fork with fixes

[mooseyboots]

i forked this and the old emacs-oath lib, and pushed the known fixes to them:
https://codeberg.org/martianh/tumblesocks, and
https://codeberg.org/martianh/emacs-oauth.

[gleitz]

This is excellent, thank you!

I added both libraries to my path and I was good to go

[gcr]

awesome !! do you mind if i merged your fixes into my branch?

[mooseyboots]

of course not! i only forked because i didn't think you ever responded to activity here. the fixes i used are in issues here in your repo.

for tumblesocks we can just merge, but for oauth, you'll need to update the readme saying that users need to first load the patched fork of it (point to mine, or you fork a version of mine so it's all here under your username, and point to that), then load tumblesocks.

[gargle]

It doesn't work anymore. oauth in elpa is 20230522.1900 now and it works in a different way. I'm still trying to figure out what's going wrong. I cloned https://codeberg.org/martianh/emacs-oauth and added it to my load-path to get the old version back.

[mooseyboots]

that's weird, as emacs-oauth @ https://github.com/psanford/emacs-oauth was mothballed 19 April 2023.

[gargle]

It seems that emacs-oauth is now at https://gitlab.com/fvdbeek/emacs-oauth. I tried to log an issue there but I'd have to register and gitlab asks for my credit card info :/

[mooseyboots]

https://gitlab.com/fvdbeek/emacs-oauth/-/issues/1.

i found the commit that changed the function we had patched, but it doesn't force the unibyte, so looks like our patch still required. let's see what the maintainer says.

[gargle]

Let's see, because it's complicated. It seems that the maintainer deleted all of the 'nonce' stuff in oauth.el and now assumes sasl is available. I remember doing (setq oauth-nonce-function 'oauth-internal-make-nonce) in my .emacs file. oauth-nonce-function is set to sasl-unique-id.

[fvdbeek]

I am the new maintainer of emacs-oauth, and it seems I broke the package by depending on a buggy HMAC-SHA1 implementation. It should be fixed by commit #4b8e9c6e.

I also removed the "nonce" stuff, all meant to simplify things. But I found an even better function to use than sasl-unique-id, so I could drop the dependency on sasl in commit #c55e402f. I assumed that sasl was omnipresent. @gargle: do you think that was a misconception?

I moved the repository to Gitlab, because that is my platform of choice. I never had to provide credit card info. Here I read that a debit/credit card is needed for a developer to run pipelines. @gargle: is it mandatory to create an account? Because then I should reconsider my preference.

[gargle]

Good morning! @folkert, many thanks for looking into this! I'm going to test your solution later on today. I checked the documentation and I too believe now that sasl is omnipresent. @folkert, gitlab started to ask for credit card info early in May 2021. https://forum.gitlab.com/t/preventing-crypto-mining-abuse-on-gitlab-com-saas/52911 explains the reason behind this move. It seems mandatory now even though many people complain. I'll let you all know if I run into any issues with my test. greetings, Johan

…

On Wed, 5 Jul 2023 at 23:59, Folkert van der Beek ***@***.***> wrote: I am the new maintainer of emacs-oauth, and it seems I broke the package by depending on a buggy HMAC-SHA1 implementation. It should be fixed by commit #4b8e9c6e <https://gitlab.com/fvdbeek/emacs-oauth/-/commit/4b8e9c6e5890cf4eda1268b60a2e47d1a8e6084a> . I also removed the "nonce" stuff, all meant to simplify things. But I found an even better function to use than sasl-unique-id, so I could drop the dependency on sasl in commit #c55e402f <https://gitlab.com/fvdbeek/emacs-oauth/-/commit/c55e402f82c8e26cff1c624932fa0d62155f85a4>. I assumed that sasl was omnipresent. @gargle <https://github.com/gargle>: do you think that was a misconception? I moved the repository to Gitlab, because that is my platform of choice. I never had to provide credit card info. Here <https://forum.gitlab.com/t/concern-about-gitlab-asking-for-credit-card/54479/7> I read that a debit/credit card is needed for a developer to run pipelines. @gargle <https://github.com/gargle>: is it mandatory to create an account? Because then I should reconsider my preference. — Reply to this email directly, view it on GitHub <#24 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAE35NUV25DCSQ5RNP2RLCTXOXPV7ANCNFSM6AAAAAAWTBHG3Y> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[gargle]

Good morning, @folkert, everything works as expected. I had to delete my old token. Reauthentication went like a breeze. My tumblr dashboard is showing. No complaints from my side! Thanks for the quick response, Prettige dag! Greetings, Johan

…

On Wed, 5 Jul 2023 at 23:59, Folkert van der Beek ***@***.***> wrote: I am the new maintainer of emacs-oauth, and it seems I broke the package by depending on a buggy HMAC-SHA1 implementation. It should be fixed by commit #4b8e9c6e <https://gitlab.com/fvdbeek/emacs-oauth/-/commit/4b8e9c6e5890cf4eda1268b60a2e47d1a8e6084a> . I also removed the "nonce" stuff, all meant to simplify things. But I found an even better function to use than sasl-unique-id, so I could drop the dependency on sasl in commit #c55e402f <https://gitlab.com/fvdbeek/emacs-oauth/-/commit/c55e402f82c8e26cff1c624932fa0d62155f85a4>. I assumed that sasl was omnipresent. @gargle <https://github.com/gargle>: do you think that was a misconception? I moved the repository to Gitlab, because that is my platform of choice. I never had to provide credit card info. Here <https://forum.gitlab.com/t/concern-about-gitlab-asking-for-credit-card/54479/7> I read that a debit/credit card is needed for a developer to run pipelines. @gargle <https://github.com/gargle>: is it mandatory to create an account? Because then I should reconsider my preference. — Reply to this email directly, view it on GitHub <#24 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAE35NUV25DCSQ5RNP2RLCTXOXPV7ANCNFSM6AAAAAAWTBHG3Y> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[gargle]

Hello again,

@fvdbeek, one little thing though, don't you have to define url-digest-auth-make-cnonce somewhere? I get an error and I feel embarrassed I didn't see it sooner. I'm not up to date with the emacs elisp eco system.

Greetings,

Johan

[fvdbeek]

Good morning, @folkert, everything works as expected. I had to delete my old token. Reauthentication went like a breeze. My tumblr dashboard is showing. No complaints from my side! Thanks for the quick response, Prettige dag!

Excellent, glad it is fixed. Goedenavond!

[fvdbeek]

@fvdbeek, one little thing though, don't you have to define url-digest-auth-make-cnonce somewhere? I get an error and I feel embarrassed I didn't see it sooner.

You need at least Emacs 26.1 for url-digest-auth-make-cnonce. Maybe this is a little too much asked for a library, and we first should depend on sasl for a while. So I reverted the commit.

[gargle]

I'm on emacs 28.2. sasl is ok for me, but aren't we missing a (require 'url-auth) somewhere as the module isn't loaded on my system by default. Greetings, Johan

…

On Thu, 6 Jul 2023 at 22:35, Folkert van der Beek ***@***.***> wrote: @fvdbeek <https://github.com/fvdbeek>, one little thing though, don't you have to define url-digest-auth-make-cnonce somewhere? I get an error and I feel embarrassed I didn't see it sooner. You need at least Emacs 26.1 for url-digest-auth-make-cnonce. Maybe this is a little too much asked for a library, and we first should depend on sasl for a while. So I reverted the commit. — Reply to this email directly, view it on GitHub <#24 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAE35NXAGWJERZYY67GCDTTXO4OPVANCNFSM6AAAAAAWTBHG3Y> . You are receiving this because you were mentioned.Message ID: ***@***.***>