PR #1385 - remove npm-shrinkwrap.json to work around [email protected] bug (#1384)
PR #1354 Might still be needed to be merged (it is merged with harmony, but there is no backport on master yet). The change affects carriage return to newline changes in a file that has been modified in both master and harmony though, resolve is both keep as is (so no changes on harmony except the newlines).
Still not merged as of 20/1/2017 despite being closed.
Now send as #1487
PR #1408 - Have minify() and tests use figure_out_scope() as uglifyjs CLI does
Sorry for the huge delay… Moving house and the holidays have taken up way more time than I had anticipated.
We should definitely start thinking about setting up an UglifyJS organization on Github, to make the project more manageable.
@mishoo: thoughts?
PR #1426 - add missing LHS cases which global_defs should avoid
PR #1425 - augment evaluate to extract within objects
PR #1427 - optimise binary operands with evaluate()
PR #1431 - fix mangling collision with keep_fnames
if_return optimizationmaster build failure can be fixed by #1441
negate_iifeI think there's more than enough in the PR queue for a 2.8.0 release. New patches may start to conflict soon.
~Agreed - I've reached an impasse at this point 😅~
Queue is getting large
Maybe I should do a premerge some day
@avdg I can help with any merge conflicts / test failures :sweat_smile:
You already took over the whole project :p Well thanks anyway, as I seem to be specialized in tokenizer only atm... :-)
Actually I'm thinking about this (it's git/github related)
git fetch-pr 123 234 567 (numbers obviously referring to pr id's)
git is just a collection of scripts (well, it's a bit optimized). So this would make life so much easier if a script called fetch-pr was added. Not sure if such a tool exists already.
global_defs#1484 - add .gitattributes with *.js linefeed rule
--beautify bracketize@avdg not sure if this is related to what you said in https://github.com/mishoo/UglifyJS2/issues/1383#issuecomment-278766604
Unless I'm mistaken that tool gives you an easy way to set up a local branch for each PR, rather than mixing a select number of PRs together on one branch. I don't think there's a silver bullet for that other than merging all the PRs into a single branch. Normally that branch is called master.
@kzc yes, you'll still have to merge each branch yourself. But from the looks of it you don't have to go into each PR's web page to work out where the repository/branch is, which I thought was what @avdg asked about.
Edit: this one is from GitHub https://help.github.com/articles/checking-out-pull-requests-locally/
If you want to patch a bunch of PRs into your current branch for testing (no commits) you could use this bash script:
$ cat pr-patch.sh
#!/bin/bash
[ "$#" == 0 ] && echo "usage: $0 [PR numbers]" && exit 1
for i in $*
do
echo $i ...
curl -L https://github.com/mishoo/UglifyJS2/pull/$i.patch | patch -p1 || exit $_
done
sample usage:
$ bash pr-patch.sh 1483 1479 1477 1476 1475 1473 1471 1461 1459 1457 1454 1452 1449 1448 1446 1445 1444 1443 1437
It assumes you are in the source root.
I changed the PR patch script above to exit upon first patch error.
The PRs in the previous post are compatible with each other and npm test cleanly once combined against master 7f8d72d9d37396f2da05d5d824f74bd414c30119.
PR #1451 conflicts with the others.
@avdg not sure if this is related to what you said in #1383 (comment)
I'm aware of that (maybe the access should be shortcutted somehow in documentation)
I'm having some fun with this batch file on Windows:
@echo off
git fetch https://github.com/mishoo/UglifyJS2.git pull/%1/head:pr-%1
for /f %%i in ('git log pr-%1 -1 --pretty^=format:"%%an <%%ae>"') do set author=%%i
git merge --squash pr-%1
git commit --author="%author%"
git branch -D pr-%1
Need to manually edit the commit message, but otherwise seems pretty automatic.
This command will commit a github PR to your local branch with original date, author attribution and comments:
curl -L https://github.com/mishoo/UglifyJS2/pull/1483.patch | git am -3 -b
unsafe_protopure_funcs & improve side_effectsmax_line_lenI think this Issue should closed in favor of:
This can be closed.
@kzc thanks for keeping track 😉