Re: lynx-dev SOURCE_CACHE "problem" - proposal of SOURCE_CACHE_FOR_INCOMPLETE

[Vlad Harchev]

On Sat, 8 Apr 2000, Klaus Weide wrote:

> On Sat, 8 Apr 2000, Vlad Harchev wrote:
> > On Fri, 7 Apr 2000, Philip Webb wrote:
> > > HOW the interruption occurred is irrelevant to user needs.
> > 
> >    Yes, exactly.
> 
> There is a big difference between "I pressed 'z'" and "the network
> connection broke".  I don't thing that difference is "irrelevant to
> user needs".

  Agreed. But I don't think that network connections break as frequently as
user presses 'z'.
 
> >    May be it would still be better to provide lynx.cfg option to allow
> 
> Go ahead if you feel like it, but...
> 
> > selecting either of 4 options above since there is no consenus:
> > 
> > SOURCE_CACHE_FOR_INTERUPTED:
> 
> <something>_FOR_INCOMPLETE (or actually, for _DETECTED_ _INCOMPLETE)
> would be more truth in advertising.  Make the name reflect what it
> does, not the one situation you have in mind.  Don't give the impression
> that this affects only intentional interruptions, unless you *know* that
> to be the case.

  I don't plan to detect the cause of aborting in my implementation. May be
SOURCE_CACHE_FOR_ABORTED would be more meaningful and correct name (since
support for it will go only to CacheThru_abort :)?
  Anyway, I don't think that providing documentation on the option in its
name is normal - we can provide description in "description" section :)
 
> >     values are:
> > OK             variant 1
> > PREVIOUS_OR_OK variant 1a
> > DROP                variant 2
> > PREVIOUS_OR_DROP variant 3
> 
> Not all 4 need to be offered.  A choice between 1 and 2, or 1a and 2,
> could be enough.

  May be, I will decide what to leave when I will be implementing it. 
  If anybody disagrees on names of values, comments are welcome.

> Btw, it seems 2 is the only one that offers, as an obscure feature, the
> possibility to completely remove one specific file's cache entry:  do
> something that initiates a reload-with-reparsing, then (with the right
> timing?) press 'z'.

   Why user will want to remove source cache entry? User can refresh source
cache by pressing ^R - seems it's enough.

> >  As for default, I propose PREVIOUS_OR_OK.
> >  What do you think? 
> 
> The default should be 2 or 3, IMO.  SOurce-cacheing should not by default
> introduce more situations where a corrupt copy can stick around without
> the user noticing.

    Questionable, but probably we will select the variant you propose. When
users enables source cache when configuring, they (I think) expect lynx to
cache very aggressively, so this default could confuse them. But I don't know
what the most correct default is - so let's select yours.
 
>     Klaus
> 

 Best regards,
  -Vlad