<div dir="auto">Thanks for the information!<div dir="auto"> </div><div dir="auto">I wanted to read the pixel for my own save function that I am writing, so even if it was slow I would not bother. I want to use pixman to replace Allegro4 bitmaps.</div><div dir="auto"> </div><div dir="auto">From your reply, I think that I need to convert the full bitmap to a different format that a .bmp file accepts and then write these converted bits to the stream all in one go. I still don't know how to do that. :/</div><div dir="auto"> </div><div dir="auto"> </div><div dir="auto">I am sorry, I didn't mean a direct reply, I wanted to email the mailist actually but I misclicked the reply to all.</div></div> <div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Mon, Jul 13, 2020, 04:29 Pekka Paalanen <<a href="mailto:ppaalanen@gmail.com">ppaalanen@gmail.com</a>> wrote: </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">On Sun, 12 Jul 2020 23:24:51 -0300 Erico P <<a href="mailto:eri0onpm@gmail.com" target="_blank" rel="noreferrer">eri0onpm@gmail.com</a>> wrote: > Thanks! > > I am trying to get started but I need to read and write bitmaps to see what > I am doing. I made a function to write bitmaps from my own, and made up > some function to read. I can't seem to find anything readymade that > supports all formats available in pixman (r8g8b8a8, ...). Hi, Pixman's format support is so wide and wild (from 1-bit formats where endianness affects also bit order within a byte, to floating point formats nowdays IIRC), that I doubt you will find even a file format that supports everything as is - or if you do, it's really complicated. > I have a small problem that I can't seem to figure out, is there a function > to read or write a single pixel? Maybe it returns an int, or a color (for a > 32bit bmp), or the index of the palette in case it's an indexed bitmap? It > would be really useful for me to be able to read a single pixel, is there a > function for this I am overlooking? I don't think there is, maybe because using it would make things slow so no-one wanted it. Instead, if you need to, you get the pointer to the beginning of the data, compute the appropriate offset from stride, format and position, and access the bits yourself according to the format. <a href="https://ppaalanen.blogspot.com/2016/02/a-programmers-view-on-digital-images.html" rel="noreferrer noreferrer" target="_blank">https://ppaalanen.blogspot.com/2016/02/a-programmers-view-on-digital-images.html</a> and <a href="https://afrantzis.com/pixel-format-guide/" rel="noreferrer noreferrer" target="_blank">https://afrantzis.com/pixel-format-guide/</a> might help. You might also mimick that by doing a blit with pixman_image_composite32() from your source image into a 1x1 target image in your preferred inspection format. I'd recommend using just few carefully selected formats for your input/output and use Pixman itself to convert to/from those. If you want to continue the chat, I would like to keep it on the mailing list to benefit everyone and you might get more help, too. It would be nice to CC this reply to the mailing list. Thanks, pq > > Em qua., 8 de jul. de 2020 às 13:23, Pekka Paalanen <<a href="mailto:ppaalanen@gmail.com" target="_blank" rel="noreferrer">ppaalanen@gmail.com</a>> > escreveu: > > > On Wed, 8 Jul 2020 11:43:59 -0300 > > Erico P <<a href="mailto:eri0onpm@gmail.com" target="_blank" rel="noreferrer">eri0onpm@gmail.com</a>> wrote: > > > > > Hello, > > > > > > I foundout about pixman recently as a recommended library for efficiently > > > drawing pixels, apparently used when writing a rasterizer and some other > > > use cases in GUIs too. > > > > > > I want to learn how to use it, but unfortunately the webpage states > > "There > > > is currently no documentation besides the source code itself." > > > > > > Is there any documentation available at all, maybe something someone > > wrote > > > some day for someone, anything other than just the source? Maybe a manual > > > from someone who ported it to other language - but I do intend to use it > > in > > > a mixed C/C++ environment. > > > > > > I appreciate any suggestions on getting started with it! > > > > Hi, > > > > I'm not aware of anything particularly about Pixman, unfortunately. > > You'll have to go through code that uses Pixman. > > > > If you want to do rectangular operations, > > pixman_image_composite32() is the main function for practically > > everything. I've used that one, but I never looked into the > > trapezoid or font compositing stuff. > > > > > > <a href="https://gitlab.freedesktop.org/wayland/weston/-/blob/master/libweston/pixman-renderer.c" rel="noreferrer noreferrer" target="_blank">https://gitlab.freedesktop.org/wayland/weston/-/blob/master/libweston/pixman-renderer.c</a> > > is one of Weston compositor's renderers that has a relatively > > simple task of just copying and blending window images. Weston also > > uses pixman_region32 for most region tracking and operations. > > > > <a href="http://ssp.impulsetrain.com/porterduff.html" rel="noreferrer noreferrer" target="_blank">http://ssp.impulsetrain.com/porterduff.html</a> > > explains the Porter/Duff compositing operations. Some or all of > > those are found in pixman_image_composite32(). It doesn't mention > > the use of a mask though. > > > > Since Cairo can use Pixman underneath, I think some Cairo > > documentation about the image operations are applicable as well: > > <a href="https://www.cairographics.org/tutorial/" rel="noreferrer noreferrer" target="_blank">https://www.cairographics.org/tutorial/</a> > > > > > > Thanks, > > pq > > -- Pekka Paalanen <a href="http://www.iki.fi/pq/" rel="noreferrer noreferrer" target="_blank">http://www.iki.fi/pq/</a> </blockquote></div>