Fix .pdf gotcha #227
Fix .pdf gotcha #227
Conversation
zbarimg/zbarimg.c
Outdated
| static int scan_image(const char *filename) | ||
| { | ||
| if (exit_code == 3) | ||
| return (-1); | ||
|
|
||
| int found = 0; | ||
| MagickWand *images = NewMagickWand(); | ||
| if (file_extension(filename, ".pdf")) MagickSetResolution(images, 300, 300); |
There was a problem hiding this comment.
Maybe replace the file_extension check with one based on MagickGetImageFormat.
There was a problem hiding this comment.
This is a very sensible suggestion, but SetResolution only works before the images are loaded, and GetImageFormat only works after the images are loaded.
I'm open to suggestions on how to do it anyway--I'm pretty much an ImageMagick newbie. Seems like it would be nice to have it work for stdin, for example.
There was a problem hiding this comment.
(And before you ask, if you just unconditionally set the resolution, it overrides the DPI settings in file formats that actually do set it internally)
There was a problem hiding this comment.
Hmm, I think you use MagickPingImage to get the format before doing the full load.
There was a problem hiding this comment.
Maybe need to remove the image created by ping image using something like MagickRemoveImage before the full load.
There was a problem hiding this comment.
Seems like it would be nice to have it work for stdin, for example.
For stdin I think you could use MagickReadImageBlob and MagickPingImageBlob for reading/pinging a stdin buffer.
There was a problem hiding this comment.
I've made the change for a single image. For stdin I don't think this is possible, because you can't seek back to the beginning of stdin--once it's consumed, it's consumed, in general.
There was a problem hiding this comment.
For stdin I don't think this is possible, because you can't seek back to the beginning of stdin--once it's consumed, it's consumed, in general.
I think you would read stdin until EOF into a dynamically allocated char buffer then pass the buffer to imagemagic using the blob read functions(which shouldn't remove that buffer and thus allow reuse).
Anyways that's mostly separate from this fix at least but may be a nice follow up.
There was a problem hiding this comment.
Actually, scan_image is used for "-" (which I realized as I started said followup). Currently this PR breaks all zbarimg use with stdin. I'll go ahead and figure out a fix. I'll also look into whether there's a missing test--make check shows no issues.
|
https://stackoverflow.com/questions/73656471/use-filetoblob-in-imagemagick well i'm a bit stuck using FileToBlob() if you can offer any help. |
Maybe should be something like this: void *blob = FileToBlob(filename, MagickMaxBufferExtent, &blob_length, &blob_exception); |
|
or maybe: void *blob = FileToBlob(filename, MAGICK_SSIZE_MAX, &blob_length, &blob_exception); |
|
The second argument (extent) is not causing the issues. I started it as
INT_MAX, sure, just left that out for clarity.
|
|
Proposed an ImageMagick feature to help us out. If someone has thoughts on a better way to do it, feel free to comment. ImageMagick/ImageMagick#5554 Also, went ahead and reported my issues as a bug at ImageMagick/ImageMagick#5553. I can get it to work for an 80K file but not a 100K file (note size is like 100MB), which makes me think MAYBE it's not just a me issue. In the meantime, I've made sure this doesn't break stdin, and I think it could be merged. It just won't work for PDFs on stdin yet. |
|
Confirmed the issue I was running into was a bug. ImageMagick has a patch, thanks ImageMagick! I'll offer a PR that makes PDFs work on stdin too, once imagemagick's stable release includes their fix. |
|
OK, ready for submission. Apparently I was wildly wrong about how setting the density works, and I can just do it unconditionally. This is now a one-line change. |
|
Ping on this? It's a one-line change now so it seems pretty safe. You could change the default resolution to 300 or 600 if you prefer. |
|
Merged, thanks! |
PDF does not inherently encode a resolution--it's essentially a vector format. By default, ImageMagick samples it at a tiny 72dpi resolution, so many QR codes are not read.
Default .pdf to 300dpi sample rate to fix the issue.