On 01/09/2010 01:16 PM, Abdelrazak Younes wrote:
On 09/01/2010 16:33, Abdelrazak Younes wrote:
On 09/01/2010 16:26, rgheck wrote:
Abdel, I haven't tested, but in the work I did yesterday on XHTML
output I was seeing some sub-optimal behavior involving the BibTeX
cache, namely, that the cache isn't cloned. It doesn't look to me as
if this would be fixed by what you have done; if not, then is the
solution (for now, anyway) for me to add appropriate code to the
clone() routine, e.g.:
buffer_clone->d->bibfiles_cache_ = d->bibfiles_cache_;
buffer_clone->d->bibinfo_ = d->bibinfo_;
etc?
Or rather in the Impl ctor (Buffer::Impl::Impl).
I did that for you.
Thanks.
Is there a straightforward way to know if we are in a cloned buffer? I
think it would be safest not to allow re-reading of the bibliography
data as we are doing output. There's an obvious race condition here, one
we had anyway, but now we can prevent it.
rh
