On 07/22/2013 10:44 PM, Tim Starling wrote:
Round-trip bugs, and bugs which cause a given wikitext input to give different HTML in Parsoid compared to MW, should have been detected during automated testing, prior to beta deployment. I don't know why we need users to report them.
500+ edits are being done per hour using Visual Editor [1] (less at this time given that it is way past midnight -- I have seen about 700/hour at times). I did go and click on over 100 links and examined the diffs. I did that twice in the last hour. I am happy to report clean diffs on all edits I checked both times. I did run into a couple of nowiki-insertions which is, strictly speaking not erroneous and based on user input, but is more a usability issue. As far as I know, these are caused by a couple usability kinks (1. wikitext entry in VE; 2. linktrail issues). 2. might already have been fixed in VE. 1. is trickier. But, from what I can tell, once these are fixed, they will greatly reduce the number of nowiki escape wrappers added.
This is not to mean that all is perfect in the Parsoid-VE world. But, as of now, I can say quite confidently that VE and Parsoid coordinate well to achieve an extremely high percentage of clean saves. There have been spectacular roundtrip (and parse) failures of course, and those will be there for a while, but they are the exception, not the norm. And, sometimes they are fixed by fixing templates [2] or broken wikitext in pages.
Barring nowiki issues (addressed separately above), given the state of VE and Parsoid as of today, they will not corrupt enwp pages in any significant way. Parsoid itself is regularly put to automated roundtrip testing on 160K pages [3] wherein we catch regressions and fix egregious failures. At this time, the round-trip failures we are addressing in Parsoid are edge cases. We are currently spending our efforts in improving edit support for VE (image support is the one big gap we currently have). We also have some gaps in the HTML output that we generate for wikitext which we are working to achieve parity with the existing PHP parser and this will affect the editability of those pages via VE.
Yes, we fixed some serious bugs in the first 2 weeks of deployment (rather than before deployment), but that is water under the bridge now. Round-trip errors are not an issue in the decision of the preference as far as I can tell.
The timing of the A/B test also seems to have created a lot of understandable confusion around the issue. Maybe a new A/B test is merited after some time.
Subbu.
1. https://en.wikipedia.org/w/index.php?title=Special:RecentChanges&limit=5... 2. http://en.wikipedia.org/w/index.php?title=Template:Contradict-inline&dif... 3. http://parsoid.wmflabs.org:8001/stats