Let PDF::Core determine precision of float rounding (currently 4 decimal places) #782
Conversation
Update how_to_read_this_manual.rb
practicingruby
force-pushed
the
decimal_rounding
branch
from
491a90c to
21e1624
Compare
|
Note to self for CHANGELOG: This pull request will not break external behavior, but it will change how PDFs are represented at the binary level. We don't promise binary compatibility between releases, but a note to warn people who do their testing that way is in order. |
|
We had decent test coverage on the Prawn side of this, so this patch can probably be merged after prawnpdf/pdf-core#10 has been tidied up. That said, it's a scary change with global effects. @packetmonkey can you at least do a non-trivial test rendering using this branch against one of your projects? |
practicingruby
changed the title
|
Sure, I'll try to get that test run today or monday depending on meetings. I have a few fairly complex dynamically calculated products (using a lot of dry runs) that would hopefully ferret out any major problems. |
|
I just compared a fairly complex document that uses a lot of dry-runs to calculate heights and dynamically position elements on the page. After eyeballing it and also using xscope to independently check the measurements on screen, everything looks identical to me on the decimal rounding branches. I asked someone else spot check it as well and she found nothing wrong either. My tests also continued to pass which includes tests to make sure that the dimensions of various elements are calculated correctly, which shouldn't be (and weren't) impacted by this, but it was comforting to continue to see none the less. |
|
Hooray! We will merge this week, then. |
practicingruby
force-pushed
the
master
branch
from
f5e08bd to
6848965
Compare
|
Squashed and merged in 0249316 |
This attempts to integrate prawnpdf/pdf-core#10.