We have a convenient BufferSizeFromStrideAndHeight utility function that takes stride (usually width * byte count) and height. Add a utility function to take all three.

Comment on attachment 8786455 [details] Bug 1299164: Part 1. Add BufferSizeFromDimensions method to complement BufferSizeFromStrideAndHeight. https://reviewboard.mozilla.org/r/75406/#review73564 ::: gfx/2d/DataSurfaceHelpers.cpp:268 (Diff revision 1) > size_t > BufferSizeFromStrideAndHeight(int32_t aStride, > int32_t aHeight, > int32_t aExtraBytes) > { > - if (MOZ_UNLIKELY(aHeight <= 0)) { > + return BufferSizeFromDimensions(aStride, aHeight, 1, aExtraBytes); Not a huge fan of passing stride to height. Maybe we should name it stride in BufferSizeFromDimension as well? I'll leave this up to you to decide.

Comment on attachment 8786456 [details] Bug 1299164: Part 2. Use BufferSizeFromDimensions method, as well as some of the others that check for valid size. https://reviewboard.mozilla.org/r/75408/#review73566

Comment on attachment 8786455 [details] Bug 1299164: Part 1. Add BufferSizeFromDimensions method to complement BufferSizeFromStrideAndHeight. https://reviewboard.mozilla.org/r/75406/#review73564 > Not a huge fan of passing stride to height. Maybe we should name it stride in BufferSizeFromDimension as well? I'll leave this up to you to decide. I agree, the readability is important. Given that it was suffering with trying to save a few lines of "duplicated" code, and it led to an extra function call, I separated the two functions - BufferSizeFromStrideAndHeight does not call BufferSizeFromDimensions anymore. The problem above goes away and things are a bit cleaner. I don't anticipate big changes to these functions, but if we make them to one, we should make them to both.

Pushed by msreckovic@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/c39ec15e7e21 Part 1. Add BufferSizeFromDimensions method to complement BufferSizeFromStrideAndHeight. r=bas https://hg.mozilla.org/integration/autoland/rev/53dc795121e1 Part 2. Use BufferSizeFromDimensions method, as well as some of the others that check for valid size. r=bas

I had to back this out because it appears to have caused failures like https://treeherder.mozilla.org/logviewer.html#?job_id=2981925&repo=autoland https://hg.mozilla.org/integration/autoland/rev/aefced02ac1e

Thanks for catching this and backing it out. I imagine it's for disallowing negative stride, but I'll double check before try/inbounding this again.

It's for Factory::CreateDrawTargetForCairoSurface returning nullptr when AllowedSurfaceSize fails (on 45680x65360 size.) I guess we were allowing that to succeed before.

In the context of print

Comment on attachment 8786456 [details] Bug 1299164: Part 2. Use BufferSizeFromDimensions method, as well as some of the others that check for valid size. https://reviewboard.mozilla.org/r/75408/#review77416

Pushed by msreckovic@mozilla.com: https://hg.mozilla.org/integration/mozilla-inbound/rev/e2d722f565d3 Part 1. Add BufferSizeFromDimensions method to complement BufferSizeFromStrideAndHeight. r=bas https://hg.mozilla.org/integration/mozilla-inbound/rev/1aecdb3b9974 Part 2. Use BufferSizeFromDimensions method, as well as some of the others that check for valid size. r=bas

https://hg.mozilla.org/mozilla-central/rev/e2d722f565d3 https://hg.mozilla.org/mozilla-central/rev/1aecdb3b9974