leerho commented on code in PR #535:
URL: https://github.com/apache/datasketches-cpp/pull/535#discussion_r4188315228
##########
.gitignore:
##########
@@ -42,4 +42,7 @@ _*/
!__init__.py
docs
-java
+# Cross-language test data: sketch snapshots and the datasketches-java
checkout used to generate them
+serialization_test_data/
+datasketches-java/
+*.sk
Review Comment:
The only files in C++ and Java that are **tracked** files are roughly 15
historical files that can no longer be generated from modern code. We keep
these around to just check that we can still read them. These files are
treated separately from the .sk files we generate and test via TCK.
One would think that Java and C++ would have the same historical files; they
overlap but are not the same. Java has some that C++ doesn't and visa-versa.
These historical files don't even follow our more recent naming convention. I
will present this detail as another issue.
Since these files are already tracked, having a "\*.sk" in .gitignore will
not affect them.
It should be a very rare exception that a contributor would need to add to
this 'historical' set of .sk files. And, if that were required, such a file
can be added using 'git add --force'.
Since all the .sk files are stored in subdirectories, not the root, the
entry in .gitignore must be the global '\*.sk' or '**/\*.sk' version. We
don't want any of those files to be inadvertently ending up in origin.
It might appear that having 'serialization_test_data/' in gitignore would be
redundant, but the code that generates and tests these files is spread all over
the place and not centralized. We have found instances where some sketches are
not following the convention of writing these files to designated places and
ending up in the wrong directories. The 'global' '\*.sk' in gitignore will
catch these.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
