Unable to manipulate .odt file downloaded from Google docs

[tddrmllr]

I'm trying to download a Google doc in .odt format, manipulate it with odf-report, then update the Google doc with the modified content.

I'm able to create a .odt file locally (with OpenOffice), manipulate it with odf-report, then upload it to Google just fine. But if I use a .odt file downloaded from Google docs, odf-report doesn't make the changes. Any idea why?

I experienced some trouble with Google spreadsheets in a different context roo-rb/roo#166. Perhaps Google docs has some similar nuances?

Below is a gist that outlines my proof-of-concept implementation. It also has a sample .odt file that was downloaded from Google, and that does not work with odf-report.

https://gist.github.com/tikitikifoofoo/3cef692c7fa7f66a9ffa#file-00test-odt

[atzkey]

I confirm this issue. Opening the downloaded file in the OS X's TextEdit and saving it (without making any changes) seems to fix the issue, so there are definitely some quirks with Google docs exporter, even though it generates seemingly valid files.

[frenkel]

In my testing I discoved that the Google Docs xml's contain extra text:span items around the brackets, which will result in the tags not being recognized. Saving in LibreOffice merges the tags and fixes the problem. I think the best workaround is a detection of these text:span elements and replacing them as well.

[frenkel]

I've created a pull request for this: #76

[sandrods]

Sorry for long time, but I was no able to reproduce your problem. It only happens when you break the golde rule: NEVER edit your placeholders. If you have to chage it, delete and type all over again.

[frenkel]

@sandrods Did you look at the pull request?

[sandrods]

Yes I did, but as I said above, I couldn't confirm the behavior you pointed out. I examined odts downloaded from Google Docs and could not find the issue described.
Can you confirm it is still happening?

[frenkel]

Ah I see that Google fixed the problem on their end. But wouldn't it be wise to merge the PR anyways? It can't hurt to ignore empty tags that might be added to the odt by some program that doesn't keep the source clean.