Legacy ImageMagick Discussions Archive

One more pair of examples then to show why I think it's a bug. The difference between the two commands below is one draws the image at 25x40 and then crops to 20x40, and the other draws the image at 20x40. I would expect the end images to be identical, but they are not. convert \ -size 25x40 xc:whit...

Sorry, perhaps I have poorly described the issue. The fill in the letters should be drawn in the last column of the image regardless of the stroke. The fill is drawn correctly when the stroke is set to black, red, white or presumably any coloured value. It's just not draw at all when the stroke it s...

It appears that the text drawn by annotate doesn't reach the right side of the image when the text is rotated. Instead there is a blank column on that edge. No stroken and there is a blank column: convert \ -size 100x100 xc:white \ -pointsize 17 \ -weight 400 \ -fill black \ -stroke none \ -annotate...

Hi fmw, To be precise the segfault appeared I was testing with PHP cli and the segfault was being generated during the PHP shutdown phase. Yes, it is almost certainly a problem caused by libgomp interacting either with just Imagick, or with some aspect of the PHP internal engine, but I don't have th...

Hi Fmw42 + Bonzo, You are taking a fundamentally wrong approach to security. You have identified a weakness and are trying to patch it by 'blacklisting' behaviour that shouldn't be allowed; that will always eventually fail as it requires you to blacklist every possible way the weakness can be attack...

btw It looks like the following functions are also hard-coded to be zero for the bias value; it's possible some of them should be observing the image bias value:

BlurImageChannel, EdgeImage, EmbossImage, GaussianBlurImageChannel, SharpenImageChannel, and CannyEdgeImage

Hi, I believe I have found the cause of this issue, and the fix appears relatively simple. In file magick/effect.c the following changes (or a close approximation of them) need to be done near line 889: +double + bias; + +bias=image->bias; Line 914-ish: -convolve_image=MorphologyApply(image,channel,...

For reference then, using libgomp.x86_64 version 4.4.7-11.el6 on Centos when using ImageMagick 6.9.9.0 through Imagick has issues; namely it segfaults on exit in the libgomp library. They may be caused by using OpenMP inside an application that is also managing its own threads. So yeah, I'll just ke...

So, will this be merged into the SVN repo?

btw the diff above appears to be the wrong way round.

Hi, I've just spent a few hours trying to figure out why ImageMagick was segfaulting in libgomp. Referencing this post from 2007 it seems that OpenMP crashing is a known issue. Is there a version that is known to be stable on Centos 6.4, or a set of flags that need to be set to make it stable with O...

Yep - that seems to fix it.

In that case, I have opened an issue in the bug forum - viewtopic.php?f=3&t=26655

I believe these each of the pairs of commands below should produce the same image as their pair i.e. curveT and curveQ should be the same image, and curve2T should be the same as curveQT. convert -size 800x500 xc: -fill "#B42AAF" -stroke Black -draw "path 'M50,200 T200,250 z'" cu...

I'm still a little confused by some behaviour - consider the two commands: convert -size 800x500 xc: -fill "#B42AAF" -stroke Black -draw "path 'M50,200 T200,250 z'" curveT.png convert -size 800x500 xc: -fill "#B42AAF" -stroke Black -draw "path 'M50,200 Q50,200,200,...

"I don't understand what you are asking." Every description of the function describes it as drawing a 'smooth bezier curve'. It only seems to draw a straight line. The exact question I was asking is 'how do you use it to draw a curve' ? I think possibly I've possibly found what I was misun...