--
You received this message because you are subscribed to the Google Groups "OOO Libs" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ooo-libs+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ooo-libs/70665335-29d8-47e6-a2df-97b601dbbb40n%40googlegroups.com.
2) The local directory keep a copy of all versions of the file, it does not delete the versions that are not used anymore.
Hi, after using getdown for some months, I feel more confident about the behavior I described on my first message.
Getdown never deletes a downloaded file and it makes sense. And it would be no issue keeping all those old files.
But I really noticed that that, using unversioned mode, when a new getdown.txt gets downloaded, it still assumes the list of resouces and jars from the previous getdown.txt.
For the files that kept the original name, it correctly validates the file checksum and downloads them again if the checksum has changed. However, it tries to download old files that existed in the old getdown.txt (and do not exist anymore on the server and are not mentioned anymore in the new getdown.txt). After a number of retries, getdown stops and shows an error message to the user.
When the user tries to start the application again, then getdown reads the correct (new getdown.txt) and checks the correct files (as described in the new getdown.txt).
It least, I could observe it consistently from the logs. I hope to have an opportunity to get further details.
This issue does not happen on versioned mode.