Re: [boost] Re: Broken HTML doc link for boost::function

From: Rene Rivera (grafik.list_at_[hidden])
Date: 2003-11-12 13:59:53

[2003-11-10] David Abrahams wrote:

>Douglas Gregor <gregod_at_[hidden]> writes:
>
>> On Monday 10 November 2003 10:22 am, Beman Dawes wrote:
>>> At 08:47 AM 11/10/2003, Douglas Gregor wrote:
>>>  >We generate and include the HTML documentation for releases. Checking
>>>  >in the HTML generated nightly would make the CVS repository _huge_.
>>>
>>> Why is that? Presumably there would be no change in most files, so CVS
>>> would not actually change.
>>>
>>> --Beman
>>
>> When we generate HTML from BoostBook, many entities get unspecified names 
>> (like, say, id28193809). So if someone changes something tiny in a
BoostBook 
>> document, all these names could change, making for huge HTML diffs (even 
>> adding/removing HTML files) even though the HTML hasn't changed 
>> "semantically". It's hard to get around this without having a name
mangler in 
>> the BoostBook XSL (ugh).
>
>Might be neccessary.  This is continuing to cause problems.

Don't know how the id is generated, but instead of name mangling an easier
solution is to use a hash, like adler32 if the strings are not big.

-- grafik - Don't Assume Anything
-- rrivera (at) acm.org - grafik (at) redshift-software.com
-- 102708583 (at) icq