21:22:00 <tomman> has anyone noticed some weird rendering issues with tables?
21:22:03 <tomman> here is a case: https://imgur.com/a/um8rVIM
21:22:12 <tomman> table is this: https://en.wikipedia.org/wiki/Intel_Graphics_Technology#Xe_LP_Architecture_%28Gen12%29
21:22:41 <tomman> notice that how in SeaMonkey, some of the cells spanning multiple rows still have borders inside the cells, although text is properly positioned
21:23:01 <tomman> (in fact it seems to happen in many tables at Wikipedia)
21:23:09 <tomman> on FF91ESR it renders fine
21:24:16 <tomman> FWIW, it also renders fine on good ol' 2.49.5 on XP :)
21:25:42 <tomman> --borders inside the merged cells
21:41:40 <njsg> tomman: even before opening the image, I think yes.
21:42:09 <njsg> on mediawiki at wikipedia, so same-ish site
21:44:14 <njsg> tomman: second column from the left, all but last? (well, but last and first)
21:47:34 <njsg> interesting, happens in Composer too with the red lines
22:01:17 <tomman> dunno when it started happening
22:01:35 <tomman> but have only noticed at Wikipedia (need to check other MediaWiki sites)
22:02:19 <njsg> bug 1755218?
22:02:54 <njsg> (as it goes away if I disable "border-collapse: collapse" on wikipedia)
22:04:43 <njsg> as in the attached screenshot, I get only a "partial" border with that tournament page, not the "full" burder, but after I select and deselect text around that border, I get the "full" border.
22:07:32 <tomman> reminds me of a VERY similar bug I had with a PDF generator library with border-collapse: collapse
22:07:46 <tomman> interesting enough it does not happen on SM2.49 or FF91ESR
22:08:00 <tomman> and that bug report sends me to bug 332740
22:08:04 <tomman> which is 16 years old!
22:10:19 <njsg> bug 1755218 comment 15
22:10:43 <njsg> 1485179 appears to be in?
22:19:51 <tomman> so basically border-collapse:collapse is a minefield for corner cases, except that everything turns out to be a corner case at any time?
22:20:34 <tomman> and given my experience on similar issues with other rendering engines (Flying Saucer/openhtmltopdf), seems to be hard
22:29:09 * njsg needs to procure some sleep
23:15:31 <hardys> tomman: the table is rendered fine in 2.53.9 and 2.57
23:37:07 <hardys> tomman: renders fine in 2.53.10 and 2.53.11, issue is  present in 2.53.12
23:37:37 <tomman> hardys: curious indeed, thanks
23:38:06 <tomman> pretty much approaches the timeline on when I started noticing the issue