Community
Participate
Working Groups
Source based, v_769. The help message reads as follows for its annotations processing part (added line numbers): 1 \ Annotation processing options:\n\ 2 \ These options are meaningful only in a 1.6 environment.\n\ 3 \ -Akey[=value] annotation processors options that are made to\n\ 4 \ annotation processors. key are identifiers separated\n\ 5 \ by ''.''.\n\ 6 \ -processorpath <directories and zip/jar files separated by {0}>\n\ 7 \ specify locations where to find annotation processors\n\ 8 \ If this option is not used, the classpath will be\n\ 9 \ searched for processors.\n\ 10 \ -processor <class1[,class2,...]>\n\ 11 \ Qualified names of the annotation processors to run.\n\ 12 \ This bypasses the default annotation discovery process\n\ 13 \ -proc:only run annotation processors, but do not compile\n\ 14 \ -proc:none perform compilation but do not run annotation\n\ 15 \ processors\n\ 16 \ -s <dir> specify a directory where to put the generated source\n\ 17 \ files\n\ 18 \ -XprintProcessorInfo print information about which annotations and which\n\ 19 \ elements a processor is asked to process\n\ 20 \ -XprintRounds print information about annotation processing rounds.\n\ 21 \ -classNames <className1[,className2,...]>\n\ 22 \ Qualified names of the classes that need\n\ 23 \ to be processed\n\ Remarks and suggestions: - line 3 is incorrect; suggest to add 'available' so as to get 'made available to'; - line 4 'key are' is a singular/plural mismatch; would suggest 'key is made of', or 'key is a suite of'; - line 19 would suggest using the simpler 'a processor processes'; - lines 22-23: I do not quite get what happens exactly here; are other classes compiled or not? how does that play along with other options? - the use of capital letters and dots is not consistent across the help message; my guess would be that we only use capital letters following a dot, and that we use dots only in between sentences; this would lead to line 22 starting with 'qualified', line 9 not ending with a dot, etc.
Created attachment 70176 [details] Proposed fix
Created attachment 70245 [details] updated patch Here's my version of the proposed patch; in addition to correcting typos I tried to clarify the text a bit.
Created attachment 70251 [details] Updated patch + regression tests Walter's patch + corresponding regression tests updated
All tests passed with this patch.
Closing as WONTFIX since this won't go in for 3.3 and the whole print usage should be reorganized for 3.4.
Even if print usage is rewritten, we should not loose the suggestions included in this bug.
Updated batch compiler tests also passed on Linux.
If this doesn't make it for 3.3, I'll release it for 3.3.1. It doesn't affect translations (in fact actual code might make the translation more difficult since some sentences are not very clear) since it is only a rephrasing of the English sentences. The meaning is similar.
Created attachment 73337 [details] Proposed fix + regression test Same patch, but without the change in the print usage for the warnings. I thought it was over the 80 character limit, but it just fits in. So no need for this change. Only the rewording of the annotation processing help is required.
Released for 3.4M1. Updated the regression tests inside BatchCompilerTests. Might want to backport to 3.3.1 in order to clarify the annotation processing options.
Verified for 3.4M1 using build I20070806-1800.