-
Notifications
You must be signed in to change notification settings - Fork 3k
Re-allow updates to pre-existing "invalid" named packages #7195
Comments
This is making @dominictarr sad over on #7260. I'm investigating whether a change to npm or |
not me, so much as the ???? of people who use JSONStream and need to used a better version of it. ;) |
Any update on this? (still waiting for JSONStream update ;) |
No, alas, we haven't had time to come up with a plan for this. It's probably sufficient to relax the restrictions in npm / |
this did work previously, may I ask what broke it? |
I add myself to the queue.. cannot publish new versions of JSUS ... I cannot even go lowercase, because in the meantime some dude has published a package named jsus ! So I am really stuck! JSUS is a dependency of many other packages of mine, so this is a very bad situation for me. I tried to use an older version of npm 1.5.x, but I get other errors and I cannot even login. Please re-enable publishing names with upper case characters. Thanks! |
It's on the list of things to do, but hasn't made it to the top of the list yet. If somebody wanted to submit a patch to |
@othiym23 can you give us a hit as to what fix is necessary? |
@dominictarr |
browserify versions 5.0.0 - 10.1.1 (10.1.2 is the latest) can't be installed because of JSONStream vs jsonstream on OS X and windows. It's related to the issue described in this issue. From browserify/browserify#1247 (comment):
|
cordova has a dependency on browserify 7.1.0 and now all of our published versions are failing to install. What changed? |
@stevengill the explanation is in browserify/browserify#1247 (comment). Your best bet is to update to 10.1.2. Because of how versions float, this seems to only affect v5.0.0-v10.1.1. |
Hmm, even if we release a new version of Cordova the last few major releases are going to remain broken. |
All downstreams of cordova are breaking too (phonegap, etc). Ugh |
Yes, this breaks many of my dependencies and I can no longer |
Yeah, when browserify made this change, $#!t hit the fan. Having something like JSONString == jsonstring seems like a larger issue. |
+1 |
Looks like the issue got fixed by using an old version of npm to publish JSONStream. dominictarr/JSONStream#68 |
@stevengill hey i think it's fixed can you try installing cordova again (I can't reproduce because I'm on linux) |
@dominictarr it is fixed. Thanks! |
@dominictarr thanks! |
@dominictarr, great work. Many thanks! |
After much digging around it seems that the change in npm that originally forced @dominictarr to publish JSONStream as jsonstream has been fixed and thus the jsonstream package has been deprecated. See: - npm/npm#7260 - npm/npm#7195 - dominictarr/JSONStream@31a4975
After much digging around it seems that the change in npm that originally forced @dominictarr to publish JSONStream as jsonstream has been fixed and thus the jsonstream package has been deprecated. See: - npm/npm#7260 - npm/npm#7195 - dominictarr/JSONStream@31a4975
After much digging around it seems that the change in npm that originally forced @dominictarr to publish JSONStream as jsonstream has been fixed and thus the jsonstream package has been deprecated. See: - npm/npm#7260 - npm/npm#7195 - dominictarr/JSONStream@31a4975 This reverts commit 79271de.
After much digging around it seems that the change in npm that originally forced @dominictarr to publish JSONStream as jsonstream has been fixed and thus the jsonstream package has been deprecated. See: - npm/npm#7260 - npm/npm#7195 - dominictarr/JSONStream@31a4975 This partially reverts commit 79271de
After much digging around it seems that the change in npm that originally forced @dominictarr to publish JSONStream as jsonstream has been fixed and thus the jsonstream package has been deprecated. See: - npm/npm#7260 - npm/npm#7195 - dominictarr/JSONStream@31a4975 This partially reverts commit 79271de
See fb55/css-what#7 (comment)
If folks would like to push updates to packages that have mixed-case names, or even deprecate those packages in favor of lower-case ones, we're blocking them from doing so because of our well-intended data cleanup bits and bobs.
Probably that should just rely on server-side validation? I'm not 100% sure of the right course of action here, but it feels like we inadvertently left some folks stuck and unable to do the right thing.
The text was updated successfully, but these errors were encountered: