x/build: rewrite Hg hashes in issue tracker

[hanwen]

Some old issues refer to mercurial hashes in their issue tracker. There is a mapping at https://golang.org/change/XXXXX and the git commits have this data too (as a golang-hg header in the commit), but if you don't know that it can be hard to track down.

Instead of people having to know this, one could rewrite all codesite URLs in the issue tracker to become googlesource.com URLs.

example: #2516 mentions http://code.google.com/p/go/source/detail?r=4e38cad33649.

[bradfitz]

Out of curiosity, I searched for the "code.google.com/p/go/source/detail?r=" substring and found 237 matches, using:

func TestHGLinks(t *testing.T) {
        c := getGoData(t)
        repo := c.GitHub().Repo("golang", "go")
        repo.ForeachIssue(func(gi *maintner.GitHubIssue) error {
                return gi.ForeachComment(func(c *maintner.GitHubComment) error {
                        if !strings.Contains(c.Body, "code.google.com/p/go/source/detail?r=") {
                                return nil
                        }
                        fmt.Printf("https://github.com/golang/go/issues/%d#issuecomment-%d\n", gi.Number, c.ID)
                        return nil
                })
        })
}

(added to x/build/maintner/godata/godata_test.go)

Results:

#9 (comment)
#20 (comment)
#47 (comment)
#97 (comment)
#188 (comment)
#244 (comment)
#255 (comment)
#256 (comment)
#313 (comment)
#317 (comment)
#345 (comment)
#363 (comment)
#494 (comment)
#785 (comment)
#917 (comment)
#955 (comment)
#996 (comment)
#1033 (comment)
#1101 (comment)
#1143 (comment)
#1286 (comment)
#1297 (comment)
#1317 (comment)
#1325 (comment)
#1382 (comment)
#1439 (comment)
#1450 (comment)
#1450 (comment)
#1451 (comment)
#1464 (comment)
#1465 (comment)
#1508 (comment)
#1547 (comment)
#1552 (comment)
#1650 (comment)
#1704 (comment)
#1728 (comment)
#1734 (comment)
#1744 (comment)
#1750 (comment)
#1768 (comment)
#1816 (comment)
#1826 (comment)
#1894 (comment)
#1899 (comment)
#1916 (comment)
#1945 (comment)
#1960 (comment)
#1960 (comment)
#1972 (comment)
#1972 (comment)
#1973 (comment)
#1976 (comment)
#2019 (comment)
#2021 (comment)
#2041 (comment)
#2093 (comment)
#2093 (comment)
#2093 (comment)
#2152 (comment)
#2165 (comment)
#2206 (comment)
#2237 (comment)
#2281 (comment)
#2313 (comment)
#2328 (comment)
#2344 (comment)
#2396 (comment)
#2415 (comment)
#2444 (comment)
#2450 (comment)
#2451 (comment)
#2460 (comment)
#2516 (comment)
#2580 (comment)
#2590 (comment)
#2653 (comment)
#2653 (comment)
#2715 (comment)
#2726 (comment)
#2741 (comment)
#2787 (comment)
#2833 (comment)
#2841 (comment)
#2874 (comment)
#2919 (comment)
#2946 (comment)
#2947 (comment)
#2967 (comment)
#3050 (comment)
#3135 (comment)
#3170 (comment)
#3225 (comment)
#3254 (comment)
#3258 (comment)
#3293 (comment)
#3316 (comment)
#3329 (comment)
#3345 (comment)
#3352 (comment)
#3386 (comment)
#3431 (comment)
#3504 (comment)
#3523 (comment)
#3643 (comment)
#3660 (comment)
#3683 (comment)
#3683 (comment)
#3684 (comment)
#3702 (comment)
#3749 (comment)
#3787 (comment)
#3863 (comment)
#3863 (comment)
#3874 (comment)
#3875 (comment)
#3903 (comment)
#3917 (comment)
#4021 (comment)
#4054 (comment)
#4195 (comment)
#4259 (comment)
#4276 (comment)
#4276 (comment)
#4299 (comment)
#4459 (comment)
#4461 (comment)
#4477 (comment)
#4551 (comment)
#4551 (comment)
#4553 (comment)
#4643 (comment)
#4692 (comment)
#4721 (comment)
#4723 (comment)
#4834 (comment)
#4842 (comment)
#4868 (comment)
#4889 (comment)
#4892 (comment)
#4902 (comment)
#4973 (comment)
#4979 (comment)
#5028 (comment)
#5064 (comment)
#5091 (comment)
#5139 (comment)
#5148 (comment)
#5148 (comment)
#5148 (comment)
#5178 (comment)
#5179 (comment)
#5207 (comment)
#5234 (comment)
#5291 (comment)
#5380 (comment)
#5487 (comment)
#5494 (comment)
#5512 (comment)
#5563 (comment)
#5625 (comment)
#5723 (comment)
#5742 (comment)
#5747 (comment)
#5752 (comment)
#5780 (comment)
#5791 (comment)
#5797 (comment)
#5811 (comment)
#5872 (comment)
#5928 (comment)
#5928 (comment)
#5928 (comment)
#5935 (comment)
#5935 (comment)
#5998 (comment)
#6013 (comment)
#6036 (comment)
#6103 (comment)
#6121 (comment)
#6125 (comment)
#6177 (comment)
#6178 (comment)
#6240 (comment)
#6254 (comment)
#6276 (comment)
#6490 (comment)
#6567 (comment)
#6607 (comment)
#6624 (comment)
#6656 (comment)
#6679 (comment)
#6735 (comment)
#6771 (comment)
#6784 (comment)
#6963 (comment)
#7022 (comment)
#7056 (comment)
#7085 (comment)
#7092 (comment)
#7115 (comment)
#7159 (comment)
#7246 (comment)
#7247 (comment)
#7362 (comment)
#7364 (comment)
#7364 (comment)
#7370 (comment)
#7377 (comment)
#7420 (comment)
#7506 (comment)
#7506 (comment)
#7548 (comment)
#7556 (comment)
#7557 (comment)
#7569 (comment)
#7570 (comment)
#7596 (comment)
#7621 (comment)
#7738 (comment)
#7824 (comment)
#7856 (comment)
#7896 (comment)
#7956 (comment)
#7982 (comment)
#8001 (comment)
#8001 (comment)
#8270 (comment)
#8276 (comment)
#8300 (comment)
#8426 (comment)
#8426 (comment)
#8475 (comment)
#8489 (comment)
#8638 (comment)
#8668 (comment)
#8690 (comment)

But a proper fix would build a trie out of all golang-hg hashes from the git commits,

$ git cat-file -p 7d7c6a97f815e9279d08cfaea7d5efb5e90695a8
tree e06bd601885e16ad3d72c2a8c9b411889b2e478e
author Brian Kernighan <bwk> 80352345 -0500
committer Brian Kernighan <bwk> 80352345 -0500
golang-hg f6182e5abf5eb0c762dddbb18f8854b7e350eaeb

hello, world

R=ken
DELTA=7  (7 added, 0 deleted, 0 changed)

And then look to find any 6 or 7+ hex substring of those hg hashes.

[bradfitz]

/cc @kevinburke @andybons

[jaredborner]

i generated a csv file mapping the comments/revisions to their corresponding git commits. (the mapping was generated with git log --all --pretty=raw | awk '$1=="commit"{c=$2}$1=="golang-hg"{print $2,c}')

hg2git.matched.txt

this failed to find a commit for a few of the revisions. i think it's because they came from a different repo.

hg2git.missing.txt