Call for Testing: Site Export

Hey there! I’m Jordan, a developer working on the new Site Exporter for WordPress.com and would love to hear your feedback.

What is it?

This adds a new tab on your site’s Settings page, for exporting all post content. The exported content can later be imported into another WordPress site – either on WordPress.com or your own hosted site.

WordPress.com Site Export

What to test

  • Can you export all content from your site?
  • Can you export a specific selection of content from your site?
  • Can you download the export file when complete?
  • Do you receive an email after the export completes?
  • Are you able to take the exported content and import it into a new site?
  • Is there anything you found confusing during this process?

To get started, head to https://horizon.wordpress.com/. Click My Sites, then choose Settings from the sidebar, and finally click the Export tab.

We’d particularly like to hear of any issues you might find when exporting very large sites with many posts or pages.

Please note that the export file won’t contain theme settings. This means that a new site using imported data will not necessarily look the same as the original site.

We’ll continue gathering feedback until June 21st, 2016.

Published by

Jordan West

JavaScript Wrangler @ Automattic

28 thoughts on “Call for Testing: Site Export”

      1. Hello Jordan West,

        Oddly I got no email alert or WordPress.com notification that you had replied to my comment, the only reason that I realized that you replied is that I happened to manually look at the Comments I Made section in the WP Admin area because I was wondering why no one had replied yet, and that is when I saw that someone had replied to my comment.

        I am not sure what is going on there, but that should be fixed or people will not know that someone has replied to them and/or commented on the post that they commented on.

        Now about the missing Export option at My Site | Settings, here is the address that I am using:

        https://wordpress.com/settings/general/johnjronline.wordpress.com

        I am using the latest 64-Bit version of Google Chrome in Incognito Mode with these extensions (McAfee Webadvisor (I am temporarily testing this), Web Of Trust, Adguard (I am testing this in place of Adblock Plus), and Video DownloadHelper) on the latest version of Windows 10 64-Bit.

        If a special link was required to test this, which I am not sure, I think that the special link should always be in each Horizon Feedback post to avoid confusion or that the exact steps should be given clearly in detail.

        Thank you,
        -John Jr

        Liked by 1 person

        1. Hi John, thanks for the detailed reply!

          As this feature hasn’t launched yet, it’s only available for testing on https://horizon.wordpress.com. Apologies for the confusion; I’ve updated the post to add that link.

          As for the missing notification, I’m not sure what happened there. I’ll keep an eye on it, please let me know if it happens again with this comment.

          Liked by 2 people

      2. Hello Jordan West,

        You are welcome, and thank you for adding the link to the post and I hope that this is always done in the future so that people will know where to go to test things.

        I still did not get an email alert or WordPress.com notification with your new reply either.

        I got an email alert and notification of your like, but not for your comment.

        I have gotten an email alert of each comment that Jim Grey has made on this post, but nothing when you comment and no notification when you reply to my comments.

        Maybe there is a bug when the post author comments and/or when staff members comment?

        Thank you,
        -John Jr

        Like

        1. John, I wonder if there’s any chance you hadn’t ticked the “Notify me of new comments via email” box until a little later in the discussion. I’ve checked our email logs and can find 8 comment notification emails to you for this post, which is somewhat short of the 12 comments on here besides yours but also equals the total 16 posts minus both your comments and Jordan’s comments, which could lend credence to your theory that it’s a bug with post author or staff member comments. I suppose we can test that further by seeing if you get a comment notification for this one. Can you let me know?

          Liked by 1 person

        2. Hello Daryl,

          I assume that I enabled it because that option is not there when I comment, so either I enabled it like I usually do, or that option was missing from this post; but I assume that it is enabled because I get email alerts and WordPress.com alerts for everyone except the author of this post so far.

          I did get an email alert and WordPress.com notification for your comments today, so if you are a WordPress.com staff member like I assume, that means that maybe the bug happens when the author of the post comments.

          -John Jr

          Like

  1. Hi! I exported my main site, Down the Road (blog.jimgrey.net nee jimgrey.wordpress.com). I’ve been writing in it for almost 10 years, so I figured this would be a good test of a blog with a long history.

    I got the download link; it downloaded a zip that contained a folder that contained 3 .xml files. When I went to import these files, it gave me this screen showing three authors for my blog:

    http://screencast.com/t/IiTBH3fk

    Thing is, I’m the only author on my blog. Now, when I look at the xml I see that Jeremy Duvall has an @automattic.com email, so maybe he is someone who gave me support at some point. Velda’s email address is velda@supernovia.com so I don’t know who she is or where she comes from. Perhaps she’s another Automattician who just used that domain while working on a problem with my blog.

    It is unexpected to see multiple authors when I’ve never authorized additional authors on this blog. If Automatticians providing support in certain ways end up showing up as authors that’s fine, but it will cause users to go “hunh?” and you might need a way of explaining that to them.

    Liked by 2 people

    1. Hi Jim! Thanks very much for helping us out, your 10 year old blog is really a great case to test.

      I agree that it’s unexpected to see other authors. It does look like Velda and Jeremy helped you out at one point so I’ll do some further investigation.

      Like

      1. I’ve reported this a while ago in the forums. 🙂

        You guys are right, if one of the Automattic employees have accessed one of your posts and there was a save (or even an autosave) then it will show up in the revision history and then may wind up in your export.

        Liked by 2 people

  2. I kicked off the import at 8 am EDT and at 3:15 pm EDT came back to the computer and saw that it was still importing. I stopped the import and went to the site to which I was importing (jimgreytest.wordpress.com) and found that nothing had imported. I thought that stuff would go live on that site as it was being imported. So I wonder if the import was doing anything at all.

    Liked by 1 person

    1. Sometimes super big blogs, or blogs that have some strange characters in them, don’t import cleanly. From looking at the export file for your big blog, I see some characters in the export that shouldn’t be there, which will cause the import to fail. It’s actually a known case, and we’ve got an issue open to handle removal of these kinds of characters better on export. It’s not specific to the new Calypso export experience. It’s also not a terribly common case, so it’s not one we consider super high priority, though we would like to (and do plan to) fix it. Thanks for the feedback and the test case!

      Like

  3. Last night I exported my other, newer blog that has far less content (softwaresaltmines.wordpress.com), and imported it into my test blog. It worked, and quickly. So I cleared all that content out of the test blog, reexported my main blog (jimgrey.wordpress.com), and tried importing again. It processed overnight but never actually started importing content. So it looks like there’s an upper limit of size/scope of a blog beyond which the importer just chokes.

    Liked by 2 people

  4. I used the export all option for my blog and it said exporting for maybe two hours without it ever exporting, and so I refreshed and I tried it again and it worked the second time so I was able to download it and it was sent to my email address.

    I was using the same browser and operating system setup that I mentioned earlier.

    I do not have a test blog to import it to so I was not able to test that, and I did not test exporting only certain things.

    -John Jr

    Like

    1. This is something we’ve seen intermittently before and will be digging into a bit more, to see if we can reproduce reliably. Can you tell me whether you had already tested a few exports when this happened, or was this on your first try?

      Liked by 1 person

      1. Hello Daryl,

        It was my first try with the new beta export feature, but I never have this problem with the current WP Admin export option that I use every day.

        *Off-Topic*

        I found an old export file for my blog that was possibly from before I accidentally deleted all the comments on my blog, I am not sure though, is there a way that you and/or I can see if that export file contains all of my comments before the year 2013?

        If it does contain those comments, I would like to import only the comments if that is even possible from a full export file, but I have never done this before so I would need help.

        -John Jr

        Like

      2. John Jr, I tried a fair bit today to reproduce this export failure error and had no luck. I thought it might have to do with network interruptions, but testing scenarios around that resulted in a big red error box rather than just a stuck export. If you happen to be able to reproduce again, I’d be super interested in learning what if anything other than a spinner appears on the screen. I’d also be keen to know if anything appears in the browser development console (instructions for checking here).

        Liked by 1 person

        1. Hello Daryl L. L. Houston,

          I just tried it again and after a brief wait it worked and I was able to download the export file, and I was able to successfully download the export file that was sent to my email address; but there were 16 errors in the Google Chrome development console:

          Failed to load resource: net::ERR_BLOCKED_BY_CLIENT
          https://www.google-analytics.com/analytics.js Failed to load resource: net::ERR_BLOCKED_BY_CLIENT
          https://pixel.wp.com/g.gif?v=wpcom-no-pv&x_newdash_pageviews=route&t=0.9946070367041397 Failed to load resource: net::ERR_BLOCKED_BY_CLIENT
          https://pixel.wp.com/g.gif?v=wpcom-no-pv&x_reader_views=following_load&t=0.20930302811500434 Failed to load resource: net::ERR_BLOCKED_BY_CLIENT
          https://nrpc.olark.com/nrpc/c?c=create&s=3146-815-10-9343&v=l1IJiVW99nBUzgS…ss.com%2F2016%2F06%2F14%2Fcall-for-testing-site-export%2Fcomment-page-1%2F Failed to load resource: net::ERR_BLOCKED_BY_CLIENT
          https://pixel.wp.com/boom.gif?v=0.9&u=https://public-api.wordpress.com/pinghub&json={%22beacons%22:[%22pinghub.restproxy.ws_supported.chrome:1|c%22]} Failed to load resource: net::ERR_BLOCKED_BY_CLIENT
          https://pixel.wp.com/boom.gif?v=0.9&u=https://public-api.wordpress.com/pinghub&json={%22beacons%22:[%22pinghub.restproxy.cookie.chrome:1|c%22]} Failed to load resource: net::ERR_BLOCKED_BY_CLIENT
          https://pixel.wp.com/boom.gif?v=0.9&u=https://public-api.wordpress.com/pinghub&json={%22beacons%22:[%22pinghub.conn_open_dc.dfw.chrome:647|ms%22]} Failed to load resource: net::ERR_BLOCKED_BY_CLIENT
          https://log.olark.com/jslog/log.png?version=framestore&location=https%3A%2F…43&level=warn&timestamp=1482534283761&properties=%7B%7D&recent_logs=%5B%5D Failed to load resource: net::ERR_BLOCKED_BY_CLIENT
          https://nrpc.olark.com/nrpc/c?j=olark-11482534284183&&c=create&q=1813.84182…648c703eb70132ff220&version=api-1.2.1&pretty=true&_rnd=0.06820111531367345 Failed to load resource: net::ERR_BLOCKED_BY_CLIENT
          https://pixel.wp.com/g.gif?v=wpcom-no-pv&x_newdash_pageviews=route&t=0.8753502305486915 Failed to load resource: net::ERR_BLOCKED_BY_CLIENT
          https://nrpc.olark.com/nrpc/c?j=olark-11482534284183r2&&c=create&q=1813.841…9648c703eb70132ff220&version=api-1.2.1&pretty=true&_rnd=0.9734997385474373 Failed to load resource: net::ERR_BLOCKED_BY_CLIENT
          https://nrpc.olark.com/nrpc/c?j=olark-11482534284183r1&&c=create&q=1813.841…9648c703eb70132ff220&version=api-1.2.1&pretty=true&_rnd=0.3181155013024206 Failed to load resource: net::ERR_BLOCKED_BY_CLIENT
          https://log.olark.com/jslog/log.png?version=framestore&location=https%3A%2F…3&level=error&timestamp=1482534292322&properties=%7B%7D&recent_logs=%5B%5D Failed to load resource: net::ERR_BLOCKED_BY_CLIENT
          https://pixel.wp.com/g.gif?v=wpcom-no-pv&x_newdash_pageviews=route&t=0.7103165142625483 Failed to load resource: net::ERR_BLOCKED_BY_CLIENT
          https://pixel.wp.com/g.gif?v=wpcom-no-pv&x_newdash_pageviews=route&t=0.14961502792532078 Failed to load resource: net::ERR_BLOCKED_BY_CLIENT

          Like

        2. Hmm, that sounds like something in your browser is blocking a bunch of network requests generally. It’s probably not related to Export itself but could be related to a proxy, security software, or a transient network glitch.

          Liked by 1 person

        3. You are probably right, it is probably the Adguard extension, I have the English list and Spyware (blocks various types of trackers) list and Malware list enabled.

          -John Jr

          Like

  5. I exported only posts from 2007 from jimgrey.wordpress.com and imported them successfully into jimgreytest.wordpress.com. So I wonder if it’s sheer size of a longtime blog that’s the challenge.

    Like

  6. Hi Jordan,

    I’ve tested on Ubuntu 14.04.4 LTS running Firefox 47.0 and gave the new calypso export feature a go.

    Can you export all content from your site?
    Works fine. 🙂

    Can you export a specific selection of content from your site?
    Works fine. 🙂

    Can you download the export file when complete?
    I was able to download an export file on every test I made. 🙂

    Do you receive an email after the export completes?
    Yes, I’ve received an email for every export I initialized. 🙂

    Are you able to take the exported content and import it into a new site?
    I’ve only tested with a very small file where I exported some posts from a specific month. I can say, overall this worked fine apart from some confusing aspects I’ll describe in the next part.

    Is there anything you found confusing during this process?

    I was confused at first about where to find the screen for exporting specific content. It took me some some seconds to find the drop down arrow on the right. The current layout doesn’t make much sense to me in my eyes. (Screenshot)
    The DropDown arrow should be limited to the part about exporting specific content. I’ve made a mockup to illustrate what I mean.

    When I decide to export only specific content (posts, pages, feedback) and I do so, the screen shows two spinner controls. One above near the “Export all” option and one below where you export specific content. (Screenshot) Doesn’t make much sense to me. If I decide to export only specific content, only the lower spinner should work.

    The content of the DropDown controls (Status, StartDate, EndDate) hasn’t been localized yet. The translation for all the states (pending, draft, …) are already listed in GlotPress. Same goes for the dates. The month abbreviations can be found in GlotPress.

    Exporting the posts for a specific month requires using both dropdowns (Start- & EndDate). Otherwise if I only specify a StartDate, the export feature will download all posts following the start date. How about instead of a Start- & Enddate dropdown, you show one “Pick a time range” control (combined startdate & enddate) and one dropdown which lets me download the content for a specific month. That would save some clicks when creating backups for certain months.

    If I pick some weird combinations (StartDate + EndDate + Category) when trying to export specific content which would result in not finding exportable content, the export dialogue doesn’t show a (red) warning notifying me about it but instead offers me to download a zip file. Only after opening this zip file will I find out it’s empty.

    The filename for the zip file currently looks like: BLOGURL-YYYY-MM-DD-hh_mm_ss-“32-random-chars-string”.zip
    Why the long random number? I guess the filenames would be unique enough with the “seconds” being part of the filename. Can you give me a reason for why we would need those complicated, long file names?

    I find it strange that an export feature exports files that can’t be imported on its own. I would have to extract the zip file at first before importing. How about extending the importer so that it can work with zip files as well?

    Uploading the export file doesn’t automatically start the import or at least the part of the process where I can assign different authors. When I tested importing the file, I thought the import was completed when the file upload was done. Checking my testblog for the newly imported posts and not finding them left me confused. It’s only then when I checked the import screen again to find out you have to push another button to start the import after the file has been uploaded. (Screenshot)
    Why do you wait for user input which can only be pushing that “Start Import” button? Or asked the other way around, why would I not push the “Start Import” button when I decide to upload the WXR-file?

    When I tested the import, it was successfully completed but the screen always displayed the message “Finishing up the import”. (Screenshot).
    I had the impression the process was stuck because nothing else happened. So I stopped the import and checked the result. All categories, tags and posts were imported successfully it seems. So why the “Finishing up the import” message?

    Cheers
    Martin K. aka iqatrophie

    Liked by 1 person

    1. Thanks for your thorough feedback, Martin! I’ll address a few things specifically.

      I’ll share the UI feedback with a designer. We’re tending to reuse components to try to be consistent across the app, but I see what you mean re the down arrow. I don’t know that we’ll change that pre-launch, but I’ll surely pass the feedback along, not only for Export but as general feedback about the UI in Calypso. I’m not sure we’ll implement a combined date range picker, but I’ll surely direct our designer to your comment.

      I’ll see what I can figure out about the translations of the date dropdowns.

      The matter of picking combinations of date/category/etc. and finding an empty export seems likely to be an edge case, so it’s not something I think we should consider holding launch for. I suspect the core (non-wordpress.com) exporter is subject to the same issue. I’ll make a note of it for the future.

      There are dull legacy implementation details behind the random characters in the filename. I’ll raise an issue to investigate whether it’s plausible to improve that (we need the random stuff, but maybe we could move it to a directory rather than junking up the filename itself).

      The inability to import the zip file itself is complicated because the core importer would also have to be updated, and I think it may be complicated further by having to support zip file expansion on multiple platforms, etc. I agree that it’d be nice to be able to upload the zip file. I’ll raise an issue for the future to see if that’s something we might be able to support on wordpress.com at least.

      Regarding the import process, I believe the need for the second start button arises in part out of a need to validate the file. If somebody tries to upload an invalid file, we show an error message. I believe there were other conscientious decisions around making each step of the process discrete and easily resumed at a later time, though I don’t suppose auto-forwarding to the next step in the process (author mapping) after upload would contradict those decisions. I’ll share this feedback with a designer for further thought.

      The “finishing up import” issue is the result of a delayed post-import process that handles replacing image urls in posts. Once the content seems to have been completed, we kick off some asynchronous processes that scour posts for references to images at their original locations and replace them with new image urls. For larger sites, this can take some time, so there’s a delay. For smaller sites, the delay isn’t optimal. Basically, it’s a known issue that is lower priority than a lot of other known issues. We could reduce the delay, but it is sadly more complicated than it would seem to need to be.

      Really appreciate all the great feedback!

      Liked by 1 person

      1. … and I think it may be complicated further by having to support zip file expansion on multiple platforms

        Is zip file expansion so different from zip file creating? Since the export feature creates zip files, don’t you have to make sure all those multiple platforms support zip file creation as well?

        Like

Leave your Feedback

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s