Lazy load unicode names

[zzzz-qq]

Lazy load unicode names when they are used first time rather than when ipython startup.

[Carreau]

@LucianaMarques that's your code, you might want to review it;

But otherwise that's looks good to me.

[LucianaMarques]

Hey there, sorry for the late reply.

If tests are working and this will improve performance, I don't see why not accept it :) That's actually really nice improvement, thanks!

I would like though that you provide us more info on the profiling you did, how you did it, etc. I'm not used to profiling yet and I would certainly like to learn more about it.

I am doing some work in the completer class, if you have some more suggestions I would love to hear it.

[LucianaMarques]

I just did a review on your pull request. Please take some time to consider my comments and let's discuss them :)

Also, I like to put descriptions in my pull requests, even if it's just something like "closes issue number #x". It's important to keep things documented in my opinion.

[Carreau]

Hey there, sorry for the late reply.

No need to apologize, we are all busy.
Those were really good comment and questions.

I would like though that you provide us more info on the profiling you did, how you did it, etc. I'm not used to profiling yet and I would certainly like to learn more about it.

I don't think much profiling is necessary. It won't be "faster" in all cases, it will mostly delay the computation of the unicodes mapping until actually requested.

This mean that the startup of IPython will be a bit faster, but the first press on <tab> that needs completion will be a bit slower.

[LucianaMarques]

I have one more code comment but i'm on my phone now, please wait before merging just a little bit.

[Carreau]

I have one more code comment but i'm on my phone now, please wait before merging just a little bit.

Don't worry plenty of time we'll wait for your approval.

[LucianaMarques]

Don't worry plenty of time we'll wait for your approval.

Thanks! Just to sum up, only exclude the comment I mentioned above and add some description to the PR, and in my opinion it's ready to be merged.

[zzzz-qq]

Hey there, sorry for the late reply.

No need to apologize, we are all busy.
Those were really good comment and questions.

I would like though that you provide us more info on the profiling you did, how you did it, etc. I'm not used to profiling yet and I would certainly like to learn more about it.

I don't think much profiling is necessary. It won't be "faster" in all cases, it will mostly delay the computation of the unicodes mapping until actually requested.

This mean that the startup of IPython will be a bit faster, but the first press on <tab> that needs completion will be a bit slower.

@Carreau Really appreaciate your patience and kindness.

[Carreau]

Really appreciate your patience and kindness.

Thanks much appreciated; I wish I had more time to mentor and guide others.
Thanks to you for taking the time to send a pull-request, and allow us to discuss it; I know it can be frustrating to wait.

I just released 7.5 yesterday so this did not make the cut, but I'm hopping to get that in 7.6 next month.

[LucianaMarques]

@Carreau are we wainting for 7.6 to merge this PR? I don't have merge permission, so I was just wandering why you didn't merge this yet.

[Carreau]

Thanks; I have some time to look at ongoing PRs, merging this.

Thanks both !