Would it be feasible to collect enough data from an installation to replicate the user's environment?
A "leo.state" file would contain settings, plugins, versions. It could be used to install on a VM, or maybe parsed for known issues. On Thu, Oct 20, 2011 at 12:32 PM, Edward K. Ream <[email protected]> wrote: > > > On Oct 17, 2:41 pm, Kent Tenney <[email protected]> wrote: >> As I understand the discipline expected in the Zope community: >> >> When approaching a bug, begin by writing a test which fails >> due to the bug. Then fix the code such that the test succeeds. >> >> Patches which don't follow this convention are not considered. >> (code without tests is assumed buggy) > > I have been thinking about this comment ever since, especially as I > fix bugs, or attempt to fix bugs, without unit tests :-) > > The problem, as I see it now, is that it is, at present, impossible to > run most unit tests in the environment in which the bugs actually > appear. This can take several forms: > > 1. In the latest instance, there would have been no way to replicate > the bug because the bug had already been fixed. This is just > infuriating. Happily, this situation has been rare. > > 2. Many (most?) bugs nowadays depend on a specific set of plugins > being enabled, or a specific set of options being in effect. But I > run unit tests from a *particular* script, that runs unitTest.leo with > a *particular* set of options. Furthermore, running tests externally > is *more* restrictive, not less, because at present running tests > externally uses the leoBridge module, which in turn uses the nullGui. > > 3. Gui-related bugs cause particular problems for unitTest.leo, > because I don't like to pollute the output with the output from > various work-arounds to code that would otherwise put up dialogs and > thus pause the unit tests. > > In other words, gui-related unit tests often result in added code (in > the code being tested) that a) are almost pure cruft and b) typically > actively subvert the intention of the actual test! This does not seem > like progress: I would (usually) prefer to endure repeated (failed) > attempts to fix the bug rather than pollute the code. > > In short, a more general testing framework is needed. Until that > happens, we will have to muddle through with untested, and thus very- > likely buggy code. > > Edward > > -- > You received this message because you are subscribed to the Google Groups > "leo-editor" group. > To post to this group, send email to [email protected]. > To unsubscribe from this group, send email to > [email protected]. > For more options, visit this group at > http://groups.google.com/group/leo-editor?hl=en. > > -- You received this message because you are subscribed to the Google Groups "leo-editor" group. To post to this group, send email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/leo-editor?hl=en.
