Re: sqlite3

[cesar mena <cesar.mena@gmail.com>]

Eric Abrahamsen <eric@ericabrahamsen.net> writes:

> cesar mena <cesar.mena@gmail.com> writes:
>
>> Eric Abrahamsen <eric@ericabrahamsen.net> writes:
>>
>>> Lars Ingebrigtsen <larsi@gnus.org> writes:
>>>
>>>> Eric Abrahamsen <eric@ericabrahamsen.net> writes:
>>>>
>>>>> Well, Lars is the Gnus maintainer, not me, so all I can do is propose!
>>>>> But yes I am, though only for group marks. Right now .newsrc.eld
>>>>> combines concerns: it saves variables such as `gnus-server-alist' and
>>>>> all the topic stuff that would be better saved as customization options
>>>>> (or saved in a separate file, using the customization machinery). It
>>>>> also saves group parameters, right next to the marks.
>>
>> Yes, it's a bit of a big hammer. 
>>
>>>> It's a bit much, but I don't think those bits can be usefully separated
>>>> out.  But things like `gnus-save-duplicate-list' (which goes to a
>>>> separate, ad-hoc file) would fit well as a simple persistent variable.
>>>
>>> If `gnus-newsrc-alist' were saved as-is, but *without* the mark data,
>>> that mark data could be pulled out of a separate storage backend at
>>> startup time, with a simple group-name->mark-data mapping. The
>>> .newsrc.eld file could otherwise remain the same.
>>
>> One can edit flags through G p in the topic buffer; that interface would
>> need to be preserved.  Who knows what other secrets lurk in there.
>
> My proposal would only affect the save and load routines (zip and unzip
> data), in-session behavior wouldn't change.

Yes, but there are so many little details that would prop up (I
think). Like what happens if you couldn't load the marks for one of my
2k (I exaggerate ... slightly) groups?  Would you turn off expiry to
prevent the deletion of dormant articles? etc ... 

>>> But Qiantan's message reminded me that this proposal is actually a Gnus
>>> issue, not a sqlite issue, and shouldn't clog up this thread. I'll keep
>>> this for later!
>>
>> Yes, but in the end separating flags from .newsrc.eld is not for the
>> faint of heart, and I am not sure I see the benefits. 
>>
>> Like you said, for later.
>
> I have any number of pie-in-the-sky ideas for Gnus. If I get around to
> implementing half of them, I'll consider it a success :)

You're doing the lord's work. I appreciate it :)

Cheers!