SILE 0.14.8 (from the siletypesetter/sile:v0.14.8 docker image, to ensure my own build of current master wasn't causing the issue), but also reproduced with earlier versions (tested with siletypesetter/sile:v0.12.5)
Pretty minor issue, since there's a workaround anyway (see below), but I am not sure whether it's caused by some rounding issue or if there's something more tricky at play.
Scenario
The content in a "center", "raggedleft" or "raggedright" environment is exactly 100%lw. E.g. I observed it with a full-line table, but it would also be true with an image set at that width.
In the MWE below I am just using rules to show the issue in a simple way.
Vertical spacing isn't expected to differ from the justified case.
Observed result
Some glues or (zero?)boxes are pushed in their own line, leading to additional vertical spaces.
A workaround is to set the width to a smaller value. I was drastic above with the 99.9999...., even a mere 99.9% is usually sufficient and remains hardly visible.
And for the record, it turned out I had seen the issue already last month with an hrule too, though I implemented a quick workaround, late at night... and later forgot about it ;)
SILE 0.14.8 (from the
siletypesetter/sile:v0.14.8
docker image, to ensure my own build of current master wasn't causing the issue), but also reproduced with earlier versions (tested withsiletypesetter/sile:v0.12.5
)Pretty minor issue, since there's a workaround anyway (see below), but I am not sure whether it's caused by some rounding issue or if there's something more tricky at play.
Scenario
The content in a "center", "raggedleft" or "raggedright" environment is exactly 100%lw. E.g. I observed it with a full-line table, but it would also be true with an image set at that width. In the MWE below I am just using rules to show the issue in a simple way.
Expected result
Observed result
A workaround is to set the width to a smaller value. I was drastic above with the 99.9999...., even a mere 99.9% is usually sufficient and remains hardly visible.