Team,
It seems that multiple code lines and merge will still remain way of
life going forward also. It would be great if we can automate the
process as much as possible. For today's merge into fcs branch, some
diffs/conflicts had to be manually inspected. Following are the details.
1. Conflict due to white space changes:
We had some checkins that introduce dos style CR/LF in the file. Cvs
merge is not very intelligent as far as white spaces go. There is no way
to instruct it to ignore white spaces. These CR/LF introduces a
unneccessary conflict that needs to be manually resolved. Today
following files had this issue
C
src/java/oracle/toplink/essentials/internal/ejb/cmp3/metadata/MetadataConstants.java
C
src/java/oracle/toplink/essentials/internal/ejb/cmp3/metadata/MetadataProject.java
C
src/java/oracle/toplink/essentials/internal/ejb/cmp3/metadata/accessors/ClassAccessor.java
C
src/java/oracle/toplink/essentials/internal/ejb/cmp3/metadata/accessors/CollectionAccessor.java
C
src/java/oracle/toplink/essentials/internal/ejb/cmp3/metadata/queries/MetadataQuery.java
C
src/java/oracle/toplink/essentials/internal/ejb/cmp3/metadata/tables/MetadataSecondaryTable.java
C
src/java/oracle/toplink/essentials/internal/ejb/cmp3/xml/accessors/XMLBasicAccessor.java
C
src/java/oracle/toplink/essentials/internal/ejb/cmp3/xml/accessors/XMLClassAccessor.java
C
src/java/oracle/toplink/essentials/internal/ejb/cmp3/xml/queries/XMLNamedNativeQuery.java
C
src/java/oracle/toplink/essentials/internal/ejb/cmp3/xml/queries/XMLNamedQuery.java
C
src/java/oracle/toplink/essentials/internal/ejb/cmp3/xml/tables/XMLJoinTable.java
C
src/java/oracle/toplink/essentials/internal/ejb/cmp3/xml/tables/XMLSecondaryTable.java
C
src/java/oracle/toplink/essentials/internal/ejb/cmp3/xml/tables/XMLTable.java
Can we be little bit more careful while checking in files. I propose
that we change our editor settings to always use unix style LF instead
of dos style CR/LF. If your editor does not support this feature, please
consider using a utility like dos2unix before checking in.
2. CVS keywords like $ID$, $HEADER$:
Couple of files in the code use these keywords. It triggers a diff
between file in the trunk and branch which needs to be manually verified
to be ok. I think that these keywords in source file do not add much
value. I am planning to remove these keywords from all the files in
trunk on 08/30. Please let me know if any one thinks this is not a good
idea.
Thanks,
Mitesh