Bug found and fix provided by Earl Hood

https://bz.apache.org/bugzilla/show_bug.cgi?id=58054

Too bad this hasn't been found two weeks before.

we haven't created betas in years

- clarify what "availibility" means
- fix </td> on @postProcessGeneratedJavadocs

https://bz.apache.org/bugzilla/show_bug.cgi?id=57834

first part of https://bz.apache.org/bugzilla/show_bug.cgi?id=57834 -
allbutlast to follow

https://bz.apache.org/bugzilla/show_bug.cgi?id=43271

Apparently the remote umask still comes into play and the result won't
be exactly what we intended it to be.

Based on patch by Holly Cummins

Port of fixes for https://issues.apache.org/jira/browse/COMPRESS-314 and
https://issues.apache.org/jira/browse/COMPRESS-315

since we read resources from it, the JVM won't close it and subsequent tests
can not remove ${output} - making them fail on Windows.

Submitted by Petr Kureš @pkures
closes #8 and closes #9

1024 byte buffer size is insufficient for transfering large files over faster (> 100Mbit) networks. We are transfering files tens or hundreds MB in size and transfer speed was 1MB/s. After increasing the buffer size too 100KB we are getting transfer speeds up to 60MB/s. Please consider increasing the BUFFER_SIZE in the next Ant release. Thanks, Peter

1024 byte buffer size is insufficient for transfering large files over faster (> 100Mbit) networks. We are transfering files tens or hundreds MB in size and transfer speed was 1MB/s. After increasing the buffer size too 100KB we are getting transfer speeds up to 60MB/s. Please consider increasing the BUFFER_SIZE in the next Ant release. Thanks, Peter

This is a workaround for Bugzilla Issue 57822
https://bz.apache.org/bugzilla/show_bug.cgi?id=57822

Submitted by Roman Savko
PR: 57789