Link checker strategy

Hi,

this is the grant old problem revisited.

Does anybody ever implemented a link checker solution in Teamsite?

What I'm thinking of is:
- some perl (or any other language) script that is called from the TS-menu or workflow
- does a check of all links on all files of an edition, staging or workarea (so unlike the Smart Context Editor link checker which works on a per file basis)
- preferably it would run pre-deploy, since finding broken links on a live site is too late
- returns some report of all broken links

Does anybody have some hints, strategy, perl-modules, ready-made scripts, ready-made tools to do this?

Thanks, Ronald

Find more posts tagged with

Comments

Migrateduser

See this thread, among others. If you do a search for "link checker" you'll see that we discuss this one frequently.

RonaldV

Thanks Rob for your reply.

However the thread you refer (and all others in the devnet, I did do a search before) to is mainly focused on the link checker from the SCE and that is on a per page bases. What I'm looking for is an _edition-wide_ check.

Still, any hint are welcome.

Ronald

Migrateduser

Sorry - should have drilled down more. In any case, if you search the forums you'll find a couple of threads that deal with it.

Long story made short - if whatever link checker you use either exposes an API (like watchfire does, via a perl module -- this is probably the direction we will be going early next year), or is invocable from the command line and returns results you can parse (like the solution Smitty mentions in the above thread), it's fairly easy to either incorporate it into a cgitask or externaltask to check only the files associated with the workflow, or to add a custom menu item that would check whatever you want (select a workarea, an edition, or whatever). I can't be much more specific than that, since we haven't actually done ours yet, but I'm not anticipating any major "gotchas".

Migrateduser

Sure would be nice if IW noticed the number of customers who are wanting a Link Checking tool embedded into TeamSite and actually implement one into the product. It is such a natural enhancement to the existing functionality.

Dave Smith
Sr. Software Engineer
Nike, Inc.
(503) 671-4238
DavidH.Smith@nike.com

stefanmaier

Hi Ronald,
we have implemented a linkchecker with two basic features:
1. check all links in one page/DCR
2. check who links on an object (GIF, PDF, HTML and so on...)

Both basic features can be started through two menu items and from two buttons in a DCT.
The "who links" function is based on an verity search index, which contains all links an all webpages of one website.

Our next step is the implementation of an website-check function, which extracts all broken links of one website.

Regards
Stefan, TS552SP5

RonaldV

Hi Stefan,

thanks for your hints.

Did you use any ready-made code like perl-modules for this or did you start from scratch for this?

Ronald

stefanmaier

Hi Ronald,
we didn't use any ready-made code.

Our components are
one CGI(.ipl) to check all links in one page/DCR (first step is a link extraction algorithm, second step is link checking),
one CGI(.ipl) and a little tomcat-webapplication (1 jsp-file,1 class) to implement the "who links"-function. The webapplication implements the communication to our verity search index,
one perl script included as OD-deploy'n'run-script which extracts all links from our jsp,html-pages and generates some xml-output needed for indexing

Additionally we have a verity-cronjob running every 30 minutes, the cronjob is implemented with basic verity features.

Regards

Stefan

RonaldV

OK, thanks all.

At least you given me some ideas to look at. Or maybe I should wait until Iwov's solution will come with the product. :-)

Thanks, Ronald