Re: diffseq: Introduce EQUAL_IDX for comparing indexes instead of elements

2010-04-19 Thread Bruno Haible
Hi Eric, Eric Blake wrote: > > + It is also possible to you this file with abstract arrays. In this > > case, > > s/you/use/ Oops, thanks. Corrected. > > +#ifdef ELEMENT > >ELEMENT const *const xv = ctxt->xvec; /* Still more help for the > > compiler. */ > >ELEMENT const *const yv

Re: diffseq: Introduce EQUAL_IDX for comparing indexes instead of elements

2010-04-19 Thread Eric Blake
On 04/18/2010 04:04 PM, Bruno Haible wrote: > +++ lib/diffseq.h Sun Apr 18 23:53:39 2010 > @@ -53,6 +53,14 @@ > early abort of the computation. > USE_HEURISTIC (Optional) Define if you want to support the > heuristic fo

Re: diffseq: Introduce EQUAL_IDX for comparing indexes instead of elements

2010-04-18 Thread Andreas Gruenbacher
On Monday 19 April 2010 00:04:54 Bruno Haible wrote: > I'm applying this instead. It should be trivial for you to update > patch/src/merge.c accordingly. Sure, your modified version does the same job. Thanks, Andreas

Re: diffseq: Introduce EQUAL_IDX for comparing indexes instead of elements

2010-04-18 Thread Bruno Haible
Hi Andreas, > > patch has a historic > > feature called "plan B" for when only little memory is available; it then > > doesn't load the entire patch and file to be patched into memory. To > > support this, the file and patch are accessed through functions rather > > than directly as arrays. ..

Re: diffseq: Introduce EQUAL_IDX for comparing indexes instead of elements

2010-04-18 Thread Andreas Gruenbacher
On Sunday 18 April 2010 16:13:01 Andreas Gruenbacher wrote: > Hello, > > here is a change to diffseq whcih I need in GNU patch: patch has a historic > feature called "plan B" for when only little memory is available; it then > doesn't load the entire patch and file to be patched into memory. To >

diffseq: Introduce EQUAL_IDX for comparing indexes instead of elements

2010-04-18 Thread Andreas Gruenbacher
Hello, here is a change to diffseq whcih I need in GNU patch: patch has a historic feature called "plan B" for when only little memory is available; it then doesn't load the entire patch and file to be patched into memory. To support this, the file and patch are accessed through functions rath