- Dillo-dev - lists.mailman3.com

Underscores not drawn in pre with background color

by nick＠linicks.net

On Sat, 23 Jul 2016 10:01:59 +0100 Nick Warne <nick at linicks.net> wrote: > On Fri, 22 Jul 2016 20:12:59 +0300 > John Found <johnfound at asm32.info> wrote: > > > > > > Confirmed. I am using "Liberation Mono", but with "DejaVu Sans Mono" > > have the same problem. It is probably because of the characters > > height - the next row of text overwrites the previous. The question > > is whether this is because of Dillo or because of FLTK. > > > > The same font, used in Firefox renders correctly. > > This looks like a bug in the font: > > http://dejavu-fonts.org/wiki/Bugs/Archived > > (you need to scroll down a bit). > > I have version 2.34 installed - will see if I can reproduce it later. OK, I get the same issue. But if you change the font size, thus: <pre style="background-color: white; font-size:105%;">softpoint_with_block() { the underscores are visible. Nick -- Gosh that takes me back... or is it forward? That's the trouble with time travel, you never can tell." -- Doctor Who "Androids of Tara"

July 23, 2016

Underscores not drawn in pre with background color

by nick＠linicks.net

On Fri, 22 Jul 2016 20:12:59 +0300 John Found <johnfound at asm32.info> wrote: > > Confirmed. I am using "Liberation Mono", but with "DejaVu Sans Mono" > have the same problem. It is probably because of the characters > height - the next row of text overwrites the previous. The question > is whether this is because of Dillo or because of FLTK. > > The same font, used in Firefox renders correctly. This looks like a bug in the font: http://dejavu-fonts.org/wiki/Bugs/Archived (you need to scroll down a bit). I have version 2.34 installed - will see if I can reproduce it later. Nick -- Gosh that takes me back... or is it forward? That's the trouble with time travel, you never can tell." -- Doctor Who "Androids of Tara"

July 23, 2016

Underscores not drawn in pre with background color

by johnfound＠asm32.info

On Fri, 22 Jul 2016 17:58:44 +0200 Alexander Voigt <Hole.destructor at gmx.de> wrote: > I have not set any specific font in dillorc, so the default setting is > used. However, when I change the monospace font to > > font_monospace="Monospace" > > then the rendering problem disappears and the underscores are > correctly displayed. When I change to > > font_monospace="DejaVu Sans Mono" > > the rendering problem is present, i.e. the underscores are not > displayed. > > Best regards, > Alex > Confirmed. I am using "Liberation Mono", but with "DejaVu Sans Mono" have the same problem. It is probably because of the characters height - the next row of text overwrites the previous. The question is whether this is because of Dillo or because of FLTK. The same font, used in Firefox renders correctly. -- http://fresh.flatassembler.net http://asm32.info John Found <johnfound at asm32.info>

July 22, 2016

Underscores not drawn in pre with background color

by Hole.destructor＠gmx.de

> > IMHO, this is somehow related to the default fonts set in dillorc or > > the global set system fonts. > > +1 > > A wild shot, but a cheap one, so I'd also try that first. Change your > default font to something else and see what happens. > > Also check FLTK's test programs (IIRC there's a symbols one, and another > that's useful to try fonts). > > If FLTK also has problems check the video driver (older kernel, > older/different video driver, GRUB CLI parameter). I have not set any specific font in dillorc, so the default setting is used. However, when I change the monospace font to font_monospace="Monospace" then the rendering problem disappears and the underscores are correctly displayed. When I change to font_monospace="DejaVu Sans Mono" the rendering problem is present, i.e. the underscores are not displayed. Best regards, Alex On Fri, Jul 22, 2016 at 11:43:14AM -0400, Jorge Arellano Cid wrote: > On Fri, Jul 22, 2016 at 06:19:21PM +0300, John Found wrote: > > On Fri, 22 Jul 2016 10:17:38 +0200 > > Alexander Voigt <Hole.destructor at gmx.de> wrote: > > > > > > > > then maybe it is an FLTK issue? I'm using > > > > > > Debian Jessie GNU/Linux 4.6.0-0.bpo.1-amd64 > > > FLTK 1.3.1 > > > > > > Best regards, > > > Alex > > > > > > > IMHO, this is somehow related to the default fonts set in dillorc or > > the global set system fonts. > > +1 > > A wild shot, but a cheap one, so I'd also try that first. Change your > default font to something else and see what happens. > > Also check FLTK's test programs (IIRC there's a symbols one, and another > that's useful to try fonts). > > If FLTK also has problems check the video driver (older kernel, > older/different video driver, GRUB CLI parameter).

July 22, 2016

Underscores not drawn in pre with background color

by jcid＠dillo.org

On Fri, Jul 22, 2016 at 06:19:21PM +0300, John Found wrote: > On Fri, 22 Jul 2016 10:17:38 +0200 > Alexander Voigt <Hole.destructor at gmx.de> wrote: > > > > > then maybe it is an FLTK issue? I'm using > > > > Debian Jessie GNU/Linux 4.6.0-0.bpo.1-amd64 > > FLTK 1.3.1 > > > > Best regards, > > Alex > > > > IMHO, this is somehow related to the default fonts set in dillorc or > the global set system fonts. +1 A wild shot, but a cheap one, so I'd also try that first. Change your default font to something else and see what happens. Also check FLTK's test programs (IIRC there's a symbols one, and another that's useful to try fonts). If FLTK also has problems check the video driver (older kernel, older/different video driver, GRUB CLI parameter). -- Cheers Jorge.-

July 22, 2016

Underscores not drawn in pre with background color

by johnfound＠asm32.info

On Fri, 22 Jul 2016 10:17:38 +0200 Alexander Voigt <Hole.destructor at gmx.de> wrote: > > then maybe it is an FLTK issue? I'm using > > Debian Jessie GNU/Linux 4.6.0-0.bpo.1-amd64 > FLTK 1.3.1 > > Best regards, > Alex > IMHO, this is somehow related to the default fonts set in dillorc or the global set system fonts. > On Thu, Jul 21, 2016 at 08:14:20PM -0400, Jorge Arellano Cid wrote: > > Hi Alexander, > > > > On Thu, Jul 21, 2016 at 09:50:22PM +0200, Alexander Voigt wrote: > > > Hi Dillo developers, > > > > > > I stumbled upon a page with a lot of code snippets wrapped in a <pre> > > > block, where Dillo seems to have a rendering problem: Underscores `_' > > > are not drawn. I stripped down this real-world-example to the > > > attached minimal html file which reproduces the problem: > > > > > > The page should render as follows: > > > > > > ============================================ > > > softpoint_with_block() { > > > { echo ""; } | ./softpoint.x leshouches > > > } > > > ============================================ > > > > > > However, in the development version of Dillo it renders as follows: > > > > > > ============================================ > > > softpoint with block() { > > > { echo ""; } | ./softpoint.x leshouches > > > } > > > ============================================ > > > > > > As you can see, the underscores in the function softpoint_with_block() > > > are not drawn in Dillo. > > > > > > However, the underscores are drawn correcty if you > > > > > > * remove the style="..." specification from the <pre> block or > > > > > > * remove the line > > > > > > { echo ""; } | ./softpoint.x leshouches > > > > > > from the <pre> block content. > > > > The attached example renders OK here. > > Maybe someone else can reproduce it. > > > > -- > > Cheers > > Jorge.- > > > > _______________________________________________ > > Dillo-dev mailing list > > Dillo-dev at dillo.org > > http://lists.dillo.org/cgi-bin/mailman/listinfo/dillo-dev > > _______________________________________________ > Dillo-dev mailing list > Dillo-dev at dillo.org > http://lists.dillo.org/cgi-bin/mailman/listinfo/dillo-dev -- http://fresh.flatassembler.net http://asm32.info John Found <johnfound at asm32.info>

July 22, 2016

Underscores not drawn in pre with background color

by Hole.destructor＠gmx.de

Hi Jorge, then maybe it is an FLTK issue? I'm using Debian Jessie GNU/Linux 4.6.0-0.bpo.1-amd64 FLTK 1.3.1 Best regards, Alex On Thu, Jul 21, 2016 at 08:14:20PM -0400, Jorge Arellano Cid wrote: > Hi Alexander, > > On Thu, Jul 21, 2016 at 09:50:22PM +0200, Alexander Voigt wrote: > > Hi Dillo developers, > > > > I stumbled upon a page with a lot of code snippets wrapped in a <pre> > > block, where Dillo seems to have a rendering problem: Underscores `_' > > are not drawn. I stripped down this real-world-example to the > > attached minimal html file which reproduces the problem: > > > > The page should render as follows: > > > > ============================================ > > softpoint_with_block() { > > { echo ""; } | ./softpoint.x leshouches > > } > > ============================================ > > > > However, in the development version of Dillo it renders as follows: > > > > ============================================ > > softpoint with block() { > > { echo ""; } | ./softpoint.x leshouches > > } > > ============================================ > > > > As you can see, the underscores in the function softpoint_with_block() > > are not drawn in Dillo. > > > > However, the underscores are drawn correcty if you > > > > * remove the style="..." specification from the <pre> block or > > > > * remove the line > > > > { echo ""; } | ./softpoint.x leshouches > > > > from the <pre> block content. > > The attached example renders OK here. > Maybe someone else can reproduce it. > > -- > Cheers > Jorge.- > > _______________________________________________ > Dillo-dev mailing list > Dillo-dev at dillo.org > http://lists.dillo.org/cgi-bin/mailman/listinfo/dillo-dev

July 22, 2016

Underscores not drawn in pre with background color

by jcid＠dillo.org

Hi Alexander, On Thu, Jul 21, 2016 at 09:50:22PM +0200, Alexander Voigt wrote: > Hi Dillo developers, > > I stumbled upon a page with a lot of code snippets wrapped in a <pre> > block, where Dillo seems to have a rendering problem: Underscores `_' > are not drawn. I stripped down this real-world-example to the > attached minimal html file which reproduces the problem: > > The page should render as follows: > > ============================================ > softpoint_with_block() { > { echo ""; } | ./softpoint.x leshouches > } > ============================================ > > However, in the development version of Dillo it renders as follows: > > ============================================ > softpoint with block() { > { echo ""; } | ./softpoint.x leshouches > } > ============================================ > > As you can see, the underscores in the function softpoint_with_block() > are not drawn in Dillo. > > However, the underscores are drawn correcty if you > > * remove the style="..." specification from the <pre> block or > > * remove the line > > { echo ""; } | ./softpoint.x leshouches > > from the <pre> block content. The attached example renders OK here. Maybe someone else can reproduce it. -- Cheers Jorge.-

July 22, 2016

Underscores not drawn in pre with background color

by Hole.destructor＠gmx.de

Hi Dillo developers, I stumbled upon a page with a lot of code snippets wrapped in a <pre> block, where Dillo seems to have a rendering problem: Underscores `_' are not drawn. I stripped down this real-world-example to the attached minimal html file which reproduces the problem: The page should render as follows: ============================================ softpoint_with_block() { { echo ""; } | ./softpoint.x leshouches } ============================================ However, in the development version of Dillo it renders as follows: ============================================ softpoint with block() { { echo ""; } | ./softpoint.x leshouches } ============================================ As you can see, the underscores in the function softpoint_with_block() are not drawn in Dillo. However, the underscores are drawn correcty if you * remove the style="..." specification from the <pre> block or * remove the line { echo ""; } | ./softpoint.x leshouches from the <pre> block content. Best regards, Alex

July 21, 2016

ImageMagick for SVG and other image types?: Pseudo Code

by jgaffney＠lawrenceville.org

I was going to attempt a sort of proof-of-concept patch, and still include a mix of code and pseudo-code below, but see that a change would be required in the way a_Misc_get_content_type_from_data() (in src/misc.[hc]) is called from elsewhere in Dillo -- but this would not require much change, since, as far as I can see, a_Misc_get_content_type_from_data() is called from only two places in the rest of the source code. The problem is that the data being parsed is from a Dstr, but only the str and the value of the len members of this Dstr are passed to a_Misc_get_content_type_from_data(), so that the Dstr is not modifiable. If a pointer to the Dstr were passed instead, it would be modifiable, which would then allow for conversion of image types other than GIF, PNG, and JPG to be converted to PNG for the purposes of display by replacing the original Dstr data with the data for the converted image. (This would mean that Dillo would be using and caching the PNG into which the original image was converted, rather than the original, but this is only proposed as an interim means of at least displaying other, currently unsupported, image types.) In a mix of code and pseudo-code: Along with the configure.ac block to use the Magick++-config utility to test for the Magick++ library and its version and to set the corresponding compiler and linker flags, and a call in main() to Magick::InitializeMagick(NULL), replace the current src/misc.c block starting at line 161 } else if (Size >= 2 && !strncmp(p, "\xff\xd8", 2)) { /* JPEG has the first 2 bytes set to 0xffd8 in BigEndian - looking * at the character representation should be machine independent. */ Type = DT_IMAGE_JPG; st = 0; /* Text */ } else { /* Heuristic for "text/plain" ... */ with something like } else if (Size >= 2 && !strncmp(p, "\xff\xd8", 2)) { /* JPEG has the first 2 bytes set to 0xffd8 in BigEndian - looking * at the character representation should be machine independent. */ Type = DT_IMAGE_JPG; st = 0; } else if (ConvertImage(ModifiablePtrToData, & DataLen)) { Type = DT_IMAGE_PNG; st = 0; /* Text */ } else { /* Heuristic for "text/plain" ... */ and adding to src/misc.h #include "miscconv.hh" with two new files: a header miscconv.hh with #ifndef __MISCCONV_HH__ #define __MISCCONV_HH__ #ifdef __cplusplus extern "C" { #endif #include <stdlib.h> int ConvertImage (void *, size_t *); #ifdef __cplusplus } #endif #endif and, for the source file miscconv.cc, something like #include "miscconv.hh" #include <Magick++.h> extern "C" int ConvertImage ( char * Data, size_t * Size ) { try { Magick :: Blob blob ( Data, Size ), PngBlob; // This ctor will throw an exception if Data is not image data: Magick :: Image image ( blob ); // From the magic used by the `file' command, it seems that identifying SVG // data by its magic would be pretty involved, and as far as I can tell // there is no way in Magick++ to check the type of the image data // (including whether it is image data to begin with) without constructing // an Image -- which, if the data is any kind of image recognized by // ImageMagick, seems automatically to convert the image to some internal // format. And in consequence it seems that any security policies, such as // limits on image sizes or types, would have to be set via the local // ImageMagick policy (in the ImageMagick policy.xml in /etc or the like). // Now that the Image has been constructed, if desired its native type can be // determined by, e.g., Image::format(). Otherwise, one can // just proceed with the conversion of images of any format. // The following call could also throw an exception: image . write ( & PngBlob, "PNG"); // And now we just need to rewire things for the converted image: Code here to deallocate Data, allocate a new pointer according to PngBlob.length(), replace Data with the new pointer, and copy PngBlob.data() to it. } catch ( Magick :: Exception & e ) { // The ImageMagick docs cover the rather wide array of possible warnings and // errors, and a MSG could be crafted from these, but for the present purposes // any exception means the failure to return a displayable PNG. Code to do whatever with e.what() ... return 0; } return 1; } John Gaffney <jgaffney at lawrenceville.org> writes: > At this point Dillo supports the display of GIF, PNG, and JPG, but SVG > is used heavily by sites like, for example, Wikipedia. Has the > Magick++ library in ImageMagick been considered for supporting these > other image formats? The FLTK people seem enthusiastic about the > Imago library in spite of its alpha state of development; librsvg, > which looks, at least to my untutored eye, like it requires a fair > amount of voodoo to use, is apparently no longer actively supported > and would at any rate only provide SVG support; but ImageMagick is > pretty standard with Linux distributions, and the Magick++ interface > makes conversions from any format to PNG for the purposes of display > extremely easy. In src/misc.c, at the end of the image part of the > else-if ladder in a_Misc_get_content_type_from_data(), one could add > another "rung" that uses Data and Size to create a Magick::Blob, and > from this a Magick::Image, then use Magick::Image.write() to write > this to another Magick::Blob of type PNG, and then replace Data and > Size with those of the resulting Magick::Blob. A try-catch block > would catch the cases where Data is not an image. Altogether this > sort of "cheat" would require very few lines of code to implement. > And the Magick++-config utility could supply the needed -D..., -I..., > -l... flags to the configure script. > > I'm not suggesting this is an ideal long-term solution, but in the > interim would at least allow the display of other image types. > > The most recent mention of this on the mailing list seems to be > Jorge's indication on 22 April 2016 that > > No, there's no SVG image support in dillo yet. > > It should not be hard to implement though, but it needs an effort > similar to GIF or PNG or JPEG code. Patch welcomed! ;) > > But surely someone has thought of ImageMagick before. Are there > objections to using it? And if so, maybe a flag could be set by the > configure script to conditionally compile this in for those who want > it? (If this is pursued, the configure script should probably also > check that the version of ImageMagick is recent enough to have been > patched against that serious exploit in just this sort of conversion > code.) > > A test program shows that the invocation of ImageMagick and > conversions really are as simple as the steps outlined above. But I > would be happy to make an attempt at a patch, if you don't mind this > being undertaken by an amateur. > > > _______________________________________________ > Dillo-dev mailing list > Dillo-dev at dillo.org > http://lists.dillo.org/cgi-bin/mailman/listinfo/dillo-dev

July 20, 2016