Please do not reply directly to this email. All additional comments should be made in the comments box of this bug report.
Summary: Review Request: xtide - Calculate tide all over the world
https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=211626
------- Additional Comments From pertusus@free.fr 2006-10-30 07:35 EST ------- (In reply to comment #67)
- tcd-utils is a separate tarball, a small one, but needed for building
your "xtide" rpms. So, you decided to include it, thereby defeating the purpose of the upstream split. Among other side-effects, you need to push new xtide, libtcd, libtcd-devel, xttpd, tcd-utils packages whenever one out of two upstream packages changes. Further, tideEditor, which is part of tcd-utils and not put into a separate rpm, adds a dependency on the rather large Qt. This reduces the benefit of a separate tcd-utils rpm, as now Qt is needed when installing the small *tide_db tools.
I my opinion this is the result of a broken upstream split, since upstream should have put tideEditor with xtide, and not with the *tide_db tools. Maybe this comes from historical reasons (because the original tcd-utils author isn't the xtide author), and not from technical reasons.
- And finally, XTide, split into multiple small rpms, is "enhanced" with
a strict dependency on a 40M data rpm.
Without wvs, xtide or tideEditor are ugly. Admitedly the 2 high resolution wvs data files are certainly not very usefull for xtide, but removing them from wvs adds complexity.
The current package split is neither obvious nor convenient for the users. Seriously, considering the special target group of XTide and friends, I wouldn't even mind an all-in-one package with *optional* data.
wvs cannot really be optional, there are no coast boundaries without it.
I don't really like having xttpd installed together with the graphical tools, but I wouldn't object a all-in-one package (or with only libtcd, libtcd-devel subpackages), since indeed it is more convenient.