DEBUG HTML ---------- This collection of command procedures create a set of debugging files for 'typical' web pages and will work on VAX or Alpha. The DCL may not completely bug-free or may not even be very useful in your situation, but should be a template for performing validation testing on web sites. debug-make.com calls each of the remaining files to create the following. Note that all filenames are lowercased by the procedure, primarily because my host server is a unix system. Although these files are generated and used on a VMS system, I also need to use then on a case sensitive system too. The pages I use these on are text rich and mainly HTML-2 without frames. The procedures are not particularly tidy and could probably be made more efficient, a run can take a several hours with the huge number of files and references on my site. The web site these have assisted with have over 2000 images and 250 html files with many internal and external links. The command procedures make the maintenance and testing much easier, it couldn't be done manually. All files are referenced by a file called debug.htm which is copied to the directory being tested if it is not there. links to all images - image-links.htm ----IMAGE-LINKS.COM------------------ This provides a clickable alphabetic indexed list of all GIF and JPG files and includes creation dates and size in k. It is also an example of an automatic HTML generator with embedded anchors. load unused GIF or JPG links - unused-gif.htm and unused-jpg.htm ----UNUSED-GIF.COM UNUSED-JPG.COM------------------------------- This is a page of all the pictures which exist in the directory, but none of the HTML files refer to them, so they are never used. (The command procedure that creates this has a boolean to create a clickable list only). This can take a while to execute as it locates each image file then searches all the HTML files for each. link unused any - unused-any.htm ----UNUSED-ANY.COM-------------- This contains a list of all files in the directory which are not referred to by HTML files. This can take a very long time if there are many files, all the html files are searched for every single file in the directory. links not found or located locally or online testing - check-ref.htm ----CHECK-REF.COM--------------------------------------------------- All the HTML files are scanned for HREF and SRC and any which cannot be matched with a local file are placed here. Most likely they are external references, but they may be local typos or missing files. It also checks for mixed case, which can cause problems for unix web servers if it is a typo. SOme other errors are also detected, missing quotes or empty references. Each HTML file is read in turn and the reference is isolated to a file specification which is tested for existence. The code in fact does screen a couple of common references which are not files e.g. mailto. Execution should not take too long as all files are read in only once. links to all HTML files - link-html.htm ----LINK-HTML.COM---------------------- As suggested by the title. Very quick execution as the results of an F$SEARCH are placed into a HTM file. links to all files - link-all.htm ----LINK-ALL.COM----------------- This makes everything 'clickable' with creation dates. This is fairly quick as well. references within HTML files - references.htm ----REFERENCES.COM------------------------- This test cross references all files. Taking each file, it lists what other HTML files also reference it. It has a preset low threshold, and lower than this it displays the total count in red. It is designed to find HTML pages that are not linked to the main body of the pages, or have few links to them overall. This can take a while, it locates each file in turn and searches the remaining files for references. HTML not containing reference - notcontained.htm ------------------------------------------------ This simply examines all HTML files for a reference to, say index.html. It allows you to check that all files link back to the homepage for example, or are referenced by an indexing file. This is fairly quick as each HTML file is searched once in turn. HTML not referenced by file - notreferenced.htm ----------------------------------------------- This is the opposite of the previous test, it check to see which files are not linked by, say, the main index page. This can take a little longer as it searches all files once for each HTML file. While running these procedures file access conflict errors are normal. Pressing control-y at any time is safe and results in only a partial htm file. How to use ---------- Either use directory from the freeware disk, or copy all the .COM files and the debug.htm file to a directory of your choice. Simply @ or submit with parameters the debug-make file, it will find all the others itself. You may individually call each file, or submit it to batch with the correct parameters, or edit the debug-make.com to run the components you wish. That's it, I hope you find something of use here. If you develop something interesting you think I should keep with the distribution then send the updates to me. I'll maintain a current distribution on http://www.python.demon.co.uk and I would appreciate you dropping me a note if you find this useful. Nic Clews nic@python.demon.co.uk