CDR Tickets

Added myself as a watcher.

As part of the WCMS release the HTML had been cleaned up and as part of that some of the IDs that existed in the old HTML document have now changed.

I've fixed the problem on DEV related to the links jumping to locations within the document. I forgot if fragment refs were supposed to work when pointing to other documents as well. With the new HTML output this may not always be achievable as part of the PP report.
The program modified is

• PublishPreview.py

The PublishPreview has been fixed on DEV and is ready for review:

• R13044: PublishPreview.py

I'm getting python errors in Pub Preview on DEV.

A problem occurred in a Python script.

D:\cdr\Log\tmp27lsz_.html contains the description of this error.

A problem occurred in a Python script.

D:\cdr\Log\tmpgg576v.html contains the description of this error.

I'm seeing the same error message right now. It appears to be an error coming from Gatekeeper.
I'll let you know when it's OK to test again.

The problem on Gatekeeper/Percussion has been fixed. Please try again to review this.

Sorry I missed this message. Just tested this on DEV and I'm still getting a Python error:

A problem occurred in a Python script.

D:\cdr\Log\tmpubdoqd.html contains the description of this error.

I see that this is in the Mailer Hotfix, so I tried it on QA too. Got a Python error there too:

A problem occurred in a Python script.

D:\cdr\Log\tmptvgt38.html contains the description of this error.

Verified on QA.

This does not appear to be working on STAGE.

Unfortunately, with the Gatekeeper code on DEV upgraded to the NVCG and the QA server currently not available I'm unable to investigate this in more detail at the moment.
I've looked at the PP output and the version of the PP program but everything appears to be OK on STAGE. I'm afraid this may not be fixed as part of the Mailer hot-fix.

Volker,

Do you expect these changes will be promoted with the upcoming security remediation release or the NVCG-related CDR release? Just trying to get an idea of when this will be fixed so I can let everyone know.

Thanks,
Robin

I can have a quick look on QA to see what's going on (last time QA wasn't available to check it out) but I'm currently unable to give you an estimate on when this may be fixed.

I'm having trouble running PublishPreview on the QA tier at the moment most likely because the DEV and QA tiers are already using the code for NVCG. I'll have to check with Blair or Aarti if a server with the current enrironment is still available for testing.

Volker's going to check whether this got into Subversion (to explain how it didn't get onto PROD with the rest of the Ampere branch).

I have confirmed that revision 13044 of /branches/Ampere/Inetpub/wwwroot/cgi-bin/cdr/PublishPreview.py did actually get promoted to production.

Bob, did you also confirm that the problem was fixed? One of the problems with this issue was that the fix worked on DEV and QA but not anymore on STAGE. When trying to double-check was was going on using QA the PP on that tier was broken - a result of the servers moving to Shady Grove and changing IP addresses.
I think we need to go back to QA and start over with this ticket. Feel free to assign it to me and I'll take a look. My WCMS load for Egyptian Mau isn't too bad.

This does need to be reopened. Summary Frag Refs still aren't working in PP. Thanks for taking a look (when you have time), Volker.

I found and fixed the problem with the publish preview. I'm assuming that the last CDR and WCMS releases crossed each other. The HTML coming from Gatekeeper is now slightly different.

I checked three summaries in PubPreview on Dev and all summary frag refs worked fine. Verified on Dev.

~BKline We'll have to coordinate our changes to PublishPreview.py so that we don't create a conflict. My changes are very minor and easily redone in case a conflict exists.

By the way, this is how I've disabled the GlossaryTerms in the PP report without removing the looks-like-a-link link:

    for gloss in myHtml.cssselect('a.definition'):
        gloss.set('href', '')
        gloss.set('onclick', 'return false')

~volker: I'll be checking in changes to PublishPreview.py some time today (assuming I can ever get it to talk to me on DEV - it's taking forever). I see there's a diff between what's in Curie and what's in the CGI directory:

--- PublishPreview.py   2015-06-17 14:41:27.803847200 -0400
+++ d:\Inetpub\wwwroot\cgi-bin\cdr/PublishPreview.py    2015-06-11 12:12:43.170937300 -0400
@@ -1,6 +1,6 @@
 #----------------------------------------------------------------------
 #
-# $Id: PublishPreview.py 13191 2015-06-17 17:40:01Z bkline $
+# $Id: PublishPreview.py 13175 2015-05-29 18:03:10Z bkline $
 #
 # Transform a CDR document using an XSL/T filter and send it back to
 # the browser.
@@ -287,7 +287,8 @@
     # Modify SummaryRef links
     # These links are identified by the inlinetype attribute
     # ------------------------------------------------------
-    for x in myHtml.xpath('//a[@inlinetype="SummaryRef"]'):
+    #for x in myHtml.xpath('//a[@inlinetype="SummaryRef"]'):
+    for x in myHtml.xpath('//a[@objectid]'):
         # The SummaryRef links don't contain an href attribute
         # because Percussion isn't able to add this to the PP
         # output.  Without the href attribute the text is not

Is that a change that needs to be preserved?

Yes, it's already in my Curie sandbox and will hopefully be committed before I go home, I hope it doesn't create a conflict with your changes.

Would you like me to commit it right away?

I think you've got plenty of time, given how long it's taking me to get PublishPreview to do its thing. :-)

The following program has been saved to subversion:

• R13204: PublishPreview.py

Summary Frag Refs are working well on QA so far. However, clicking "back" after clicking on a summary fragment ref brings us to the top of the summary rather than where we were when we clicked the link. We reported the same problem on Cancer.gov during the initial NVCG testing and I think it was determined that this is a Javascript problem.

Just wanted to note that summary frag refs to a specific section of a different summary are working and the back button will take you back where you began.

Verified on QA.

Looks like the problem with the back button bringing you to the wrong place is being tracked in the WCMS project, specifically OCEPROJECT-3114.

Please verify on PROD and close this issue.

This is verified for the most part. We just need time to do more testing.

Summary Frag Refs seem to be working OK for me. William, do you think we can close this issue?

Yes, we can close this issue. Thanks!