New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
introduse the ability to postpone a downloaded update #4263
Comments
Comment 1 by zahari_bgr on 2014-07-11 04:59 I've created a branch named "t4263" in the following repository: This ticket originates from ticket #4057 and this branch is based on t4057. In ticket #4057 you will find an attached diff with the initial changes for this ticket. |
Comment 3 by briang1 on 2014-07-11 09:18 |
Comment 4 by zahari_bgr on 2014-07-17 08:36
I've also added some documentation for the above options. I now check the pending version against the new version when performing an update check. I've also changed a little the update notification GUI:
Finaly, I removed the checks for isInstalledCopy() and let UpdateDownloader() to be used for all update downloads - I think it is now safe and it will eliminate some confusion and browser issues. I've created a test build from my testing branch with this and my other changes, which could be downloaded from: When you check for updates from this snapshot, you will get nvda_testing-3.exe, which is exactly the same, except for the version number. |
Attachment pendingUpdates.diff added by zahari_bgr on 2014-07-18 04:43 |
Comment 5 by blindbhavya on 2014-10-03 11:56 |
Comment 6 by driemer.riemer@... on 2015-05-29 20:19 Code review: It looks like you disable everything but the edit box for a directory in secure mode. Shouldn't this be disabled as well?
Great work. |
I'm so sorry - I haven't seen that there's code review. I don't mind if someone takes this over - I rarely use Windows those days. Same apply for my other contributions, which are still waiting code review. I'll be glad to see my ideas complemented by others and merged in NVDA, since I implemented them with the hope they'll be of help to users. Now while I'm thinking of it, there were also a few other update-related patches, and in one point I was thinking of separate Update preferences dialog. Of course, this is in case NVAccess approves those propositions, and probably should be a subject of another ticket. |
Do you mind putting this into a pr? That way it can be officially review. Sent from a mobile device.
|
I'm not interested at this time. |
We can't do any review on the current code as the attachment seems to be lost in the migration to github. @@zahyur, do you still have the original code somewhere? |
https://bitbucket.org/zahyur/nvda/ |
Thanks for this. I've pulled in the branch and tried to do a rebase to current master. Some little merge conflicts i've tried to fix as good as possible. @derekriemer: Do I understand it correctly that you were willing to do a code review? I'd be happy to collaborate on getting this code into good shape, as I belief that it might be a very handy feature. |
Note that I'm not actually authorized to do official code review. I was On 8/20/2016 8:28 AM, Leonard de Ruijter wrote:
Websites: email me at derek.riemer@colorado.edu mailto:derek.riemer@colorado.edu |
But if you have questions about or disagree with what I said, just On 8/20/2016 8:28 AM, Leonard de Ruijter wrote:
Websites: email me at derek.riemer@colorado.edu mailto:derek.riemer@colorado.edu |
I was just wondering whether you'd be interested in working together on
this project to make it suitable to be included in official releases.
That is, if the core developers are interested in including a sanitised
pull request.
|
Sure. |
Have looked into this again. I have the following thoughts:
|
Happy to support postponing of downloaded updates, but I don't think we should allow the user to choose the location at all. If a user wants to download a build for later use, this should be downloaded from the web, rather than relying on an auto update mechanism.
|
Hi, |
I should have been clearer: I agree the temporary directory isn't
suitable once you can postpone updates. However, I still don't think
this should be configurable, as it increases complexity and is the kind
of configurability that might confuse users.
|
Hi, Best wishes, |
I belief things should be as follows:
1. Downloads should be saved to the temp folder upon download.
2. If one decides to posphone, copy de download to the updates folder under userconfig
3. If the latter fails, tell the user that the download can't be saved and that he either should install directly or download again later.
|
I also agree that this is the better way to go. On 9/4/2016 12:59 PM, Leonard de Ruijter wrote:
Websites: email me at derek.riemer@colorado.edu mailto:derek.riemer@colorado.edu |
Hi, Best wishes, |
I'm not sure I see a need to use the temp dir at all, as copying for
postpone seems rather pointless. Eventually, we'll probably want an
option to do automatic background downloads which would make the temp
dir even less viable.
|
For portable copies on flash drives, no temp dir would mean that the download is saved to the flashdrive and than opened from it. In the case of a direct download-update, this would mean that there is a decrease in performance when compared to an update from an installer on the local drive.
|
True, but right now, we don't do update downloads for portable copies
anyway.
|
The current code for this ticket allows for this. It downloads and than starts with the --launcher parameter instead of --install-silent
|
|
Agreed; just keep until installed or next one is downloaded.
|
@jcsteh @LeonarddeR Could you please update us about the status of the code review for this feature? |
* initial work on postponeing updates * Removed undesired code from the source and userdocs and added some copyright notices * Review actions * Some major code updates to make this actually work, hopefully * Fixed translator comment * Some more fixes, still some testing to be done * Support updating of portable copies * Review actions * More review actions * Actions for user guide * Fixed issue where a right paren was placed at the wrong position * Properly evaluate whether to show or hide the pending updates menu item on every time the menu is opened. Binding this evaluation to the item itself made the item never to show up. * Don't use the file variable name when looping through possible old updates. This broke the loading of the state file. Added state file loading debug logging while at it. * Properly destroy the UpdateResultDialog when installing a postponed update * When updating a portable copy, run NVDA with the -m command line argument * Fix KeyError on updating * make sure that the update state file will be updated to contain pending update info * When updating a portable copy, place paths that are provided as command line params between quotes to deal with paths containing spaces * Fix crash after restart with a postponed update * No longer add removeFile to the update state, as it is used nowhere * Call prePopup before raising the UpdateAskInstallDialog and make it modal * Use runScriptModalDialog for the UpdateAskInstallDialog * Clear the don't remind version when postponing an update. If we do not do this, a user will never be reminded about a postponed update automatically at all
Reported by zahari_bgr on 2014-07-11 04:43
Hi,
currently, when one downloads an update when using installed copy, they have no other choice except to run the update within the current session.
This is not always the desired behaviour, cause the user may want to do other stuff before that and may need to restart either NVDA or the system. Redownloading is not very nice, cause not everyone have a persistant internet connection, not to menssion the bandwith and traffic.
I propose to be introduced a mechanism for postponing the install of the update and executing it on user's demand.
The text was updated successfully, but these errors were encountered: