Re: [SM-DEVEL] Ship quicksave plugin by default

[Marc G. K. <ma...@gr...>]

Paul Lesneiwski wrote:
>

>
> Thijs Kinkhorst wrote:
>
>> Hello people,
>>
>>
>> I propose to make the quicksave plugin, that uses JavaScript to save
>> your message while composing, in the case of problems, a packaged plugin.
>> Looking at the download numbers this makes sense, and I think it
>> addresses a very common problem.
>>
>> Let me know what you think.
>>
>
> I would only hesitate because I think the plugin gives the false
> impression that it will save everything all the time.  And that is
> certainly not the case.  Javascript could be disabled, and moreover, due
> to cookie size limits, longish emails will get unexpectedly cut off (which
> probably makes some people very irate).
>
> (An aside here: I have a new version of the timeout_user plugin in
> development that saves FULL messages to drafts when users are timed out by
> the timeout_user plugin... not useful in all the ways that quicksave is,
> but it is good to know about.)
>
> Furthermore, the failure detection code is *VERY* sensitive; it
> currently fires on the onLoad event (IIRC) and compares the current message
> body to what it has in its cookie(s).  Due to the vast number of ways that
> the message body can be reloaded and changed (adding sig, adding
> attachment, adding recipients from addr book, doing some other unknown
> things with another plugin, etc etc), the "Do you want to recover" message
> can (and does) pop up when it should not.  For casual users, this
> typically is not a problem (and I don't really see many bug reports at
> all), but power users have definitely noticed this very annoying problem.
> I get it semi-frequently, and it is not a big deal to
> me to know that it is a false positive and just press Cancel... but is this
> something we want to make core?
>
> I do have some plans to rewrite that comparison code using regexp
> instead of string comparisons, which may increase JavaScript/Browser
> version dependencies, but will hopefully help the false alerts somewhat.
>
> Oh, and the other thing is that users who enable quicksave on slower
> computers will see a noticable slowdown in their system when typing
> messages while it tries to write a new cookie to disk every second or
> so....
>
> I don't know how well (or not well) those caveats match up to the
> potential problems that are part of the rest of the core, but there you
> have it....
>
> - paul
>

I agree that it's a problem and we should come with a solution for the
session timeout and based on the remarks of Paul i'm not sure if the
quick_safe plugin is the best solution (peronally I never looked at the
quick safe plugin so i don't know exactly how it behaves).

A long time ago I added some code that probably is broken at the moment
but the theory behind it is saving the $_POST vars in a new session on
timeout and continue with composing the message after the user
revalidated. Of course we should give some sensible feedback information
about a timeout and that the message is fully recovered if the user login
again.

What's the opinion about my approach ?

Regards,

Marc Groot Koerkamp


-- 
"It requires wisdom to understand wisdom: the music is nothing if the
audience is deaf." - Walter Lippman