Re: Bug Reports (snapshot builds)
Righteo, got it 
High Quality Resources for Collectible Card Games and Home of the CCGHQ Team
https://www.slightlymagic.net/forum/
https://www.slightlymagic.net/forum/viewtopic.php?f=52&t=6333
That can be avoided, if you parse mtg-data.txt or xml and get card's collection number. Won't work for every set (because early sets don't have numbers strictly assigned to cards), but for most will do.friarsol wrote:Otherwise we'd be trying to hit magiccards.info 12000+ times in short order everytime the script would run.
That's sad. Wish they were.friarsol wrote:I don't believe flip cards are handled by SetInfoScript, so that just proves my point that having scriptors hardcode this info will lead to issues.
Yea we have a separate script that's run to clear out anything for particular sets, that's what moomarc was referring to early. It's a bit heavy-handed, since you need to re-run SetInfo afterwards. Maybe if I have some free time I'll make that a bit more elegant.Max mtg wrote:There might be a special mode to run the script so that it re-evaluates all the cards.
Yep, it's because the SetInfoScript was written before Alternate card faces were made available, so there's no consideration for that style of script.Max mtg wrote:That's sad. Wish they were.friarsol wrote:I don't believe flip cards are handled by SetInfoScript, so that just proves my point that having scriptors hardcode this info will lead to issues.
myk wrote:There are some notes in the code that the svn revision is not always reported properly. This is because the svn revision is pulled from the root directory when doing a build, and a lot of people just do their svn updates from some sort of GUI, which usually only updates whatever directory tree you r-clicked on when you selected 'update'. We use the buildnumber plugin to get the svn revision for us in the pom. I was looking at its documentation, and it seems we can get a consistently updated value if we set its doUpdate parameter to true. Any bug reports that come in to this thread will then have accurate SVN revision labels. Another plus is that it would ensure the res directory is always up to date, preventing issues like what Almeida saw. The potential downside, though, is that it will attempt to merge whatever is on the server, and if there are local changes, this can be a hassle. What do you guys think? Would enabling the doUpdate param be a good idea?
Ok, here's a patch below that seems to do what I want. I didn't commit it because I wasn't sure if we should be using CardCharacteristic in this way, or if there's a different spot in changeZone where we'd prefer the color reset. Feel free to confirm anyone who is familiar. If noone has any strong preference, I can patch this in and add Deathlace tomorrow night.friarsol wrote:Edit: Clarification, Spells going from the Stack to any zone doesn't reset colors. So perhaps unintentionally, going from Stack to Battlefield is functional. But Stack to anywhere else (no matter what type of spell) the color doesn't reset.
It should not affect them at all. It would be the equivalent of doing an 'svn up' command before running 'mvn package'. For people who are just syncing svn to stay bleeding edge, it would actually be less work compared to now, since they would no longer need to do explicit svn updates. It would just be a potential annoyance if (1) local changes are unmerged and conflict with something that was just checked in by someone else or (2) you're doing revision bisecting trying to find the first bad commit while tracking down a bug. We could solve (2) by introducing an optional commandline parameter (e.g. 'mvn -DdoUpdate=false package'), but I do fear that (1) might cause some angst. Those devs who do not want svn to be automatically updated could create a properties file and set doUpdate to false in there, and we could read that from the pom, but of course if that were to become popular to do, why have the autoupdate at all. We could try it and see how it goes -- it might not become a problem at all. Let me test it out locally for a bit to see if it causes issues.Chris H. wrote:How would this impact my building and deploying the snapshots and the betas?myk wrote:Would enabling the doUpdate param be a good idea?
myk wrote:It should not affect them at all. It would be the equivalent of doing an 'svn up' command before running 'mvn package'. For people who are just syncing svn to stay bleeding edge, it would actually be less work compared to now, since they would no longer need to do explicit svn updates. It would just be a potential annoyance if (1) local changes are unmerged and conflict with something that was just checked in by someone else or (2) you're doing revision bisecting trying to find the first bad commit while tracking down a bug. We could solve (2) by introducing an optional commandline parameter (e.g. 'mvn -DdoUpdate=false package'), but I do fear that (1) might cause some angst. Those devs who do not want svn to be automatically updated could create a properties file and set doUpdate to false in there, and we could read that from the pom, but of course if that were to become popular to do, why have the autoupdate at all. We could try it and see how it goes -- it might not become a problem at all. Let me test it out locally for a bit to see if it causes issues.
edit: not sure if this is an option, but this particular kind of problem doesn't exist with git.