Can't find exsisting site #206

New Issue

u9000 · 2021-03-10T00:47:02Z

u9000 commented

2021-03-10 00:47:02 +00:00

When trying going to ~~gemini://srht.pages~~gemini://srht.site, bombadillo says [5] Permanent Failure. Site not found even though that site exsists. I confirmed by trying the same URL in a different gemini browser and it worked fine.

When trying going to ~~gemini://srht.pages~~`gemini://srht.site`, bombadillo says `[5] Permanent Failure. Site not found` even though that site exsists. I confirmed by trying the same URL in a different gemini browser and it worked fine.

sloum commented

2021-03-10 02:19:48 +00:00

When I visit gemini://srht.pages in Bombadillo or AV98 both of them report a DNS error. This is distinct from a response of [5] Permanent Failure. Site not found, so I am a bit confused. @u9000 can you confirm that the intended host is indeed gemini://srht.pages? At present I am unable to replicate the issue as described.

When I visit `gemini://srht.pages` in Bombadillo or AV98 both of them report a DNS error. This is distinct from a response of `[5] Permanent Failure. Site not found`, so I am a bit confused. @u9000 can you confirm that the intended host is indeed `gemini://srht.pages`? At present I am unable to replicate the issue as described.

sloum added the

needs-info

label 2021-03-10 02:20:04 +00:00

u9000 commented

2021-03-10 02:36:14 +00:00

Poster

I'm so sorry; I meant gemini://srht.site.

I'm so sorry; I meant `gemini://srht.site`.

sloum commented

2021-03-10 04:00:02 +00:00

@u9000 Thanks! Ok, so I have looked into this a bit... but am not happy with what I am finding. Which is, not a lot. I have verified you are correct that Bombadillo does not get a 2x series status code back from this host while Amfora and AV98 both do.

I have verified that the string that Bombadillo sends as the request is correct (it is). I have also verified that every time I get a 51 status code back. Which means we get the server and the server responds... it just doesn't seem to know what we are looking for. Which points back to the request... which I verified is correct. I am at a bit of a loss here. Bombadillo works fine with every gemini host I have ever tried with the exception of srht.site. I'm wondering if the server is being odd in some way that is affecting Bombadillo but not the other clients or if maybe the spec has changed and it is not being backward compatible?

I'll do some digging. At first I thought that it might be related to said host's owner threatening to black hole certain IP addresses or clients that did not go his way on a certain issue... but I do not believe that to be the case here and I believe he retrackted those threats (or at least walked them back quite a bit).

Anyway, I'll keep digging, but if anyone else has a chance to look into the source (@asdf if you are around you have tended to be very successfull at squashing these sorts of bugs - as an aside: it has been awhile, I hope you are well) it would be much appreciated.

@u9000 Thanks! Ok, so I have looked into this a bit... but am not happy with what I am finding. Which is, not a lot. I have verified you are correct that Bombadillo does not get a 2x series status code back from this host while Amfora and AV98 both do. I have verified that the string that Bombadillo sends as the request is correct (it is). I have also verified that every time I get a `51` status code back. Which means we get the server and the server responds... it just doesn't seem to know what we are looking for. Which points back to the request... which I verified is correct. I am at a bit of a loss here. Bombadillo works fine with every gemini host I have ever tried with the exception of `srht.site`. I'm wondering if the server is being odd in some way that is affecting Bombadillo but not the other clients or if maybe the spec has changed and it is not being backward compatible? I'll do some digging. At first I thought that it might be related to said host's owner threatening to black hole certain IP addresses or clients that did not go his way on a certain issue... but I do not believe that to be the case here and I believe he retrackted those threats (or at least walked them back quite a bit). Anyway, I'll keep digging, but if anyone else has a chance to look into the source (@asdf if you are around you have tended to be very successfull at squashing these sorts of bugs - as an aside: it has been awhile, I hope you are well) it would be much appreciated.

sloum added

help wanted

and removed

needs-info

labels 2021-03-10 16:14:37 +00:00

sloum added the

in progress

gemini

labels 2021-03-10 17:41:38 +00:00

sloum commented

2021-03-10 17:44:44 +00:00

Alright. I have found the issue. When Bombadillo sends a request it always includes the port in the request. Every server, or so it would seem, except the one for srht.site accepts this. srht.site works fine if the port is omitted. I believe this to be an issue with the server and not with Bombadillo. I have reached out to Drew at sourcehut to discuss the issue and see if it is something he wants to adjust on his end. If so, this should be fixed when that adjustment gets made. If not, then I will think about what to do and make code changes within Bombadillo likely amounting to: dont send a port if the port is 1965 and only send one for ports that vary from that standard gemini port. I would prefer to not make this change as I believe Bombadillo to be spec compliant in this area currently and have had no issues with it until this. However, if it becomes necessary I will make the change. I'll update here when I hear back from Drew and advise a path forward.

Alright. I have found the issue. When Bombadillo sends a request it always includes the port in the request. Every server, or so it would seem, except the one for `srht.site` accepts this. `srht.site` works fine if the port is omitted. I believe this to be an issue with the server and not with Bombadillo. I have reached out to Drew at sourcehut to discuss the issue and see if it is something he wants to adjust on his end. If so, this should be fixed when that adjustment gets made. If not, then I will think about what to do and make code changes within Bombadillo likely amounting to: dont send a port if the port is `1965` and only send one for ports that vary from that standard gemini port. I would prefer to not make this change as I believe Bombadillo to be spec compliant in this area currently and have had no issues with it until this. However, if it becomes necessary I will make the change. I'll update here when I hear back from Drew and advise a path forward.

👍 1

sloum removed the

help wanted

label 2021-03-10 17:45:03 +00:00

Sign in to join this conversation.