John Levon <[EMAIL PROTECTED]> writes:
| On Thu, Aug 04, 2005 at 06:44:39PM +0200, Lars Gullik Bj?nnes wrote:
>
>> | It's impossible to achieve both testing of performance and debugging
>> | then. I can't run two LyXes at once you know. You've chosen debugging
>> | when we're not in a heavy surger
On Thu, Aug 04, 2005 at 06:44:39PM +0200, Lars Gullik Bj?nnes wrote:
> | It's impossible to achieve both testing of performance and debugging
> | then. I can't run two LyXes at once you know. You've chosen debugging
> | when we're not in a heavy surgery phase, bit in a tweaks and fixes
> | stage.
[EMAIL PROTECTED] (Lars Gullik Bjønnes) writes:
>>> But there are two levels of this debug support, normal and pedantic
>>> and we turn on both. I am not sure about what we loose by only using
>>> one of them. That might be a solution.
>>
| | Ah. Can you investigate?
>
| I can ... if the gcc gocs
John Levon <[EMAIL PROTECTED]> writes:
| On Thu, Aug 04, 2005 at 05:21:53PM +0200, Lars Gullik Bj?nnes wrote:
>
>> | Hmmph. It's hard to gauge where LyX is wrt performance if this is on by
>> | default - daily use of the latest LyX is how much stuff gets caught.
>>
>> turn off all debuging if try
On Thu, Aug 04, 2005 at 05:21:53PM +0200, Lars Gullik Bj?nnes wrote:
> | Hmmph. It's hard to gauge where LyX is wrt performance if this is on by
> | default - daily use of the latest LyX is how much stuff gets caught.
>
> turn off all debuging if trying to gauge whre LyX is wrt performance.
> As
John Levon <[EMAIL PROTECTED]> writes:
| On Thu, Aug 04, 2005 at 04:24:56PM +0200, Lars Gullik Bj?nnes wrote:
>
>> yes I think so,
>> and we should already know to not use this for profiling builds.
>
| Hmmph. It's hard to gauge where LyX is wrt performance if this is on by
| default - daily use o
On Thu, Aug 04, 2005 at 04:24:56PM +0200, Lars Gullik Bj?nnes wrote:
> yes I think so,
> and we should already know to not use this for profiling builds.
Hmmph. It's hard to gauge where LyX is wrt performance if this is on by
default - daily use of the latest LyX is how much stuff gets caught.
Ca
Jean-Marc Lasgouttes <[EMAIL PROTECTED]> writes:
| Lars> We should at least report this to the gcc dudes. Or at least
| Lars> figure out if stdlib-debug is supposed ot have runtime impact.
>
| Well, if it does run-time checking of the validity of iterator it sure
| has impact. Actually, what shows
> "Lars" == Lars Gullik Bjønnes <[EMAIL PROTECTED]> writes:
Lars> Jean-Marc Lasgouttes <[EMAIL PROTECTED]> writes: |
Lars> It turns out that bug 1922 (navigating allsymbols.lyx is
Lars> extremely | slow) can be solved by configuring with
Lars> --disable-stdlib-debug.
>>
Lars> | That raises a q
Jean-Marc Lasgouttes <[EMAIL PROTECTED]> writes:
| It turns out that bug 1922 (navigating allsymbols.lyx is extremely
| slow) can be solved by configuring with --disable-stdlib-debug.
>
| That raises a question: is it really wise to turn this on by default?
We should at least report this to the
10 matches
Mail list logo
