I have the following SCP task that copies files to a remote server: <target name="deploy" depends="makeDist"> <scp todir="${remote.user}:${remote.password}@${remote.server}:${remote.dir}" trust="true" verbose="true"> <fileset dir="${dist}" /> </scp> </target> When executing this taret using the jsch-0.1.30.jar, the process fails after a number of files get copied. Leaving the fileset alone will result in having the same number of files copied over prior to the hang. Changing the fileset to a new set of files (call it X) will result in a different number of files being copied prior to the hang. However, if repeated the failure will always occur after the same number of files for fileset X. It seems the failure condition is reached after a certain number of bytes are transferred as opposed to a certain number of files. No error messages are displayed and the build does not exit, it just hangs. I've confirmed that this does NOT occur with jsch-0.1.29 or jsch-0.1.28.
Hi, I'm an author of jsch and I'm interested in reproducing this problem. Is it reproducible on Ant 1.7.0? Can you estimate how much of bytes should be transferred to reproduce it, a few Kbytes, Mbytes, Gbytes?
Created attachment 19296 [details] a patch for ant 1.7.0 jsch 0.1.30 has improved the stream handling internally and it must be flushed explicitly.
*** Bug 41465 has been marked as a duplicate of this bug. ***
*** Bug 42271 has been marked as a duplicate of this bug. ***
*** Bug 46107 has been marked as a duplicate of this bug. ***
*** Bug 47856 has been marked as a duplicate of this bug. ***