Using inkscape-21256 snapshot (because release version has bug causing core dump when attempting to change fill color as reported numerous times).
src/display/nr-arena-shape.cpp has #include <fenv.h>
But I don't see any use of floating-point definitions, FE* macros, or fe*() functions.
I commented that line out and inkscape built fine. (Also some platforms don't have any fenv.h.)
Also src/live_effects/lpe-dynastroke.cpp uses fmin and fmax.
Some platforms don't have that. So I added to src/live_effects/lpe-dynastroke.cpp :
#define fmin(a,b) ((a)<(b)?(a):(b)) #define fmax(a,b) ((a)>(b)?(a):(b))
Also work/inkscape-21256/src/display/nr-filter.cpp refers to fmax but it is commented out.
Hi all,
I hate to butt into a conversation here, but Jeremy is exactly correct on this one. Our inkscape-specific isnan.h and its ilk really need to be rolled up into a single .h that has all of the switches/aliases/etc for all of the float-classification needs that we have. It should all be grouped into a single .h so that downstream developers don't need to worry about it. Map your needs to Inkscape-specific names so that there is no fear of collision.
Did we not agree to this very exact thing >2 years ago?
By the way, 0.47 is looking very good. with Bulia as the warden, how can it fail? :-)
bob
p.s.:
("butt" is the correct spelling for this metaphor. like being hit by a male goat :-)
On 4/29/2009 4:17 PM, Jeremy C. Reed wrote:
Using inkscape-21256 snapshot (because release version has bug causing core dump when attempting to change fill color as reported numerous times).
src/display/nr-arena-shape.cpp has #include<fenv.h>
But I don't see any use of floating-point definitions, FE* macros, or fe*() functions.
I commented that line out and inkscape built fine. (Also some platforms don't have any fenv.h.)
Also src/live_effects/lpe-dynastroke.cpp uses fmin and fmax.
Some platforms don't have that. So I added to src/live_effects/lpe-dynastroke.cpp :
#define fmin(a,b) ((a)<(b)?(a):(b)) #define fmax(a,b) ((a)>(b)?(a):(b))
Also work/inkscape-21256/src/display/nr-filter.cpp refers to fmax but it is commented out.
Wait. Bulia is not a warden for this release... he's just pushing it along. ;-) I think I'm the only official warden currently, unless I don't remember another volunteer (entirely possible).
On Apr 30, 2009 8:24 PM, "Bob Jamison" <rwjj@...127...> wrote:
Hi all,
I hate to butt into a conversation here, but Jeremy is exactly correct on this one. Our inkscape-specific isnan.h and its ilk really need to be rolled up into a single .h that has all of the switches/aliases/etc for all of the float-classification needs that we have. It should all be grouped into a single .h so that downstream developers don't need to worry about it. Map your needs to Inkscape-specific names so that there is no fear of collision.
Did we not agree to this very exact thing >2 years ago?
By the way, 0.47 is looking very good. with Bulia as the warden, how can it fail? :-)
bob
p.s.:
("butt" is the correct spelling for this metaphor. like being hit by a male goat :-)
On 4/29/2009 4:17 PM, Jeremy C. Reed wrote: > Using inkscape-21256 snapshot (because release ver...
Even better. Your words are my command, Josh. :-)
bob
On 5/1/2009 1:01 AM, Josh Andler wrote:
Wait. Bulia is not a warden for this release... he's just pushing it along. ;-) I think I'm the only official warden currently, unless I don't remember another volunteer (entirely possible).
These problems are still in latest snapshot (inkscape-21470).
So I opened two bugs:
https://bugs.launchpad.net/inkscape/+bug/383246
https://bugs.launchpad.net/inkscape/+bug/383247
Jeremy C. Reed
echo 'EhZ[h ^jjf0%%h[[Zc[Z_W$d[j%Xeeai%ZW[ced#]dk#f[d]k_d%' | \ tr '#-~' '-.-{'
participants (3)
-
Bob Jamison
-
Jeremy C. Reed
-
Josh Andler