In LyXText::setHeightOfRow, I read:
/* this must not happen before the currentrow for clear reasons.
so the trick is just to set the current row onto this row */
int unused_y;
getRow(row_ptr->par(), row_ptr->pos(), unused_y);
What does this do??? As far as I c
On Tue, Nov 20, 2001 at 04:39:32PM +0100, Jean-Marc Lasgouttes wrote:
> What does this do??? As far as I can read, getRow does not modify
> anything but unused_y (which is truly unused). It does return a row,
> but this is just lost.
Does it sill work if you throw it out? *grin innocently*
> T
On Tue, Nov 20, 2001 at 05:16:57PM +0100, Lars Gullik Bjønnes wrote:
> | Btw, I just noticed that Row hast 'unsigned short' interspersed with
> | 'unsigned int'. This does not buy a single byte of memory, and on IA32
> | shorts are slower by a factor of three or so than 'normal int'...
>
> Almost
> "Andre" == Andre Poenitz <[EMAIL PROTECTED]> writes:
Andre> On Tue, Nov 20, 2001 at 04:39:32PM +0100, Jean-Marc Lasgouttes
Andre> wrote:
>> What does this do??? As far as I can read, getRow does not modify
>> anything but unused_y (which is truly unused). It does return a
>> row, but this i
> The only change I noticed is that, for a very large file, total time
> counted by gprof goes from 5.68s to 3.26s :)
Uuh... that is really bad... this goes exactly into the opposite
direction than the one 1.2.0 development has followed all the time
(apart from temporary dark times in Bozano or s
> "Andre" == Andre Poenitz <[EMAIL PROTECTED]> writes:
>> The only change I noticed is that, for a very large file, total
>> time counted by gprof goes from 5.68s to 3.26s :)
Andre> Uuh... that is really bad... this goes exactly into the
Andre> opposite direction than the one 1.2.0 developme
On Thu, Nov 22, 2001 at 10:19:15AM +0100, Jean-Marc Lasgouttes wrote:
> Actually, I think I know what happened. LyXText::GetRow used to have
> the side effect of setting the currentrow member of LyXText, which was
> some kind of cache. Sometimes in the 1.1.5 development cycle, this
> currentrow ha