Re: [Sugar-devel] Sugarbot working on latest Sugar
Excerpts from Tim McNamara's message of Sun Aug 08 02:06:21 +0200 2010: Is there interest in pushing an updated sugarbot to git.sl.o? Sure, go ahead! If you need commit rights, please write to systems@ and I'll transfer the project to you. The original author, a GSoC student, has already stated having no interest to work on it any further. I have spent some time over the last three days getting Sugarbot to run in sugar-jhbuild. After a few code changes, and lots of reading, I have been successful. Great! And thanks for working on this. Automated UI testing is our most pressing need right now (IMO at least). It enables faster, easier and better testing of code changes (including upstream changes). Think... Sugarbot's widget identification system worked well when text-based labels were common place. Recent Sugar versions have moved to a highly icon-based naming system. Developers will either need to make use of gtk.Widget.set_name() or another widget identification algorithm will need to be established. There are several things under the hood that I think would be useful. I think some of what I would like to do would be classed as code style changes, rather than substantive though. I don't have time to answer more in-depth, but one thing I'd like to point out is that Gnome is using dogtail for UI testing which seems to be rather similar to SugarBot so we might want to use that instead. IIRC it even comes with some tools to make writing tests easier (a macro recorder thingie and a widget tree exploration GUI). As much as I'd hate to see a GSoC project being never put into real use, the advantage of having someone else maintain the code is rather important for now. But maybe we can reuse some code or ideas from SugarBot to make it easier to write test cases for activities? Sascha -- http://sascha.silbe.org/ http://www.infra-silbe.de/ signature.asc Description: PGP signature ___ Sugar-devel mailing list Sugar-devel@lists.sugarlabs.org http://lists.sugarlabs.org/listinfo/sugar-devel
Re: [Sugar-devel] Sugarbot working on latest Sugar
On Wed, Aug 11, 2010 at 17:17, Sascha Silbe sascha-ml-ui-sugar-de...@silbe.org wrote: Excerpts from Tim McNamara's message of Sun Aug 08 02:06:21 +0200 2010: Is there interest in pushing an updated sugarbot to git.sl.o? Sure, go ahead! If you need commit rights, please write to systems@ and I'll transfer the project to you. The original author, a GSoC student, has already stated having no interest to work on it any further. I have spent some time over the last three days getting Sugarbot to run in sugar-jhbuild. After a few code changes, and lots of reading, I have been successful. Great! And thanks for working on this. Automated UI testing is our most pressing need right now (IMO at least). It enables faster, easier and better testing of code changes (including upstream changes). Think... Sugarbot's widget identification system worked well when text-based labels were common place. Recent Sugar versions have moved to a highly icon-based naming system. Developers will either need to make use of gtk.Widget.set_name() or another widget identification algorithm will need to be established. There are several things under the hood that I think would be useful. I think some of what I would like to do would be classed as code style changes, rather than substantive though. I don't have time to answer more in-depth, but one thing I'd like to point out is that Gnome is using dogtail for UI testing which seems to be rather similar to SugarBot so we might want to use that instead. IIRC it even comes with some tools to make writing tests easier (a macro recorder thingie and a widget tree exploration GUI). As much as I'd hate to see a GSoC project being never put into real use, the advantage of having someone else maintain the code is rather important for now. But maybe we can reuse some code or ideas from SugarBot to make it easier to write test cases for activities? The biggest difference between dogtail and sugarbot is how they discover the widgets. Dogtail uses AT-SPI and sugarbot uses the introspection features in Python. We anyway need to make Sugar work under AT-SPI because of accessibility so would be good to see movement in that area. Regards, Tomeu Sascha -- http://sascha.silbe.org/ http://www.infra-silbe.de/ ___ Sugar-devel mailing list Sugar-devel@lists.sugarlabs.org http://lists.sugarlabs.org/listinfo/sugar-devel ___ Sugar-devel mailing list Sugar-devel@lists.sugarlabs.org http://lists.sugarlabs.org/listinfo/sugar-devel
Re: [Sugar-devel] Sugarbot working on latest Sugar
On 12 August 2010 03:17, Sascha Silbe sascha-ml-ui-sugar-de...@silbe.orgwrote: Excerpts from Tim McNamara's message of Sun Aug 08 02:06:21 +0200 2010: Is there interest in pushing an updated sugarbot to git.sl.o? Sure, go ahead! If you need commit rights, please write to systems@ and I'll transfer the project to you. The original author, a GSoC student, has already stated having no interest to work on it any further. Sugarbot repo is currently reserved for the IRC bot. I'll think of a new name for the Sugarbot testing project start a new project from the GSoC SVN. Affination is one of the processes in raw sugar processing. What about that? I have spent some time over the last three days getting Sugarbot to run in sugar-jhbuild. After a few code changes, and lots of reading, I have been successful. Great! And thanks for working on this. Automated UI testing is our most pressing need right now (IMO at least). It enables faster, easier and better testing of code changes (including upstream changes). [snip] I don't have time to answer more in-depth, but one thing I'd like to point out is that Gnome is using dogtail for UI testing which seems to be rather similar to SugarBot so we might want to use that instead. I would like to use Sugarbot during a transition phase. Two reasons: 1) Things should work with it now. To support Dogtail (or LDTP), Sugar's components will need to support AT-STI. I don't know how hard that will be. 2) Test scripts written for Sugarbot are be almost identical to scripts written for Dogtail. Therefore, testers' time taken to build scripts for Sugarbot will not be wasted. I'm not sure how Dogtail works under the hood. The project was initially dismissed by the Sugarbot author as being too hard to look into during GSoC.[1] [1] http://gsoc-sugarbot.blogspot.com/2008/05/dissecting-guitest.html ___ Sugar-devel mailing list Sugar-devel@lists.sugarlabs.org http://lists.sugarlabs.org/listinfo/sugar-devel
[Sugar-devel] Sugarbot working on latest Sugar
Is there interest in pushing an updated sugarbot to git.sl.o? I have spent some time over the last three days getting Sugarbot to run in sugar-jhbuild. After a few code changes, and lots of reading, I have been successful. Some changes that needed to be made from the 0.1 release: - a few syntax errors - update sugarbot.py to access jarabe.model.bundleregistry.get_registry, rather than sugar.shell.registry and update associated methods Left to do: Think... Sugarbot's widget identification system worked well when text-based labels were common place. Recent Sugar versions have moved to a highly icon-based naming system. Developers will either need to make use of gtk.Widget.set_name() or another widget identification algorithm will need to be established. There are several things under the hood that I think would be useful. I think some of what I would like to do would be classed as code style changes, rather than substantive though. Thoughts questions welcome. Tim ___ Sugar-devel mailing list Sugar-devel@lists.sugarlabs.org http://lists.sugarlabs.org/listinfo/sugar-devel