Hi David, I see the following - which indicates Mac OSX awk does not include gensub ?
Marko ran into the same problem. OSX uses a suuuper old version of awk. Installing the latest GNU command line tools will solve the issue: https://www.topbug.net/blog/2013/04/14/install-and-use-gnu-command-line-tools-in-mac-os-x/ is it possible to start hadoop automatically as part of the process Unfortunately not. Gremlin Server is started automatically if it's not already running, but that only works, because it's part of the project and we know where to find it. But we cannot know the location of your Hadoop installation. The computer unit tests will use the local job runner if Hadoop wasn't started - that works because HDFS is not needed. However, if you want to run the integration tests, you'll have to start Hadoop. Or do I have to install my own copy of hadoop ? > Hadoop 1 or 2 ? Yes, Hadoop 1 (I'm using 1.2.1 <https://archive.apache.org/dist/hadoop/common/hadoop-1.2.1/>). Do I have to shut down the standalone Hadoop when running a full build and test > cases ? No. I usually start Hadoop right after starting my computer and then have it running forever. hadoop namenode -format /usr/local/hadoop/bin/start-all.sh Done. How does the document build process "find" my standalone hadoop ? > Do I have to set up some environment variables that the document build picks > up ? The environment variable CLASSPATH should contain the path to your Hadoop configuration directory. In fact mine contains only this directory: daniel@cube ~ $ echo $CLASSPATH /usr/local/hadoop/conf Cheers, Daniel On Tue, Jun 9, 2015 at 2:19 PM, David Robinson <[email protected]> wrote: > Hi Daniel, > > Thank you for providing the short-cut. I use Mac OS for development. > > When I run the command you provided above, > I see the following - which indicates Mac OSX awk does not include gensub ? > (it does appear to include gsub): > > * source: > ./tinkerpop3/incubator-tinkerpop/docs/src/the-traversal.asciidoc > target: > > ./tinkerpop3/incubator-tinkerpop/target/postprocess-asciidoc/the-traversal.asciidoc > progress: initializing./docs/preprocessor/preprocess-file.sh: line 64: > > ./tinkerpop3/incubator-tinkerpop/target/postprocess-asciidoc/the-traversal.asciidoc: > No such file or directory > awk: calling undefined function gensub > input record number 1, file > ./tinkerpop3/incubator-tinkerpop/docs/src/the-traversal.asciidoc > source line number 26 > awk: calling undefined function gensub > source line number 33 > > > On the full document build process: > > If I do want to build all of the documents, is it possible to start hadoop > automatically as part of the process like the > test cases appear to do ? > > Or do I have to install my own copy of hadoop ? > > Hadoop 1 or 2 ? > > Do I have to shut down the standalone Hadoop when running a full build and > test cases ? > > How does the document build process "find" my standalone hadoop ? > Do I have to set up some environment variables that the document build > picks up ? > > Thank you, > > > > > On Mon, Jun 8, 2015 at 5:32 PM, Daniel Kuppitz <[email protected]> wrote: > > > Hi David, > > > > bin/process-docs.sh will process *all* *.asciidoc files. That includes 1 > or > > 2 files with Hadoop code samples, hence Hadoop must be running. > > To avoid those checks and process only a single file, you can use this > > command: > > > > docs/preprocessor/preprocess-file.sh > > `pwd`/gremlin-console/target/apache-gremlin-console-*-standalone > > `pwd`/docs/src/*xyz*.asciidoc > > > > > > Cheers, > > Daniel > > > > > > On Mon, Jun 8, 2015 at 10:33 PM, David Robinson <[email protected]> > > wrote: > > > > > I am working off of the latest "master" and wanted to add a few changes > > to > > > the "Has Step" documentation in the the-traversal.asciidoc file. > > > > > > Following the README.asciidoc, it says this: > > > * Build AsciiDocs: `bin/process-docs.sh` > > > > > > The process-docs.sh and related sh files around this were just updated > in > > > the past few hours and now require Hadoop to be running to rebuild the > > > ascii docs ? > > > > > > Is it really true that to make a minor typo fix or an update to a > > > non-hadoop specific step, that we now need to have Hadoop running on > our > > > development machine ? Or is there a new way to generate the documents > > that > > > I am not aware of ? > > > > > > This is the message I see when running process-docs.sh: > > > >> Hadoop is not running, be sure to start it before processing the > docs. > > > > > > If the new process really does require Hadoop, may I request that we > not > > do > > > this or find a way to split up the traversal document so that > non-Hadoop > > > related updates don't require Hadoop to be running ? > > > > > > Thank you, > > > > > >
