Closed winhamwr closed 7 years ago
How does the spec say we should be handling these?
How does the spec say we should be handling these?
I didn't read the spec.
Why not? How then do you know that we are actually "correctly" handling margin positions? Simply not erroring is not sufficient.
Why not? How then do you know that we are actually "correctly" handling margin positions? Simply not erroring is not sufficient.
I can switch the update note to say that we are no longer erroring.
I can switch the update note to say that we are no longer erroring.
That is preferred over the current note.
How does the spec say we should be handling these?
What's the website you gentlemen use to browse the spec? Our documentation only links to the spec itself, which is pretty horrible reading. It would be great to have a link to that spec browser site in the contributtion docs.
I didn't read the spec.
I have the same question as Kyle. Why not? Pydocx is based on the spec. If the spec says to ignore non-conforming attributes, rather than truncate them, that's just as easy for us to do.
What's the website you gentlemen use to browse the spec? Our documentation only links to the spec itself, which is pretty horrible reading. It would be great to have a link to that spec browser site in the contributtion docs.
Honestly, I don't normally read the spec (for better or for worse).
I have the same question as Kyle. Why not? Pydocx is based on the spec. If the spec says to ignore non-conforming attributes, rather than truncate them, that's just as easy for us to do.
Because I don't want to spend an hour to find and figure out what the spec says to maybe save one pixel.
What's the website you gentlemen use to browse the spec? Our documentation only links to the spec itself, which is pretty horrible reading.
I don't use any website. I always go to the PDF. It's the most authoritative source I know of, and it is very detailed. I don't think I've ever read a spec or reference that wasn't horrible reading.
Because I don't want to spend an hour to find and figure out what the spec says to maybe save one pixel.
The job of pydocx is to do that so that every future user doesn't have to worry about things like this. Have you tried and failed to find what you're looking for or just not tried?
The job of pydocx is to do that so that every future user doesn't have to worry about things like this. Have you tried and failed to find what you're looking for or just not tried?
I haven't tried. If Kyle could link me to the PDF, I will timebox looking for this for an hour.
It's linked here: https://pydocx.readthedocs.io/en/latest/conformance.html
You'll need to download Part 1 of the fourth edition. The PDF is contained within the zip.
I will timebox looking for this for an hour.
That seems like a good timebox use, to me.
This tells me that if this number is off by one, it'll be off by less 1/1440th of an inch. Which doesn't seem significant to me. Aside from that, I can't really tell if this should end up being an int
or a float
Looks good!
@kylegibson care to weigh in? You are much more familiar with reading and understanding the spec for this.
it's a pity that in github we do not see the release 0.9.10 which has solved this bug.
@Simplici - Just created a release, thank you! https://github.com/CenterForOpenScience/pydocx/releases
Some .docx files that we're seeing in the wild are using a non-integer value for the
indentation_hanging
attribute. It looks something like:<ind hanging="504.00000000000006">
This causes a crash when we try to cast to int.
I'm not sure if these documents are violating the spec, but we've seen 123 cases of this in the last 8 days.
Traceback