Patch covers all of HEAD except modules/ and themes/. Will mark CNR once I get all of HEAD covered.
| Comment | File | Size | Author |
|---|---|---|---|
| #52 | 606526-rollback-D7_full.patch | 4.25 KB | scor |
| #50 | 606526-rollback-D7.patch | 3.67 KB | dave reid |
| #47 | 606526_combined_ends.patch | 54.42 KB | axyjo |
| #47 | 606526_line_ends.patch | 49.1 KB | axyjo |
| #47 | 606526_file_ends.patch | 5.31 KB | axyjo |
Comments
Comment #1
axyjo commentedAll of HEAD covered.
Comment #3
axyjo commentedUsing the right CVS diff options would probably help with getting the patch to apply.
Comment #4
dries commentedI'll commit this as soon the code is frozen. Don't want to inadvertently break other patches because of this.
Comment #6
scor commentedThis patch removes all the trailing spaces from HEAD. script included.
Comment #7
yched commented/me dreams of an automated trailing space removal on commit - or of a test that would reject any patch with trailing spaces.
Comment #8
scor commented@yched I have the feeling a lightweight coder-like script might be part of PIFR 2, well, at least I hope so :)
Comment #9
webchickThis oughtta make some people cranky. ;) But hey, it is after code freeze, so... committed to HEAD! ;)
Comment #10
scor commentedI forgot to include *.inc and *.php file in my first script. patch against HEAD and updated script attached.
Comment #11
scor commentednow with *.js (toolbar.js got caught).
Comment #12
axyjo commentedThis script should ideally go into the testbot's list of things to check. It'd help with Dries' implementation of step 7 here http://buytaert.net/8-steps-for-drupal-8
Comment #13
axyjo commentedHere's an updated patch and script that checks for newlines at the end of files as well.
Comment #15
axyjo commentedAdded the --no-prefix option so that it'll work now.
Comment #17
scor commented@axyjo : make sure to not include any change to misc/ui/* since there are libraries and we are not altering these... (git checkout misc/ui/ before rolling the patch)
splitting (1) trailing spaces and (2) newline at end of files into two patches might be a good idea as it would help a lot with patches reviews.
Comment #18
axyjo commentedHere's a patch that should hopefully work, thanks to DamZ on IRC. Rolled against the git mirror of the cvs repository instead of a local git project created after checking out HEAD from cvs.
Comment #19
axyjo commentedGood idea, #17. Didn't refresh the page before I hit submit, so I'll reroll now with two distinct patches.
Comment #20
scor commented@axyjo: re newlines at end of files, files which already have a new line at the end get an extra new line.
So I agree with hunks like
but not
text.js has now 2 newlines in the end.
EDIT: gdiff | grep "No newline at end of file" | wc -l returns 15 so it should be quick to select these hunks from the patch and discard all the others; or modify the script to take this into account ;)
Comment #21
axyjo commentedGot it :D. Drastically reduces the size of the patch as well. Two separate patches posted for review, the combined patch is for committing purposes. I also updated the script a bit, so that's here as well.
Script assumes you've got a git clone of the repo in the drupal subdirectory in your current working path.
Comment #22
axyjo commentedForgot this.
Comment #23
axyjo commentedHere it is, updated based on the latest HEAD.
Comment #24
axyjo commentedHmm. I wonder if reverting the modules/simpletest/files directory will help the file ends patch pass. Here are the patches when those files are excluded. Line ends patch remains the same, so it isn't uploaded this time around. Since the combined patch has both of them, I'll just upload the file ends patch. Updated script:
Comment #25
axyjo commentedAwesome! Since that script got the green light from the testbot, here are the files again. file_ends and line_ends are just for the reviewer's sake and the actual file to patch/commit with is combined_ends.
Comment #26
scor commentedplease ignore this patch for now, I'm just trying to see if we can get more of the test files cleaned up. It's fair to leave the css_input_without_import.css.optimized.css as it is since it's the output of drupal_load_stylesheet(). Also found modules/simpletest/files/html-2.html (the html extension was not in the script, and that's the only .html file in HEAD).
EDIT: can't reproduce the failing UploadTestCase on my localhost.
Comment #28
scor commentedNevermind. Reuploading the combined patch from #25
Comment #29
brianV commentedIt all looks good to me, and the testbots are greenlighting it.
I think this is RTBC.
Comment #30
axyjo commentedUpdated patch. Left at RTBC since both brianV and scor have looked at it.
Comment #31
scor commentedThis is a library so I'm not sure we should alter it.
Comment #32
axyjo commentedAgreed. Added an exception to jQuery files in the script. Here's the patch without it.
Comment #33
scor commentedlooks good.
Comment #36
scor commentedrerolled
Comment #37
scor commentedrerolled patches (including script).
Comment #38
axyjo commentedRerolled patch with scor's new script.
Comment #39
dave reidAdding tag and subscribing...
Comment #40
boombatower commentedLets get this in, looks good (simple).
Part of #666022: [meta issue] Style cleanup.
Comment #41
dries commentedCan we extend this script to deal with #516576: Convert CRLF to LF in all jquery ui files as well? I'd rather run such a script locally, than accept a monster patch where it is hard to verify all the changes. Just askin' ... :)
Comment #42
scor commentedSince the jQuery UI files are external libraries, I don't think we should modify them in any way, and that's why the script in this issue ignores
misc/ui/andmisc/jquery.*. Is there any performance improvement in changing CRLF to LF other than saving one char per line and complying with our coding standards?Comment #43
axyjo commentedAren't all jQuery libs basically 1 LOC due to compression/packer? So that's upto 20 lines total in a file. 20 bytes saved by changing CRLF to LF times 41 (number of jQuery files in misc/ and misc/ui/) = 820 bytes saved overall. Is it really worth it in terms of file size?
+ scor brings up a good issue about jQuery being an external library.
Comment #44
scor commentedhttp://drupal.org/coding-standards does not specify whether our coding standards should apply to external libraries as well. I believe it should not but let's see what other people think.
Comment #45
axyjo commentedAgreed with scor. This reduces our work since we don't have to patch the files every single time an external library updates (even though we freeze the version in our tree).
Comment #46
axyjo commentedWith coder integration now in the test bot, I really think that this should be fixed ASAP. Bumping the priority up and bumping the status down until I do a re-roll.
Comment #47
axyjo commentedHere's that re-roll I promised. Setting back to RTBC since it's the same script that's used.
Comment #48
dries commentedCommitted to CVS HEAD.
Comment #49
scor commentedsee follow up patch for dreditor #656456: Auto-detect trailing white-space and missing EOF newline in patch review so we don't have to reopen this issue again...
Comment #50
dave reidThe changes to UnicodeTest.txt are causing the current HEAD failures since two lines had their characters removed.
http://drupalcode.org/viewvc/drupal/drupal/modules/search/tests/UnicodeT...
http://qa.drupal.org/pifr/test/32
Rolling back the changes to UnicodeTest.txt puts me back to 1 fail with Search simplify test.
Comment #51
dave reidComment #52
scor commentedfull reroll of Unicodetest.txt, which solves all 3 failures of the Search simplify test.
The file UnicodeTest.txt was introduced after the last time the testbot ran the tests sucessfully on patch #38. For some reason the testbot didn't report back on the previous patch #47.
Comment #53
catch(note to anyone who thinks the patch is broken, Drupal.org thinks patch files are iso-something, if you manually change to utf-8 in your browse it'll look fine).
Comment #54
webchickCommitted. Thanks!
Comment #55
webchickEr. That. :)