Guild Wars 2

We have a work around, but we still don’t have a good understanding of the cause.

The work around is described in this post:
/forum/support/support/Can-t-finish-last-file-of-Nov-5th-patch/5740636

To help me understand the issue, I’d appreciate if anyone who continued to have this problem would PM me in the forums. I’d like to discuss your situation to help me come to a better understanding. Thanks.

We are working with our content delivery provider on resolving this.

We think we have flushed the corrupted files.

Are people still blocked on this? If blocked, you can try adding this temporarily to your Guild Wars 2 shortcut:
-assetsrv drwkwiejruot6.cloudfront.net

Please let us know if the problem went away. Or if not, did the work-around solve it for you. You can also reply directly to me in the forums.

The fix I described above is for the errors containing the exception code c0000417 in it, as hcopter.3269 included in his post above.

Oh My God.8423, this c0000417 issue shouldn’t be an issue on MAC. I’ll repost when we know more about that “Failed to start Coherent host” issue, described in your post.

We are pretty confident these patching crashes have to do with 3rd party software (such as malware, antivirus, and proxies) hooking into the Winsock API.

This Microsoft KB article explains how to reset TCP/IP:
https://support.microsoft.com/en-us/kb/299357

Please let us know if this does not solve the problem.

We have flushed potentially corrupt assets. Please try again.

I apologize for the inconvenience this undoubtedly caused.

Old API keys should again be working. My apologies for the disruption.

Maintenance complete.

Maintenance will begin at 3:30pm PST and service should be restored within 10 minutes.

Some public API disruption can be expected during this downtime.

Also, I’d like to work with someone who regularly sees this patching problem. Please private message me on the forum and lets see if we can figure this out.

Hi,

Whoever is having this problem, can you please run this built in Windows utility from a command prompt (cmd.exe):
> pathping assetcdn.101.arenanetworks.com

It will take a few minutes to run. Copy and paste the resulting output into an email to support@guildwars2.com with subject, “Patch download stuck at 0 kbps.”

Thanks. I hope we can get this figured out shortly.

Ah rats! Yes, turns out there was another aspect of megaservers I didn’t understand and wasn’t apparent in the development environment. We will need to temporarily disable that API until I can get it fixed for real.

The API fixes will go out with our upcoming 5/20 release. I wasn’t sure of the date until recently.

Regarding a megaserver-compatible live events API: For a few reasons, there are no plans to build this in the short/medium term, unfortunately.

Hi All, thank you for the report. Turns out there were more APIs affected by megaservers that I originally thought. I am working on a fix and hope to have it out shortly.

anzenketh.3759:

Lishtenbird.2814:

Stefan, you might not be the right person to ask it, but…

Could you tell us if this problem (transmuted weapons receiving name from stats item and not skin item) is a temporary bug or an intended change? We, the aesthetics-driven people, are really saddened by this change and would appreciate any official news. Thanks!

The API bug and the transmutation bug(if it is a bug) were unrelated. The API bug was patched last hotfix.

Hi Lishtenbird,

Yes, this isn’t the correct forum. But I went ahead and inquired internally here to someone who might know. In short, this can be considered a bug, but one that is not easy or fast to fix. There is a desire to address this along with some other improvements to the tool tips of transmuted items. But there is no ETA for this.

As to any ETA for other API releases, unfortunately I don’t have one right now either.

Thanks all!

Ya, you guys found it before I could document it. Was part of the item “name” fix and a result of the wardrobe feature. I’ll get some documentation up on the API wiki today (unless someone beats me to it).

As you can see it isn’t very interesting yet since you can only cross-reference skins that are the ‘default_skin’ of certain items. But will be more useful once we ship APIs that show personalized items, such as the /characters API.

Hi All,

We realized late in our release cycle that today’s wardrobe changes broke the “name” field for some of the items coming back from /item_details.json. I am working on a fix for this which should be out soon in a subsequent hotfix. I apologize for any inconvenience.

Hi All,

As some people have noted, the current /events.json API is not compatible with the upcoming Megaserver technology. As such, the events API will be phasing out in its current form as maps get migrated over to megaservers.

For now, instead of the events API please follow the guidance described here: https://www.guildwars2.com/en/news/the-megaserver-system-world-bosses-and-events/

We will consider implementing a timer API based on this schedule for the future.

Hi all. Yes, we are still working on the v2 APIs. It has been a little bit slow going for me due to having been pulled aside for higher priority items over the last few months. But progress is being made.

smiley.1438:

The french one seems indeed a bit weird – at least inconsistent to the others. So in french would be possibly “Cercle de feu” -> “Territoires frontaliers (Cercle de feu)”. Since these names apparently are taken directly from the game (including bugs) i guess theres some string replacement done here. Further explaination by a dev would be appreciated

I investigate this a bit. It turns out as you suspected there is some concatenation at play here, as dictated by some special markup and is different for each language translation. As in:

• EN: %str2% Borderlands
• FR: Territoires frontaliers (%str2% )
• ES: Tierras fronterizas de[pl:""] %str2%
• DE: %str1% Grenzlande

Unfortunately, what we render in the API is after the concatenation markup is trimmed out (and presumably replaced with an empty string). I did a brief investigation on how I might be able to render the name with markup, but it will require more work than I can afford right now. I’ll keep this noted as a bug.

thank you!

Hi Smiley,

It turns out those items were “in-development” items that either were rejected or supposed to be replaced with other items. We are not sure exactly how they got “discovered”. In the meantime I’ve removed them from the API. If they show up again, we’ll have to track down how they are getting discovered.

As for the additional files.json entries: I’ll try to get what I can included as soon as possible and will reply when updated.

Hi Smiley,

Thanks for the report. We are looking the problem with the ((parenthesis-named)) items.

I finally fixed the problem with the missing items, recipes, events from the previous reset we had about a month ago. I was hoping they would get re-discovered, but as people noticed this wasn’t happening in a timely fashion. It wasn’t a trivial thing to re-seed this, but it is done now. Let me know if you are still noticing anything missing. Thank you.

Hi Lolage,

Yes, I expect ultimately we would allow guild leaders to obtain guild representation and other basic guild activity of their members.

Regarding the ‘Eve’ approach to API keys: we instead are going with the OAuth2 approach, in which it is up to the 3rd party app to request required information permissions and the user may allow or deny these permissions en masse.

There are some new types of various things that have not yet been provided API display names. I will be fixing this shortly.

Thanks for the report.

Hi Sorenrye,

Interesting idea, sort of under the ‘bulk APIs’ umbrella. It is something we may consider in the future. I think our bulk/paging features will suffice when they are released, hopefully soon.

A bunch of data behind the scenes resets on new builds (which occurred this morning) and takes a few minutes to get back to normal. This is the cause of the missing world names.

Things appear to be back to normal as I write this.

Things should be coming back to normal. Thanks for the notice.

I’m taking a look. Thanks for the heads up.

Our OAuth2 implementation is basically complete. But we are awaiting actual APIs which take advantage of it before it is released. These APIs are in active development.

Thanks!

We are looking into some options for bulk access and filtering. Another option is to optionally allow specifying ids in a POST body. Though, we probably will have some sort of max id count limit so the URL constraints may not matter.

Hi Gil,

That error refers to “Guild not found”. So this appears to be a non-existent guild ID. I will followup with the WvW team.

For now you’ll just have to work-around this error. Thanks for the report.

Offering alternative data representations (such as XML) is planned, but not anytime soon. In the meantime I’d recommend using a simple JSON to XML translation library, of which there are several. There are even web-based translators you can play around with.

Hi Felicja,

The events.json and event_names.json are flushed on each new release, and get populated as events happen in real time. While the event_details.json enumerates all events ever seen. Some of your cited events such as the “Defeat the Aetherblade Captain” may no longer be happening in newer releases, which is why then don’t get populated in the events.json.

Thanks for the feedback guys.

Dr. Ishmael, your argument is interesting: just return 200, 404, and 5XX and let the caller parse the results for details.

However, I do think keeping 200 for truly ‘success’ cases makes sense because internally we have one-and-only-one ‘success’ code used across the gw2. So this will always translate correctly and it doesn’t prevent anyone who wants to treat 4XX and 200 the same, relying instead on parsing the JSON for errors.

smiley.1438:

The best way would actually be to provide all the WvW icons via the https://api.guildwars2.com/v1/files.json API

Yes, we are going to expose these via the files.json API fairly soon.

Nobrok, is this also what you are waiting on when you ask, “any word on getting the other icons”?

Hi There,

I’d like to address this 500 error problem a bit:

My proposal is to only return 500 on a relatively small set of known internal errors. Otherwise, I assume it is user error and return 400. I thought about trying to map to more granular HTTP errors, but as Cliff posted above it is a difficult and un-maintainable job.

With my proposed fix, it is still possible that I missed a few “internal” errors which will come back as 400. But as before, the error details will be provided in both 400 and 500 cases. I can always add more errors to the “internal” set as needed. In general, there are many more “user-error” than “internal-error” type codes in our system.

When authenticated APIs come out, I will likely introduce the 401 error code as well.

Any thoughts on this?

Hi,

I’m not familiar with this, but I want to help if possible. From my brief research, can’t ‘hue’ be computed from the RGB values?

As for material: what are you referring to when you say ‘vibrant’? Is there some in-game concept that I can cross reference for this?

I’ll inquire into the ‘set’ suggestion. My initial look indicates we probably can’t provide much detail there.

Thanks!

Hi coderjoe,

Yes, we do have plans to offer guild and character APIs in an authenticated fashion (via OAuth2) – but secured to a particular user. For example, you would programmatically be able to obtain last-login/representation information for yourself, not any arbitrary user in your guild.

It sounds like what you would like is some enhanced permissions for guild leaders to obtain members’ guild stats. We will definitely keep this in mind when implementing the guild APIs. Thank you for the suggestion.

I’ve posted about this “neutral” bug on another thread: https://forum-en.gw2archive.eu/forum/community/api/WvW-Ruins-after-Patch

In short, the WvW team is planning a fix, but I don’t have an ETA at this time. Thank you for the report.

Hi All,

Thank you for the report. I’ve made the WvW team is aware of this problem and they are planning a fix on their end which will trickle down to the API as well. Unfortunately, I don’t have an ETA on this. If I learn of a solid ETA, I will let you know.

Thanks.

Hi All, I’ve just deployed the “missing event update” fix. There was an unexpected problem which caused some downtime while I got this deployed. It should be updating correctly from now on and the EU event updates problem should be gone for good.

Sorry for the disruption.

Ya, offering different data formats is on our list, but rather low priority at the moment. We’ll be sticking with JSON as our primary data format for the foreseeable future. Thanks!

Hi Lazarus, good suggestions.

I am curious about your ideas for #1. At the time we roll out the OAuth-based APIs or soon after there will probably be an opportunity to also revisit the existing un-auth APIs. Feel free to PM me your ideas or post on the forum itself.

Regarding your idea #2, we have talked about doing this at some point. We wanted to get the fully factored information out there first, and then consider ways to fetch things in bulk later.

For your #3 changedSince, tracking new items or recipes etc is easy – but probably of marginal value since you guys can already do that yourself. Implementing a changedSince for existing items is actually hard given where the API sits relative to the other components. It is something I will think about though.

Thanks!

It appears a day-1 bug, I’ll get this fixed. Thanks!

Is this a regression you are seeing or a new machine you are querying from? I assume the latter. Perhaps the machine you are querying from doesn’t have some standard root certificates installed. Ensure the GeoTrust Global CA cert in installed on your machine.

The fix is prepared. My goal is to be able to release updates and fixes with reasonably fast turn-around. But due to some process changes we are implementing internally, I need to get some infrastructure setup first before I am able to get this released. I expect updates going forward will be roughly on a two-week cadence, similar to our content releases.

I saw this too. This is a transient problem during a game update release, which just happened. I’m already seeing more world names.

I can look into this as part of the CORS support on the mapping API.

Yes, we are still subject to EU event inaccuracies until I get the dropped event update problem resolved.

This was likely due to an API outage we had which was related to the GW2 outage, resolved about an hour ago. I’m now seeing events updating on my end. Please let me know if you are still having problems.