Using MadCap's CLI to build a PDF or webdoc?

[sdcinvan]

Hello all,

I am days away from purchasing MadPak. In my proposal process, I was asked if:
a) Is it possible to pull down (from source control - we use git), a MadCap project to make some text edits to the XML
b) Compile the project directly from MadCap's CLI

I am presuming that MadCap uses some form of makefile?

Never mind asking why... ha ha... I am just interested in an accurate answer.

The sort of why (if you are curious) is perhaps it is because I work in a 99% pure software engineering company where I nearly the only person using Windows and a GUI!


Thank you.... and especially thank you if you can provide a link to more information (I search the forums and found some references to much older versions of Flare - I think around v4).

[sdcinvan]

FYI, there are detailed CLI instructions here: http://forums.madcapsoftware.com/viewto ... &sk=t&sd=a

But this post is from 2006. Is this post still relevant for MadCap v9?

[Psider]

I've never used the command line, but perhaps this help topic covers your needs?

http://webhelp.madcapsoftware.com/flare ... d_Line.htm

[sdcinvan]

I've never used the command line, but perhaps this help topic covers your needs?

http://webhelp.madcapsoftware.com/flare ... d_Line.htm

Thank you for that.

So far, so good. I discovered that some (or much) of the work needs to be done inside Flare... as in configuring the Targets.

Next, I built a shell script to basically build any project available (anything pulled from the git repo).

Code: Select all

#! /bin/sh

mkdir -p output
for x in $(find -name *.flprj) ; 
do
  madbuild.exe -project ${x}
  
  dir_path=`dirname ${x}`
  cp -rv ${dir_path}/Output/*/* output
done

It is working quite well... except I just learned that Flare can record build logs. I just need to investigate how to initiate those logs.

[sdcinvan]

I don't know if this is useful to anyone but here is the final change to my MadBuild shell script:

Code: Select all

#! /bin/bash

SCRIPT_DIR=$(dirname $(readlink -e $0))
cd ${SCRIPT_DIR}

mkdir -p output
for x in $(/bin/find . -name *.flprj) ;
do
  madbuild.exe -project ${x} -log true

  dir_path=`dirname ${x}`
  rsync --exclude=Temporary -avzrP ${dir_path}/Output/*/* ${dir_path}/Project/Reports output
done

When I push content to our repo, any changed documents will get automatically built and placed in a public directory. Each project in its named folder and now a MadBuild log folder, as well.

Note: When using MadBuild, I cannot empathize enough the importance of creating a log (using the CLI parameters "-log true"). This is needed for MadBuild to communicate any problems that may occur during the publishing cycle.

[Msquared]

I believe MadBuild returns error codes. I'm sure there is something in the online help about the error codes. I'm not at my PC at present to check.

But be warned if you check them. I had a least one instance where something went wrong in my build (possibly some missing files), and the error was significant enough to make the log, but not significant enough to change the MadBuild return status from success.