Speaking of GSoC work needing polishing...
It has recently come to my attention that the tref elements are not working. To be honest, I am very surprised because the last time I looked at this, which was when I committed the work 6+ months ago, it was working. Not sure what happened or when.
I think it is very important to fix this for the release, as I had entered release notes and provided samples way back when. I am looking at this over the next few days and will submit a patch that could be tested and hopefully submitted to the release branch.
I was just wondering - is there any way to know when it stopped working? For example, are the SVG compliance test results archived, and if so where? It should have passed sometime in the summer and stopped since then (if it never passed then I'll be really worried!).
Thanks! Gail
On Mon, Feb 25, 2008 at 11:34 AM, Gail Carmichael <gail.banaszkiewicz@...400...> wrote:
I was just wondering - is there any way to know when it stopped working?
Usually this is found by bisection: if at date X it worked and at Y it didn't, check at date (X+Y)/2; if it works then, make that date the new Y, otherwise the new X, and repeat. This converges pretty fast, although you will still spend many hours recompiling because our codebase changes very intensively.
automated tests attached to svn commit hook would be much better
Does anybody here know how much effort would be needed in order to establish this practice?
Juca
On Mon, Feb 25, 2008 at 6:18 PM, Felipe Sanches <felipe.sanches@...400...> wrote:
automated tests attached to svn commit hook would be much better
Does anybody here know how much effort would be needed in order to establish this practice?
I like the idea in principle, but I have no idea if this is possible with our SVN host (sourceforge).
Even if it was just a regularly scheduled process on seperate site it would be helpful, if it could narrow the search field down to a day or 2 to start with. Could be done much the same way the automated win builds that have been done before.
On Mon, Feb 25, 2008 at 6:04 PM, bulia byak <buliabyak@...400...> wrote:
On Mon, Feb 25, 2008 at 6:18 PM, Felipe Sanches <felipe.sanches@...400...> wrote:
automated tests attached to svn commit hook would be much better
Does anybody here know how much effort would be needed in order to establish this practice?
I like the idea in principle, but I have no idea if this is possible with our SVN host (sourceforge).
-- bulia byak Inkscape. Draw Freely. http://www.inkscape.org
This SF.net email is sponsored by: Microsoft Defy all challenges. Microsoft(R) Visual Studio 2008. http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/ _______________________________________________ Inkscape-devel mailing list Inkscape-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/inkscape-devel
Felipe Sanches wrote:
automated tests attached to svn commit hook would be much better
Does anybody here know how much effort would be needed in order to establish this practice?
Juca
You could write a script to tie a test to the receipt of an inkscape-cvs@...156... message, I guess.
bob
On Feb 25, 2008, at 5:42 PM, Bob Jamison wrote:
Felipe Sanches wrote:
automated tests attached to svn commit hook would be much better
Does anybody here know how much effort would be needed in order to establish this practice?
Juca
You could write a script to tie a test to the receipt of an inkscape-cvs@...156... message, I guess.
Or even a low-tech approach with a cron job firing off "svn status - u" every once in a while.
Of course, the trick is to not trigger a second checkout-build-test run while once is still in progress.
Mon, 25 Feb 2008 19:18:01 -0300 "Felipe Sanches" <felipe.sanches@...400...> kirjoitti:
automated tests attached to svn commit hook would be much better
Does anybody here know how much effort would be needed in order to establish this practice?
I'd say: lots. Automated builds are easy (even though tying them to svn commit isn't possible on SourceForge). The big thing is: running automated tests is of very little use, if we don't have a wide array of tests. The make check we have now is nice, but it tests only a small area of Inkscape codebase.
participants (7)
-
Bob Jamison -
bulia byak -
Felipe Sanches -
Gail Carmichael -
john cliff -
Jon A. Cruz -
Niko Kiirala