On Sun, May 08, 2005 at 21:31:21 +0200, Tobias Oetiker wrote: > libart does not do anything 'expensive' before the gfx_render_png > step ... you can verify this by creating eps output ... this is > about 8 times faster than png output ... > > so what you need to implement is > > gfx_render_pnggd Should have posted this earlier, but the results were so disappointing that I never got around to it. First of all, with gd + freetype the increase in speed was insignificant (~10-15% or so) so I decided to use gd without freetype. The downside of this is that it's not trivial (AFAIK) to support arbitrary rotations of the text (and rotations that aren't even multiples of 90 degrees look pretty bad), and --zoom will make the text look disastrous, so it's not supported. I did some benchmarking and the result is a 2x speedup or so, which unfortunately isn't enough considering the downsides of the patch. There's no patch of configure or docs, simpy because I don't see any need for the patch to be included. Personally, I'll go with a patched libart instead. Btw, something I noticed with PNG output (that is, not GDPNG): if background color has an alpha channel, the color information is discarded (changes to white). For example, #ff000033 doesn't give you semi-transparent red, it's semi-transparent gray. Feature or bug? Also, it appears as if rrd_graph.c calculates the width of "M", but rrd_graph_helper.c inserts an "m". Perhaps this fixes it (or maybe it should be the other way around, "m" is probably not as high as "M", even though the traditional em-space is the width of "M"). --- rrd_graph_helper.c.orig 2005-05-13 12:23:21.000000000 +0200 +++ rrd_graph_helper.c 2005-05-13 12:23:35.000000000 +0200 @@ -401,7 +401,7 @@ (*eaten)--; linecp[*eaten]=' '; (*eaten)--; - linecp[*eaten]='m'; + linecp[*eaten]='M'; if (rrd_parse_legend(linecp, eaten, gdp)) err=1; -- Attached file removed by Ecartis and put at URL below -- -- Type: text/plain -- Size: 12k (13287 bytes) -- URL : http://lists.ee.ethz.ch/p/gdpng.patch -- Unsubscribe mailto:rrd-developers-request@list.ee.ethz.ch?subject=unsubscribe Help mailto:rrd-developers-request@list.ee.ethz.ch?subject=help Archive http://lists.ee.ethz.ch/rrd-developers WebAdmin http://lists.ee.ethz.ch/lsg2.cgi