User:Jholsenback/WikiDocGen
The Setup
In the directory /a/home/jholsenback/testbed
this is the expected directory structure:
documentation
mac
images
unx
images
win
images
You should be in the above mentioned directory to run the scripts, however, the batch-files wander a bit but always end up back home.
The Files
These are the files that make up the wikidocgen process:
batchMacFiles
- a bash script that processes the Mac OS documentation set
batchWinFiles
- a bash script that processes the Windows documentation set
batchUnxFiles
- a bash script that processes the Unix documentation set
getWikiPages.php
- gets the files from the POV-Wiki and processes them
mkContentsPages.php
- builds the various table of contents files and an index.html file for a given doc set
mkWikiTOC.php
- rebuilds this listing if any of these contents pages below gets changed.
See additional notes at bottom for more about this! mkImagePackage.php
- builds the images directory structure and copies files from the POV-Wiki
getStyleSheet.php
- gets the povray37.css style sheet and copies it to a given doc set directory
common.php
- frequently used functions associated with the wikidocgen process
utilities.php
- frequently used functions associated with the wikidocgen process, an attempt to organize
documentation/WikiImages
- a symbolic link to the POV-Wiki images directory
documentation/favicon.ico
- copied to a given doc set directory
Besides the content files on the POV-Wiki these files are also part of the process:
Running The Process
ssh access to lib.povray.org is required, so is this directory structure. All the shell and PHP scripts that make up the process are available revision control under povray > tools > wiki-docgen. Once you're logged and all setup has been done, do the following:
Note: This example deals with generating the Unix/Linux doc set, but the other platforms (MacOs and Windows) are the similar, they are just processed into different directories.
cd testbed
./batchUnxFiles
cd documentation/unx/
find . -type f -exec chmod 444 '{}' \;
tar -cvvf 24Oct2011unx.tar Arrow*.* favicon.ico images/* povray37.css *.html
gzip 24Oct2011unx.tar
find . -type f -exec chmod 644 '{}' \;
chmod 444 *.tar.gz
cd ..
vi docsets.html
Typically it's okay to run each platforms batch scripts one right after another, then the individual platform directories get bundled into a tar file. Lastly they get compressed using gzip.
Note: If you've previously run the process make sure that the file permissions are correctly set before running the process again. See step 7 above.
Notes
The most important part of this process is that it is driven off these files:
- Windows Table of Contents
- Mac OS Table of Contents
- Unix Table of Contents
- Tutorial Table of Contents
- Reference Table of Contents
The GatherTOCfiles
and BuildDocMap
functions reads and processes those files. They are located in utilities
and common
respectively, they depend on the class DocumentMap
definition located in common
. During batch processing if an exception message is seen, most likely something isn't correct with an associated table of contents entry and how it appears in the content body. The title portion of the toc entry and the heading entry in the content body MUST be and exact match. Batch processing will continue and the offender will appear in the content body as wiki markup.
Hint: look for the wiki tag Documentation:
The content on the Wiki is organized such that a given section can span more than one file, the script getWikiPages.php
effectively concatenates the files together, so look near the bottom of that script to find a place to tap into the stream as it were!
The file that mkWikiTOC.php
produces needs to be copied into the table of contents file. Since this file spans several sections, you need to edit the entire page (top most edit button). Look for these comments embedding in the file delimiting where the changes should go.
<!--BEGIN CHANGES BETWEEN HERE---> <!--END CHANGES BETWEEN HERE--->