Deadline For Next MOULa Update - May 21.
Re: Deadline For Next MOULa Update - May 21.
Not sure there's a reason to fidget about unicode until it can be truly supported by the whole game...
Re: Deadline For Next MOULa Update - May 21.
Indeed. Incremental improvements are nice, but most of this stuff is all tied up in important fixes to many of the same subsystems (fonts, python support, etc.).Hoikas wrote:Not sure there's a reason to fidget about unicode until it can be truly supported by the whole game...
Story of my life.JWPlatt wrote:I'm not sure how rational it is to spend time on slim hopes unless you have nothing left to do
Re: Deadline For Next MOULa Update - May 21.
Story of Uru.JWPlatt wrote:I'm not sure how rational it is to spend time on slim hopes unless you have nothing left to do
-
- Member
- Posts: 317
- Joined: Sat Dec 13, 2008 10:54 am
Re: Deadline For Next MOULa Update - May 21.
D'Lanor wrote:Hmm, that fix has quite a few multiplayer implications. Though I suppose it has been thoroughly tested on Gehn by now.Christian Walther wrote:Although, if we touch that Python code anyway, I’d be more inclined to change it into what is reportedly a better fix than revert it to the previous state.
Edit: I took a closer look at that fix and I'm happy to report that it appears to be rock solid. At first I had my doubts because it seemed as if the function which restores the bookshelf for any remaining players was partly skipped. This however is already taken care of by a different ptNotify message elsewhere. So this gets a thumbs-up from me.
- I have theoretically and experimentally (using two avatars with odd KI numbers above 2^24) verified that boq’s client-side notify fix does fix the bookshelf problem (what comes out of PtGetClientIDFromAvatarKey() is a Python integer and we now handle these correctly in the notify messages).
- I have also looked at and tried boq’s bookshelf fix and have arrived at the exact same conclusions as in D'Lanor’s edit that I saw just now. (I would prefer if people would make new posts when they add new information rather than edit old ones, as edits don’t show up in the Atom feed.)
Dropping to bed now, will respond to everything else tomorrow.
Edit: Redid the commits on a named branch, having multiple heads on the default branch is confusing.
Last edited by Christian Walther on Sun May 06, 2012 8:57 am, edited 1 time in total.
Re: Deadline For Next MOULa Update - May 21.
First, thanks for spending over a half hour doing a great job at the AGM!Christian Walther wrote:...So, +1 for both reverting Cyan’s workaround and adding boq’s fix from me too...
As for this topic, Christian, can you send me something for Mark, or post here, to explain what he needs to know (background and implementation details) in regards to reverting the workaround and adding the fix? Thanks.
Perfect speed is being there.
-
- Member
- Posts: 317
- Joined: Sat Dec 13, 2008 10:54 am
Re: Deadline For Next MOULa Update - May 21.
Over half an hour? Eek! I wasn’t supposed to take more than 20 minutes! And on reading the chat log I even found more comments from the audience that I would have liked to respond to…JWPlatt wrote:First, thanks for spending over a half hour doing a great job at the AGM!
Not quite sure what you have in mind. Is what I wrote in the pull request I just made good? He really doesn’t need to know anything to apply the fix, it just comes with the package he will pull from CWE-ou/MOULSCRIPT-ou (and he can also forget about the patch we sent him separately, as that is included as well). If he wants to go more into the background, he can read the discussion that went on in various places that I’ve hopefully all linked to in the pull request. (And the pull request will be linked to from the commit message, so it should all be discoverable.)As for this topic, Christian, can you send me something for Mark, or post here, to explain what he needs to know (background and implementation details) in regards to reverting the workaround and adding the fix? Thanks.
Oh, by the way, when you next talk to Mark, could you remind him to push the 906 build number update to CWE? It would make things easier for him if we could get that into CWE-ou before we stack our new stuff on it.
Re: Deadline For Next MOULa Update - May 21.
Sure thing, Christian.
Thanks for the PR link. That looks good. I want to save Mark (and CAVCON) all the time I can, so I wouldn't ask him to read discussions because it probably wouldn't happen - at least not quickly. I try to include all info I can in a concise as possible form without sending Cyan off on link hunts for info, though I would include reference links for completeness.
Thanks for the PR link. That looks good. I want to save Mark (and CAVCON) all the time I can, so I wouldn't ask him to read discussions because it probably wouldn't happen - at least not quickly. I try to include all info I can in a concise as possible form without sending Cyan off on link hunts for info, though I would include reference links for completeness.
Perfect speed is being there.
Re: Deadline For Next MOULa Update - May 21.
The big news for today is that this is now done and ready for rarified to get it over to CWE-ou.Christian Walther wrote:Oh, by the way, when you next talk to Mark, could you remind him to push the 906 build number update to CWE? It would make things easier for him if we could get that into CWE-ou before we stack our new stuff on it.
Perfect speed is being there.
Re: Deadline For Next MOULa Update - May 21.
Just a reminder, as if you needed it, that tomorrow is the last day before Cyan picks up changes committed to CWE-ou and MOULSCRIPT-ou. If something isn't ready, no big deal - it can go into the next update.
Perfect speed is being there.
-
- Member
- Posts: 317
- Joined: Sat Dec 13, 2008 10:54 am
Re: Deadline For Next MOULa Update - May 21.
Indeed, no reminder needed – I did the “big push” mere minutes before you posted:
Everything we’ve been testing on Minkata is in CWE-ou/MOULSCRIPT-ou now, ready to be picked up by Cyan.
There are tags that mark the start (exclusive) and end (inclusive) of the revision range that Cyan is supposed to pull: cyan-2012-05-21 (C, M), ou-2012-05-21 (C, M). Note that on CWE-ou, one of the commits in this range is a merge from CWE, make sure to leave that out so you don’t accidentally get the same change applied twice. It’s only the build number update though, so that should be easy to spot. (I tried to find a way around this complication, but didn’t find a satisfactory one.)
Pull requests and Crucible reviews have been closed, JIRA issues resolved where possible.
The next step is to reset the -ou-minkata repositories to the new -ou states. That involves stripping revisions from the repositories, which only rarified can do. I have started by pushing the revisions to be added to MOULSCRIPT-ou-minkata, but noticed that that gets them in a different order to MOULSCRIPT-ou (because some were already part of MOULSCRIPT-ou-minkata before), which is a bit confusing, so it’s probably better to strip everything that’s ordered differently first and then push it again. The same on CWE-ou-minkata. Once that is done, pull request #12 can be merged to CWE-ou-minkata. (The other open pull requests probably aren’t ready yet.)
Everything we’ve been testing on Minkata is in CWE-ou/MOULSCRIPT-ou now, ready to be picked up by Cyan.
There are tags that mark the start (exclusive) and end (inclusive) of the revision range that Cyan is supposed to pull: cyan-2012-05-21 (C, M), ou-2012-05-21 (C, M). Note that on CWE-ou, one of the commits in this range is a merge from CWE, make sure to leave that out so you don’t accidentally get the same change applied twice. It’s only the build number update though, so that should be easy to spot. (I tried to find a way around this complication, but didn’t find a satisfactory one.)
Pull requests and Crucible reviews have been closed, JIRA issues resolved where possible.
The next step is to reset the -ou-minkata repositories to the new -ou states. That involves stripping revisions from the repositories, which only rarified can do. I have started by pushing the revisions to be added to MOULSCRIPT-ou-minkata, but noticed that that gets them in a different order to MOULSCRIPT-ou (because some were already part of MOULSCRIPT-ou-minkata before), which is a bit confusing, so it’s probably better to strip everything that’s ordered differently first and then push it again. The same on CWE-ou-minkata. Once that is done, pull request #12 can be merged to CWE-ou-minkata. (The other open pull requests probably aren’t ready yet.)