that style and always have opening braces on the SAME line.What I actually said was that if we decide for a near-Sun-Style it makes more sense to go with pure-sun-Style instead,as that is built into IDEs already -- even if it produces a style a do not like much. -Markus -----Original Message----- From: Vladimir Sitnikov [mailto:sitnikov.vladimir@gmail.com] Sent: Sonntag, 27. Dezember 2015 16:02 To: Gavin Flower; Dave Cramer; Markus KARG Cc: List Subject: Re: [JDBC] PgJDBC: code reformat I did try reformatting with "opening braces on new line". It turned out it is incompatible with checkstyle: https://travis-ci.org/pgjdbc/pgjdbc/jobs/98976190 There are 17 "... have incorrect indentation level" failures. Most of which are due to enum initialization in https://github.com/pgjdbc/pgjdbc/blob/197175039068446a15c30d2b5e949f1eae08515d/pgjdbc/src/main/java/org/postgresql/hostchooser/HostRequirement.java#L16-L29 I have just two requirements: 1) IDEA's autoformat should be able to produce the desired style (minimal modifications are bearable) 2) The style should be enforced, in other words, checkstyle (or whatever tools is used) should be able to catch most of thestyling issues. As Garvin/Dave/Markus vote for "braces on new line", would you please suggest how to fix checkstyle errors? For instance: /home/travis/build/pgjdbc/pgjdbc/pgjdbc/src/main/java/org/postgresql/jdbc/PgResultSet.java:2439: error: 'object def lcurly' have incorrect indentation level 6, expected level should be one of the following: 2, 4. https://github.com/pgjdbc/pgjdbc/blob/197175039068446a15c30d2b5e949f1eae08515d/pgjdbc/src/main/java/org/postgresql/jdbc/PgResultSet.java#L2439 PS. I do not care where the braces are. However, this makes me cry: https://github.com/pgjdbc/pgjdbc/blob/197175039068446a15c30d2b5e949f1eae08515d/pgjdbc/src/main/java/org/postgresql/Driver.java#L353-L373 There are probably better ways of writing that, but those "new lines for opening brackets" make it much worse. PPS. >The other style looks cluttered, and is harder to see Gavin, I feel your pain, however I still think you are trying to solve "badly written code" with "reformat code" approach. That just does not work. Compare this one: https://github.com/pgjdbc/pgjdbc/blob/197175039068446a15c30d2b5e949f1eae08515d/pgjdbc/src/main/java/org/postgresql/jdbc/PgResultSet.java#L2462-L2512 with this one: https://github.com/pgjdbc/pgjdbc/blob/197175039068446a15c30d2b5e949f1eae08515d/pgjdbc/src/main/java/org/postgresql/core/v3/QueryExecutorImpl.java#L264-L311 Do you see how QueryExecutorImpl.java#L264-L311 is much readable? PgResultSet#getFastLong is very hard to follow no matter which way you format the braces. I believe, "readability" comes from proper segmentation (code blocks vs methods) and proper variable naming. Vladimir
pgsql-jdbc by date: