-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/7753/#review13699
-----------------------------------------------------------


Any comments on this? Is this a feature we would like to have?

- rocklund


On Oct. 27, 2012, 3:22 p.m., rocklund wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/7753/
> -----------------------------------------------------------
> 
> (Updated Oct. 27, 2012, 3:22 p.m.)
> 
> 
> Review request for wave.
> 
> 
> Description
> -------
> 
> This patch depends on https://reviews.apache.org/r/7683/ and needs slight 
> modification to work properly together with https://reviews.apache.org/r/7353/
> 
> This is a draft of a patch that makes the digest list update faster with a 
> new digest/inbox item in the search/digest list when the user creates a new 
> wave. This is done by adding a temporary local digest that listens to wavelet 
> updates until the real digest comes in from a search result. Then the the 
> temporary digest gets removed and replaced by the real one.
> 
> It also fix "WAVE-374: Browser 'backwards' button doesn't update wave list" 
> by selecting waves when they are about to be opened. This also results in the 
> digest properly being selected after a browser "refresh".
> 
> I need some input/feedback on the following questions:
> 
> 1. Since the temporary local digest stays until the real digest is returned 
> to the client from a search result it will not disappear if the user have an 
> active search that never finds the new wave. Currently any other search than 
> inbox or "all" would result in only showing the temporary digest. The 
> temporary digest would however disappear as soon as the user searches for 
> inbox or "all" again. Would this be an OK limitation for now or should it be 
> solved somehow:
>    - For example by not creating a temporary digest if the search is not 
> inbox or "all"
>    - Include the newly created wave in the searches that is being sent with 
> 15 sec interval so that the search can report when the new digest is properly 
> created on the server-side
>    - Listen for the server-side creation message in the newly created 
> wavelet/digest itself.
>    - Any other suggestion?
>    - How did Google Wave work when you created a new wave that was not 
> visible in your current search? (not that we need to copy google wave but it 
> could be a possible solution)
> 2. The same problem arises if the user creates a new wave and then quickly 
> change the current search from "all" or inbox to something else.
>    - Could be solved by some of the above suggestions.
>    - It could also be solved by removing the temporary digest if the search 
> result gets changed from "all" or inbox.
> 
> 
> While this is mostly a fix that is needed now when we don't have live search 
> I think it is something that we would want even when live search is in place. 
> It will always take some time for the newly created wave to get properly 
> created on the server side, so you always want some temporary digest to 
> appear fast.
> 
> 
> Any feedback and opinions are welcome.
> 
> Thanks
> 
> 
> This addresses bug WAVE-374.
>     https://issues.apache.org/jira/browse/WAVE-374
> 
> 
> Diffs
> -----
> 
>   src/org/waveprotocol/box/webclient/client/SimpleWaveStore.java b668d32 
>   src/org/waveprotocol/box/webclient/client/WebClient.java 863ae6c 
>   src/org/waveprotocol/box/webclient/client/WindowTitleHandler.java 2eeec81 
>   src/org/waveprotocol/box/webclient/search/Search.java 152c142 
>   src/org/waveprotocol/box/webclient/search/SearchPresenter.java b136c88 
>   src/org/waveprotocol/box/webclient/search/SimpleSearch.java 5b018d4 
>   src/org/waveprotocol/box/webclient/search/WaveStore.java 2111f64 
> 
> Diff: https://reviews.apache.org/r/7753/diff/
> 
> 
> Testing
> -------
> 
> Tested on a local server but could use some more test when we've decided what 
> to do with the above questions
> 
> 
> Thanks,
> 
> rocklund
> 
>

Reply via email to