Now I understand where you were going. I misunderstood the point in the process where it was, I think.
Anyway, that said, nothing to contribute to the last post. I'm really not sure what all would be useful to send, aside from what has already been mentioned. :)
This was something I talked with Kaelten and Net about... I really like this... if it's implemented properly (no didn't read the whole thread).
First thing it should do before updating would be to check if the version being run is the latest and greatest. If not, bitchslap the user, if so report the error to me.
This was something I talked with Kaelten and Net about... I really like this... if it's implemented properly (no didn't read the whole thread).
First thing it should do before updating would be to check if the version being run is the latest and greatest. If not, bitchslap the user, if so report the error to me.
Define "latest and greatest" though. The new Curse system is to have Alphas, Betas and Releases, and you're going to get a lot more bug reports on Betas and Releases because most people won't bother with Alphas (especially until the Curse Client gets the ability to pull Alphas from WowAce).
Making someone checkout and update a project using an SVN client and then providing them with a fancy automated bug reporting tool would be laughably inconsistent.
If an addon is forced to have a ## Version: tag in the TOC, or a line added to X-metadata by the packager to indicate which commit id that version is, then it would be much easier to work with.
If an addon is forced to have a ## Version: tag in the TOC, or a line added to X-metadata by the packager to indicate which commit id that version is, then it would be much easier to work with.
The mandatory ToC fields seemed to be something that they (Curse) didn't want to get involved with. I had suggested that to them before to make the updater a bit easier to use.
For theory sutff: as for associating addon to bug report, i'v found anything that has a "in C code [C] ? " that any thing beyond that is completely useless. So any addon named before that point in the stack trace, should be marked as being part of the bug report but the report would show up for the addon where it failed at, IE the first line of the bug.
For theory sutff: as for associating addon to bug report, i'v found anything that has a "in C code [C] ? " that any thing beyond that is completely useless. So any addon named before that point in the stack trace, should be marked as being part of the bug report but the report would show up for the addon where it failed at, IE the first line of the bug.
:: Prays what he said makes sense::
This is incorrect. There are many situations where the
"in C code [C] ? "
is relevant, because the function call chain passed through C code temporarily.
Example:
AddonA calls frameA:Show(), that triggers the OnShow() script for frameA to fire, but the caller of the OnShow() script is from C code (since Show() is a C side function), but you can see the trace it through.
Example:
AddonB calls SetMapZoom(C, Z). That triggers WORLD_MAP_UPDATE event immediately and all addons (there are HEAPS of these, but most will check if WorldMapFrame is visible before doing anything) that listen to WORLD_MAP_UPDATE fire and react. Again that passes through C code to generate the event and trigger the various Ace2/3/Rock callbacks.
I had played with some ideas localaly, and had planned to use one of the cloud services from google or ms, but just got wrapped up in a busy schedule.
Still on the fence about it personally. it has potential, but at the same time, its one of those "low reward" efforts to provide a service to developers who at best dont really care, and at worst are hostile towards it, and users who the usuall people will try to convince that its spyware.
For exaple, i think they did end up adding the toc entry to identify the version number. So there was a whine thread about that - i think thats what has me hesitant. Can you imagine if they get worked up about that, then hear, oh we are uploading your addons.txt, and saved variables to report an issue.
It requires buy-in from alot of people at alot of levels, plus lots of time. I'd prefer that some other developers join in on the effort before taking it much further, ie. a team development effort.
Define "latest and greatest" though. The new Curse system is to have Alphas, Betas and Releases, and you're going to get a lot more bug reports on Betas and Releases because most people won't bother with Alphas (especially until the Curse Client gets the ability to pull Alphas from WowAce).
Making someone checkout and update a project using an SVN client and then providing them with a fancy automated bug reporting tool would be laughably inconsistent.
Cuse said the the 'average' user should use curse release packages. So if there is an error in a 'release' package, that has been fixed in SVN, then a new release should (must?) be released as stable on the curse client.
eh, if there is a problem in a release package, then **NOTHING** is required to be done, everything is in the control of the author(s). The idea this thread is all about is IF an error comes up, should we and how could we get that to automatically report back to the author or party responsible for that part of the addon. However if the CC could easily do this then i'd be for it.. however the ticket system is a great alternative.
Rollback Post to RevisionRollBack
To post a comment, please login or register a new account.
Anyway, that said, nothing to contribute to the last post. I'm really not sure what all would be useful to send, aside from what has already been mentioned. :)
First thing it should do before updating would be to check if the version being run is the latest and greatest. If not, bitchslap the user, if so report the error to me.
Define "latest and greatest" though. The new Curse system is to have Alphas, Betas and Releases, and you're going to get a lot more bug reports on Betas and Releases because most people won't bother with Alphas (especially until the Curse Client gets the ability to pull Alphas from WowAce).
Making someone checkout and update a project using an SVN client and then providing them with a fancy automated bug reporting tool would be laughably inconsistent.
You guys can help by attaching your bugsack saved variables and your addon listing from your wtf folder. You can PM me with the info also.
This is a good idea. I'll make a ticket.
http://www.curseforge.com/projects/curseforge/tickets/816-have-packager-add-a-tag-to-the-toc-for-the-curse-version/
:: Prays what he said makes sense::
This is incorrect. There are many situations where the
"in C code [C] ? "
is relevant, because the function call chain passed through C code temporarily.
Example:
AddonA calls frameA:Show(), that triggers the OnShow() script for frameA to fire, but the caller of the OnShow() script is from C code (since Show() is a C side function), but you can see the trace it through.
Example:
AddonB calls SetMapZoom(C, Z). That triggers WORLD_MAP_UPDATE event immediately and all addons (there are HEAPS of these, but most will check if WorldMapFrame is visible before doing anything) that listen to WORLD_MAP_UPDATE fire and react. Again that passes through C code to generate the event and trigger the various Ace2/3/Rock callbacks.
Still on the fence about it personally. it has potential, but at the same time, its one of those "low reward" efforts to provide a service to developers who at best dont really care, and at worst are hostile towards it, and users who the usuall people will try to convince that its spyware.
For exaple, i think they did end up adding the toc entry to identify the version number. So there was a whine thread about that - i think thats what has me hesitant. Can you imagine if they get worked up about that, then hear, oh we are uploading your addons.txt, and saved variables to report an issue.
It requires buy-in from alot of people at alot of levels, plus lots of time. I'd prefer that some other developers join in on the effort before taking it much further, ie. a team development effort.
Cuse said the the 'average' user should use curse release packages. So if there is an error in a 'release' package, that has been fixed in SVN, then a new release should (must?) be released as stable on the curse client.