Community
Participate
Working Groups
As I was fixing bug 491304 some other enhancements occurred to me, but since they are relatively minor thought I would just document them for now, instead of continuing to work on the webpages. One item I'd like is more a "totals" row in the test summary. This is admittedly mostly to provide a confirmation that _our_ summary agrees with Huson's summary (on first page of a build). Some of the columns with numbers would look better if they were -- still centered -- but right aligned as numbers. From what I'd read about CSS this is not easy to do. Such as, see http://stackoverflow.com/questions/13992026/how-to-center-right-aligned-numbers-in-table-column-using-css I am not sure, but some of the libraries used by Solstice might provide that functionality. Most minor still, in a perfect world, the Java code in our Ant task would use no "style" attributes, but a few were left for some of the complicated cases.
Another thing I've always thought would be useful is to have a list of all "compiler summaries" available. Currently, there is only a link to the "raw summary" if there is a problem. But, occasionally someone might want to look at raw summary to confirm the version of the compiler used or perhaps to confirm exactly how something was compiled by looking at the @dot.xml file. (Hey, I did say "minor" enhancements :)
Not sure this is a "fix now", or "fix in future" issue, but I think the "blue bars" used for the heading is too dark. I would be better to match what is in the left nav bar since that would make the words easier to read, plus, would make it more obvious the "nav bar" items "went with" the headings.
Via local testing, I've noticed when there are "missing files", there are a few problems. </table>↩ </table>↩ <table><tr> <th class='cell'>Missing Files</th></tr>↩ <tr><td class='namecell'>org.eclipse.jdt.core.tests.compiler_ep46N-unit-cen64_linux.gtk.x86_64_8.0.xml</td></tr>↩ <tr><td class='namecell'>org.eclipse.jdt.core.tests.model_ep46N-unit-cen64_linux.gtk.x86_64_8.0.xml</td></tr>↩ <tr><td class='namecell'>org.eclipse.jdt.ui.tests_ep46N-unit-cen64_linux.gtk.x86_64_8.0.xml</td></tr>↩ <tr><td class='namecell'>org.eclipse.jdt.ui.tests.refactoring_ep46N-unit-cen64_linux.gtk.x86_64_8.0.xml</td></tr>↩ <tr><td class='namecell'>org.eclipse.pde.build.tests_ep46N-unit-cen64_linux.gtk.x86_64_8.0.xml</td></tr>↩ </table> A. an extra "end table" gets inserted. B. the table/cells ends up being pretty narrow, causing the "long file names names" to be wrapped, which makes them hard to read. C. if the file is completely missing then there is no row for it created in the main table. (we used to create the row, and put a "-1" in the cell). D. Should probably and a legend, or at least a 'hover', over what the various negative numbers means, such as "-3" == "the file existed, but the XML elements were not as expected" (Guess that doesn't mean much either).