I replied on the issue. Thanks for opening it, Alexander. Marcin
On Tue, Apr 12, 2022 at 5:21 PM Alexander Kriegisch < [email protected]> wrote: > I can create an issue tomorrow or so. I might even create a PR, if you > like. It might not be a great PR, given the fact that I do not know the Geb > source code well and never have contributes to the project before, but I > can give it a try. At least you have something to work with, or we have > something to discuss and, if necessary, throw away and re-do better. Would > a PR go on the current main branch or are you planning a special branch for > the work on Geb 6? > -- > Alexander Kriegisch > https://scrum-master.de > > > Marcin Erdmann schrieb am 12.04.2022 22:31 (GMT +07:00): > > But to my surprise, the Geb config is not read just once during extension >> start, but quite often, e.g. whenever the Geb test manager calls >> getBrowser in beforeTest, i.e. for each feature method in a Geb spec > > > is not completely precise. Geb config is read whenever > ConfigurationLoader.getConf() is called. It happens to be called via the > parameterless constructor of Browser, which in turn is called for the > first time getBrowser() is called on an instance of GebTestManager. > > All that aside, yes, Geb config is definitely read more often than it > needs to be. There will definitely be a performance penalty to parsing and > executing that script for every single test but this has worked like that > since forever (I started contributing to Geb in 2010 and it's how it was > back then) so changing this will be a breaking change. It will be fairly > easy to change to only load the config once per GebTestManager instance > but that would be a breaking change so would need to be done only in Geb 6 > which might be coming soon due to switching to Selenium 4 which is not > backwards compatible. If you feel like this would be worthwhile to change, > Alexander, then please open an issue in the tracker and I will schedule it > for Geb 6. > > > On Thu, Mar 31, 2022 at 10:17 AM Alexander Kriegisch < > [email protected]> wrote: > >> I am not sure you are following the Spock chat on Gitter. While talking >> to Jeremy there, I was looking into how Geb actually works. Result (I wrote >> much more there): It is a simple global Spock extension. But to my >> surprise, the Geb config is not read just once during extension start, >> but quite often, e.g. whenever the Geb test manager calls getBrowser in >> beforeTest, i.e. for each feature method in a Geb spec. By printing >> callstacks from the Geb config, this is easy to see. >> >> So my question is: What is the reason for that? Are you expecting cases >> in which the Geb config changes during a JVM session? The driver >> configurations for multiple environments are closures, so they can be >> evaluated without re-reading the file. Any system properties defined there >> should also not depend on dynamic things like time stamps, random values or >> whatever. >> >> >> Marcin Erdmann schrieb am 31.03.2022 02:45 (GMT +07:00): >> >> GebConfig.groovy is usually evaluated when >> geb.ConfigurationLoader.getConf() is executed which for example happens >> inside of the parameterless constructor for geb.Browser. If you want to >> have GebConfig.groovy evaluated early on as part of executing tests then >> one thing that comes to mind with the setup you describe is to put new >> ConfigurationLoader().conf in a static initalization block of BaseTest. >> >> On Mon, Mar 28, 2022 at 9:34 PM jc <[email protected]> wrote: >> >>> Ok, I think I know what happened. It's nothing Geb 5 changed. I worked >>> with Leonard Brünings as he helped me clean up some bad code while figuring >>> out another issue. One the lines I cleaned up was an unused line in the >>> BaseTest: static Browser browse = new Browser() >>> >>> With this line in place, even though it is unused anywhere, it seems as >>> thought the env variable set in GebConfig are evaluated in time. Without >>> this line, the GebConfig variables are not evaluated in time. I don't >>> fully understand it. I definitely want to have good code (I am aware >>> setting env variables in GebConfig is "bad", but it's working for us). So >>> if there is another way we can achieve these @IgnoreIf method getting >>> evaluated at the right time I am all for it. Any insights? >>> On Monday, March 28, 2022 at 1:47:48 PM UTC-5 jc wrote: >>> >>>> I am currently in the process of trying to update our project's from >>>> 4.0 to 5.1. Currently on 4.0, we have a BaseTest class that extends >>>> GebReportingSpec and then all other test classes extend BaseTest. In >>>> GebConfig.groovy, I have some lines that set different environment >>>> variables such as "runOnGrid" which is true/false, "browserType" which is a >>>> string, "env" which is the environment, etc.... >>>> >>>> We have some @IgnoreIf statement in our code to prevent running on >>>> certain environments like Production. For example, we have an entire test >>>> class "TestClassA" that can run on pre-prod but we want to ignore the whole >>>> class for Prod. We put an @IgnoreIf annotation on the class that looks at >>>> `System.getProperty("env") == "prod"` >>>> >>>> This is currently working beautifully in 4.0. However in 5.1 this is >>>> no longer working as `System.getProperty("env")` is coming back as >>>> null along with all of the other variables we have set. These variables >>>> are set by the time we hit the setupSpec() in BaseTest, but they are null >>>> prior to that when they were not before. So it seems like before 5.0, >>>> GebConfig.groovy was getting run earlier is the compilation process >>>> perhaps? I guess I don't know how it all works, but I just know now that >>>> with Geb 5 this method is no longer working for us. >>>> >>> -- > You received this message because you are subscribed to the Google Groups > "Geb User Mailing List" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To view this discussion on the web visit > https://groups.google.com/d/msgid/geb-user/20220412162153.C862478C0ECE%40dd19710.kasserver.com > <https://groups.google.com/d/msgid/geb-user/20220412162153.C862478C0ECE%40dd19710.kasserver.com?utm_medium=email&utm_source=footer> > . > -- You received this message because you are subscribed to the Google Groups "Geb User Mailing List" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/geb-user/CA%2B52dQSaExVtw5wBAEt0B5xF6_Z3buxVHRQFNznXuu4Ag1QgeA%40mail.gmail.com.
