This Bugzilla instance is a read-only archive of historic NetBeans bug reports. To report a bug in NetBeans please follow the project's instructions for reporting issues.
Few months ago it has been discussed and agreed that the obsoleted content of NetBeans javadoc[1] needs to be clearly marked as obsoleted[2]. After a discussion it has been agreed that replacing old pages with empty ones redirecting anyone who accidentally still has reference to the old page to new http://www.netbeans.org/download/dev/javadoc/ page[3]. Jack promissed to do it[4] if list of obsoleted files is sent to him. The list is here: http://platform.netbeans.org/servlets/ReadMsg?list=dev&msgNo=53 But nothing seems to be removed. There still seem to be the old pages[5]. As we finally fixed all broken links javadocs (issue 62058) it is silly to not fix the rest. I am opening this bug to track progress of that fixing. I believe that is has to be fixed before release and thus I am making it p2. [1] http://www.netbeans.org/download/dev/javadoc/ [2] http://platform.netbeans.org/servlets/ReadMsg?list=dev&msgNo=33 [3] http://platform.netbeans.org/servlets/ReadMsg?list=dev&msgNo=41 [4] http://platform.netbeans.org/servlets/ReadMsg?list=dev&msgNo=50 [5] http://www.netbeans.org/download/dev/javadoc/OpenAPIs/
This is actually already filed as issue 64739, but since there is more info here I'll mark that one a dup of this one.
*** Issue 64739 has been marked as a duplicate of this issue. ***
Yet new case of people accessing our old, obsolete docs: http://openide.netbeans.org/servlets/ReadMsg?list=dev&msgNo=18548 Apparently google is indexing them. We really need to get rid of the obsolete content. Opps, I've just realized that the issue is wrongly categorized. It is not problem of ide/www but apisupport's documentation. So I am moving it here. Still I am afraid there will be a need for a help from WWW or RE team as development nor docs can access the upload area.
Created attachment 27864 [details] Ruda told me that he is able to do ls but not find, so here is a little script to simulate recursive ls with plain ls, I hope it will be helpful
Sorry for the delay. This is underway, though the process is quite slow so I expect it to run for some hours (?) yet. You can see first results at eg http://www.netbeans.org/download/dev/javadoc/AWTAPI/org/openide/awt/ToolbarToggleButton.html Question : What to do with the .zip files, eg /dev/javadoc/AWTAPI.zip ? I can't replace them with a readable html file, the browser will try to download it like a zipfile. I think safest is to just remove them ? Ppl trying to grab them will just get 404s, but ... Also, there are other non-html files in these dirs, (eg css, gifs, etc) - I assume it is OK to simply remove them ?
Yes, all non-*.html files can just be removed, I think. Note: I'm not so keen on keeping the old file names around with redirects. (1) No one should have been linking to them for a long time anyway, only old mails (and we would rather have people use the Wiki FAQ than have to rely on mailing list search). (2) It is not hard to find a given new page using the master index. (3) We do not want search engines indexing the obsolete pages at all. But Yarda requested the redirects so I will leave it to him to comment.
The redirect on ToolbarToggleButton.html is exactly what I was looking for. I agree that fixing .html files is enough, the rest can be deleted. "But Yarda requested the redirects so I will leave it to him to comment." - what shall I comment? There are tons of links to our old pages and I can only repeat what has been said at [3]. Just with the exception that our new javadoc is free of old links. Still getting 404 is imho pretty amateurish. [3] http://platform.netbeans.org/servlets/ReadMsg?list=dev&msgNo=41
Created attachment 27949 [details] Btw. an awk script that takes output of ls -1RF and generates list of full paths to listed files, could be useful
Created attachment 27950 [details] Ruda sent me ls -1RF and here is the list of all html files I generated from it
Created attachment 27951 [details] Ruda sent me ls -1RF and here is the list of all html files I generated from it
So how should we prevent Google from indexing the old content?
This is mostly done. There are some directories with spaces in the names - there's an issue with us being able to address such dirs, see issue 43479. This means I am not able to remove the few zips with spaces in their filenames, nor the non-html files in those dirs. I've asked Collab to do that in issue 70588. I was able to replace the html in those problem dirs OK, thanks to Rudo for the tip. Removal of zips and non-html files in other dirs is still running. All bad html files should now be replaced by the redirect. I will attach a list of redirected files, and of removed files.
Created attachment 27953 [details] List of obsolete javadoc html files replaced by static "moved" page
Created attachment 27954 [details] List of removed obsolete javadoc supporting files
Excelent, the pages I tried seem to be redirected! I would mark this as fixed, if you do not mind. Re. "So how should we prevent Google from indexing the old content?" - Is it necessary to prevent google from indexing page like this: http://www.netbeans.org/download/dev/javadoc/OpenAPIs/org/openide/doc-files/classpath.html#overlaps No. If someones googles that page, the redirect is clear and in a few weeks the old cached content is going to be removed from google's cache anyway.
It looks fine.