On Wed, 14 May 2025 17:27:25 GMT, Lance Andersen <lan...@openjdk.org> wrote:
>> UTF-8 messed up the index, we don't want any conversion to happen. >> Warning: Entry META-INFMAANIFEST.MF in local file header is not in central >> directory >> Warning: Entry META-INF/BANIFEST.MF in local file header is not in central >> directory >> Warning: Entry METAMINF/AANIFEST.MF in central directory is not in local >> file header >> Warning: Entry METAMINF/BANIFEST.MF in central directory is not in local >> file header > >> UTF-8 messed up the index, we don't want any conversion to happen. Warning: >> Entry META-INFMAANIFEST.MF in local file header is not in central directory >> Warning: Entry META-INF/BANIFEST.MF in local file header is not in central >> directory Warning: Entry METAMINF/AANIFEST.MF in central directory is not in >> local file header Warning: Entry METAMINF/BANIFEST.MF in central directory >> is not in local file header > > Without looking into this in finer detail I am not sure what the issue is and > is why the other tests typically use the actual offset to modify. Simply put, a byte in the bytearray could be extend to more than one byte with UTF-8. The base we start to looking for the character is recorded as in the bytearray. Thus could lead we start looking at wrong place. I didn't look into exactly how this happens. ------------- PR Review Comment: https://git.openjdk.org/jdk/pull/24430#discussion_r2103056825
