XML-RPC metaWeblog.getPost and metaWeblog.getRecentPosts don't return posts' modification date.

[ocollet]

If a post has been modified, there's currently no way to get the date it happened. When retrieving a single or several posts via XML-RPC, the only date returned is the creation date.

This would help to have the modification date, especially in the case of data synchronization.

[ocollet]

Added the post modification date to metaWeblog.getRecentPosts and metaWeblog.getPost

[josephscott]

I like this change, with one exception. The only new field that we should expose is 'date_modified_gmt'. Time zone issues were a major source of confusion for many clients, which is why I added the GMT data fields. I think exposing more non-GMT data at this point will only increase that confusion.

I uploaded a modified version of your diff : ​http://core.trac.wordpress.org/attachment/ticket/19027/xmlrpc.diff - with just the 'date_modified_gmt' field.

I'd be fine with committing this modified diff.

[maxcutler]

Replying to josephscott:

I like this change, with one exception. The only new field that we should expose is 'date_modified_gmt'. Time zone issues were a major source of confusion for many clients, which is why I added the GMT data fields. I think exposing more non-GMT data at this point will only increase that confusion.

Since these date_modified fields are read-only, I'm not as concerned with exposing both GMT and non-GMT versions. If you only expose the GMT version, then XML-RPC clients have to do their own datetime math (and call getOption to figure out the blog display timezone) to determine what the displayed value should be.

[ericmann]

Replying to maxcutler:

If you only expose the GMT version, then XML-RPC clients have to do their own datetime math (and call getOption to figure out the blog display timezone) to determine what the displayed value should be.

I agree. We should expose both the GMT and local post modified fields ... but we should be consistent in how we name them. If we're using post_modified_gmt we should also use post_modified rather than postModified.

[ocollet]

I was actually trying to get consistent with the postCreated property, therefore my suggestion to have the following fields:
postCreated, postModified, post_date_gmt and post_modified_gmt.

Replying to ericmann:

Replying to maxcutler:

If you only expose the GMT version, then XML-RPC clients have to do their own datetime math (and call getOption to figure out the blog display timezone) to determine what the displayed value should be.

I agree. We should expose both the GMT and local post modified fields ... but we should be consistent in how we name them. If we're using post_modified_gmt we should also use post_modified rather than postModified.

[josephscott]

After further discussion I agree that exposing the non-GMT modified date won't hurt anything. My primary recommendation to clients is still to use the GMT field.

I've uploaded xmlrpc.2.diff which is basically the original patch with a few minor changes. The fields are date_modified and date_modified_gmt since there are no existing camelcase field names that we need to worry about in terms of backwards compatibility.

[westi]

This patch looks good - I think we should always expose both our internal GMT representation of a date and the timezone adjusted date because they won't necessarily be based on the current Timezone config of the blog - timezone changes, DST, etc.

[josephscott]

In [19883]:

Expose the post modified date (in blog TZ and GMT) via XML-RPC

Props ocollet
Fixes #19027

[koke]

Probably has the same bug as #19733 looking at the diff.