[
https://issues.apache.org/jira/browse/HADOOP-12479?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14958977#comment-14958977
]
Chris Nauroth commented on HADOOP-12479:
----------------------------------------
bq. That change to exec() doesn't break anything does it?
No, it's fine. The only other caller is the {{VersionInfoMojo}}. I chose to
do the {{Exec}} change in a way that would be backward-compatible for
{{VersionInfoMojo}}. It's a new overload of the {{run}} method that captures
stderr in a separate in/out variable. Only {{ProtocMojo}} calls the new
overload. {{VersionInfoMojo}} still calls the old one.
An alternative would have been just to dump both stdout and stderr into the
same in/out variable used by the existing {{run}} method. That could have been
dangerous, because {{VersionInfoMojo}} has specific expectations about the
output, and dumping stderr on top of that could have harmed that logic.
Thanks for the review. I'll commit later today.
> ProtocMojo does not log the reason for a protoc compilation failure.
> --------------------------------------------------------------------
>
> Key: HADOOP-12479
> URL: https://issues.apache.org/jira/browse/HADOOP-12479
> Project: Hadoop Common
> Issue Type: Bug
> Components: build
> Reporter: Chris Nauroth
> Assignee: Chris Nauroth
> Priority: Minor
> Attachments: HADOOP-12479.001.patch
>
>
> If protoc fails with a compilation error in the proto files, our Maven plugin
> won't print the details. The only way to figure it out is to repeat running
> the {{protoc}} command manually from outside the Hadoop build. This is
> because our {{ProtocMojo}} only captures stdout from the {{protoc}} command,
> and compilation errors get written to {{stderr}}.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
