Hi Erik, I tried to import your patch, and now the build fails with the following error:
make[3]: *** No rule to make target `/Users/danielfuchs/workspaces/jdk/jdk9-dev-java-time/jdk/make/data/charsetmapping/stdcs-macosx', needed by `/Users/danielfuchs/workspaces/jdk/jdk9-dev-java-time/build/macosx-x86_64-normal-server-release/support/gensrc/java.base/sun/nio/cs/_the.charsetmapping-stdcs'. Stop.
So it looks that something is still wrong :-( best regards, -- daniel On 17/02/15 15:56, Erik Joelsson wrote:
Hello, Please review this fix for the incremental build behavior of the gensrc build step. After JDK-8073152 the charsets would unconditionally get regenerated each time make was invoked. The actual culprit was a stray backslash which prevented the 'touch' on the next line from executing. I took the liberty of cleaning up some more, adding to the prerequisites lists of the rules in question and fixing whitespace according to our guideline [1]. The whitespace changes aren't visible in the webrev. I also quieted down the gensrc tool for charsets by adding the LOG_INFO macro. This will hide the output for the default warn level. To see it, run make with "LOG=info". For this to work I had to modify the tool to print its unconditional output to stdout instead of stderr. We want to keep stderr open to the log for any actual failure messages. Bug: https://bugs.openjdk.java.net/browse/JDK-8073328 Webrev: http://cr.openjdk.java.net/~erikj/8073328/webrev.jdk.01/ /Erik [1] http://openjdk.java.net/groups/build/doc/code-conventions.html
