You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hello Nemo. Thank you for making and maintaining the package. 👍
I am trying to use meteor-bootstrap with accounts-ui-bootstrap-3 (https://github.com/ianmartorell/meteor-accounts-ui-bootstrap-3). It works great in development mode, but not when meteor --production is used to run it, nor on modulus without a -D (debug) flag. '-D' with modulus means 'Debug' and the minimization if skipped.
According to this: https://meteorhacks.com/meteor-packaging-system-understanding-versioning
I think 3.3.1_5 is < 3.3.1_1 because the '_5' is A-Z sort < _1 -- if I'm reading the above-referenced depiction of versioning correctly.
accounts-ui-bootstrap-3 specifies the dependency weakly because one can also use twbs:bootstrap. That's why meteor does not raise an error and quit -- it's a soft dependency.
In essence, the '_5' may be undermining the packaging system's ability to correctly order dependencies. In fact, nemo64:bootstrap is after accounts-ui-bootstrap in .meteor/versions. I assume after means to be loaded after (I am not certain though). For some reason, this does not matter w/o the minification that --production brings.
I would have tested this and done I think that means I'd have to actually fork first and re-publish the package under my name -- so, I decided to raise it with you via an issue and get your thoughts. I may still do that in parallel...
Thanks!
Ian
The text was updated successfully, but these errors were encountered:
Hey @igoodrich, I've been helping maintain nemo64's bootstrap package.
In the version number x.x.x_y the _y is referred to as a wrap number, for packages that wrap 3rd party libraries, so it allows me to fix a bug while maintaining the full version of the upstream library. Bootstrap is currently at 3.3.5, thus the last release was 3.3.5 and when I fixed a bug in my code, I bumped it to 3.3.5_1. This practice is tersely explained here: http://docs.meteor.com/#/full/packagedescription
If I were to guess at what's happening, It looks like the accounts-ui-bootstrap-3 package you're using only uses a weak dependency if the version of nemo64:bootstrap is 3.3.1, which is very old. I suspect the fix would be updating accounts-ui-bootstrap-3 to api.use a broader range of versions available. See http://docs.meteor.com/#/full/pack_use
Hello Nemo. Thank you for making and maintaining the package. 👍
I am trying to use meteor-bootstrap with accounts-ui-bootstrap-3 (https://github.com/ianmartorell/meteor-accounts-ui-bootstrap-3). It works great in development mode, but not when meteor --production is used to run it, nor on modulus without a -D (debug) flag. '-D' with modulus means 'Debug' and the minimization if skipped.
You can see here (https://github.com/ianmartorell/meteor-accounts-ui-bootstrap-3/blob/master/package.js) that that accounts-ui-bootstrap specifies it's nemo64:bootstrap version thusly:
'nemo64:[email protected]_1'
Your current version is '3.3.1_5'
According to this:
https://meteorhacks.com/meteor-packaging-system-understanding-versioning
I think 3.3.1_5 is < 3.3.1_1 because the '_5' is A-Z sort < _1 -- if I'm reading the above-referenced depiction of versioning correctly.
accounts-ui-bootstrap-3 specifies the dependency weakly because one can also use twbs:bootstrap. That's why meteor does not raise an error and quit -- it's a soft dependency.
In essence, the '_5' may be undermining the packaging system's ability to correctly order dependencies. In fact, nemo64:bootstrap is after accounts-ui-bootstrap in .meteor/versions. I assume after means to be loaded after (I am not certain though). For some reason, this does not matter w/o the minification that --production brings.
I would have tested this and done I think that means I'd have to actually fork first and re-publish the package under my name -- so, I decided to raise it with you via an issue and get your thoughts. I may still do that in parallel...
Thanks!
Ian
The text was updated successfully, but these errors were encountered: