Better late than too late: Would it be an option to catch the exception
in JarFile.getManifestFromReference and throw another one with the name
and the caught exception as a cause? Because this is the only code that
calls that particular constructor that I know of. JarVerifier could be
a candidate or SignatureFileVerifier. And JarSigner but JarSigner could
not invoke it like it is now because not visible. I'm not aware of
future plans but if JarFile.getManifestFromReference is the only code
using jar file names in exception messages, we might as well leave
Manifest alone.
Most minor, where is throw new IOException("invalid manifest format(" i
guess there should be a space before the opening bracket at the end of
the string on line 321 to be consistent with the other messages.On Wed, 2019-01-09 at 15:42 -0500, Lance Andersen wrote: > Here is the webrev for the changes: > http://cr.openjdk.java.net/~lancea/8216362/webrev.00/index.html >
