RH11, Windows 7 Enterprise
Problem: Deployed webhelp toc has duplicate and deleted pages.
Background: Our development team pushes Help from our Atlassian bitbucket server to QA test, user test acceptance (uat), and production release, to add it to our product. Same Help file and product for over a year. Found after production. I see it was also in uat, but we missed it. We're in the next QA test cycle now, and the issue isn't seen, but I need to make sure it's fixed.
Image: Contents pages under Students Module were duplicated under Groups Module, and the Groups Module pages were gone.
Where issue was/is not seen:
Here's some things that occur to me, but I don't know how to check them:
Thank you for your thoughts!
Some of what you are talking about is very confusing. You seem to be co-mingling possibly internal terminology that makes sense only to people in your environment with basic RoboHelp stuff.
From what you have shared, it seems that in some cases you are seeing this odd duplication. What I'm unsure about is if this problem has at this point totally disappeared and you no longer see it, but you are concerned there is something that could cause it to return, or if it's still occurring for certain users.
Some developers are new. Is there an error they could have made in the deployment?
While that's certainly possible, there is no way of knowing for certain without keeping a close eye on their output to see if this problem manifests after one or more of them generate output.
You also ask:
If bitbucket went offline when webhelp files were being picked up, could a file have been corrupted, such as the whtoc.js file?
This statement is what I was referring to at the beginning. What in the world is "bitbucket"? Is it a source control system? Document management system of some type?
As for files somehow getting "corrupted", I suppose that's always a possibility. The only way to know for certain is to compare known good versions with the suspect versions.
Another somewhat confusing aspect is that you are saying that "pages" were duplicated or missing but all I see in the image are books. Are you saying that there are no pages inside the books?
Sorry Rick. I was trying to be specific for people doing the same thing and included the bitbucket source control system in case they've seen troubles there. However, I should have been more general and the forum won't let me edit the posting.
The duplicated menu nodes (seen only as books here) are duplicated down to the page level. Here's what one developer reported last night:
So we're still trying to figure it out. Yes, we're concerned that it may happen again. Thanks for your thoughts, I appreciate your experience with Robohelp.
BTW - eventually browser security issues are going to sabotage your RH11-produced WebHelp - at that point you're going to need a more up-to-date version of RH to get patches issued.
So WRT the project itself, can you please share the following screen captures?
1. The Project Manager pod with the Table Of Contents area expanded. Similar to below.
2. The Output SSL settings with Content Categories expanded and focus on the category you are using. What does the drop-down for the Table of Contents show?
One other thought I had was that it's possible that if you simply overwrite files on your server, maybe something is going screwy there. So it could be worthwhile to just delete the output from the server and begin with a fresh reload of the output.
Another thought is browser caching.
The RH multiscreen HTML5 table of contents seems super "sticky" recently. Whenever our ToC changes, a weird combo of old and new structure, and various error message display when accessing some topics. I've been advising our users when it happens to clear their browser cache, close all browser windows and wait 5min before trying again.
It couldn't hurt to see if this works for webhelp as well.
We deployed again yesterday to try to fix the duplicate menu nodes. These are the results from the two deployments:
Very glad it cleared up. However, what could have happened in Last week's deployment? I only cleared my cache after this week's deployment, so it wasn't in my RH Webhelp generation. Our devs are going to spend time on it, and we'd welcome your input.
Amber's idea of caching is the only plausible explanation that comes to my own mind. As you say you only have a single TOC, it can only mean that at some point things are getting scrambled.
Perhaps it will help to sort of fill in my own understanding of how things work here. During that WebHelp generation process, all the navigational data (TOC, Index, Search, Glossary, etc.) is broken into small chunks with sequential numbers. At the time WebHelp is opened, the scripting involved loads these smaller chunks as needed. So I could see where if some chunks weren't updated and perhaps were still hanging out in among the cached content, something like that might happen.
Hopefully it won't happen again. But if it does, seems clearing the cache of the affected machines is a good place to start!
Curious about the five minute wait. Any particular reason for that? My thought is that if the cache is cleared, it's cleared.
Captiv8r All I can think is there is some weird combo of local and server caching going on. I find that sometimes clearing the cache and re-opening immediately still shows the problem, whereas clearing and waiting for a period fixes the issue.
(I've added code to prevent caching of HTML files, but the table of contents is all in an xml (or js in xml format? I forget which) and I'm not up with how to modify that. Possibly the file could be manually excluded from caching on the server, but that is a whole nother can of worms.)
Your caching suggestions were right. Our developer advises the person who deploys the webhelp to make sure the server is putting "Cache Control: no-cache" on the files it serves. Thanks everyone!
Thanks, Rick. I only have one toc, which is the default. Overwriting on the server is an interesting thought, and I'll bring it up.