Conversation
|
Yes, there is no version of TypeScript available anywhere that has the necessary patches to make ES6 import/export work with declarations (microsoft/TypeScript#2139), so what’s committed to my branch is a custom build with that patch applied. microsoft/TypeScript#2197 was also necessary in that build but has since been merged, so microsoft/TypeScript#2139 is the only one I’m waiting for until getting a new “official” build. Watch microsoft/TypeScript#2242 for updates! |
|
^ the reply was for this diff from @csnover alright. So I will not run |
|
@csnover I suspect the |
|
We are not doing fs checks for references. Only doing it for |
…ere was no tsconfig found closes #180
|
@basarat, I was planning on using https://www.npmjs.com/package/vinyl-fs in order to simplify this scenario. You can populate a vinyl stream with the glob patterns and it will keep itself updated when you use the watch methodology. |
|
Having a quick look I think it's meaning of glob is different : order is significant. I am sure the gulp users are used to it but that is not what we promised with our filesGlob. Our is that it must match all patterns. I am leaning towards https://www.npmjs.com/package/gaze simply watching our glob and only updating our text if the file isn't open in the editor (which the child can query) |
…vices and followed the tree from there. Updated our language service to use the new ScriptInfo class. closes #176
… Will resume from home.
…do `updateText` needlessly refs #176
|
I can definitely feel the speed up ❤️ |
I blame atom. Very very badly. refs #183 (comment)
chore(watch) don't watch project file path for in memory projects
…endings. Found an alternative. Checked that it doesn't impact performance.
|
@csnover is it okay if I merge and publish this a new major version? There are lot of good performance fixes here that I'd like to get in the hands of everyone. Its been working like a charm for me. |
|
TS team was hoping to have microsoft/TypeScript#2139 landed by end of the week so can we maybe wait until Monday to see if that happens? Otherwise I don’t think I have any major objections. |
Thanks I'll wait till monday ❤️ |
…tside of the atom view refs #177
…e assume col:1 (its one based)
…t-buffer reason: just as reliable as atom itself. No more, no less ❤️
|
I've been using this over the weekend as well as on Monday. Its much better than what we have on master so publishing as a major version. |
|
I have merged into master. That's a good thing as I can ask people to just use master for the latest fixes. However I'm not going to publish till we have a valid |



Things to do. An ever updating list: