profile
viewpoint
Keegan Carruthers-Smith keegancsmith @sourcegraph Cape Town, South Africa http://people.cs.uct.ac.za/~ksmith/ South African working for @sourcegraph. Did stints in Bay Area and London for @facebook.

gchaincl/sqlhooks 381

Attach hooks to any database/sql driver

keegancsmith/FUSEPod 4

FUSEPod is a userspace filesystem which mounts your iPod into a directory for easy browsing of your songs on your iPod.

keegancsmith/advent 3

Advent of Code solutions - http://adventofcode.com/

keegancsmith/django-fileflatpages 3

Makes FlatPage fixtures easily

keegancsmith/counsel-repo 2

Emacs jump to repository using Ivy

keegancsmith/Agatha 1

Global Game Jam 2011 Entry

keegancsmith/AmarokPidgin 1

A plugin for Amarok that updates your Pidgin status message with that you are currently listening to.

keegancsmith/AmarokQuilt 1

A Full Screen mode for Amarok which displays a quilt of your album covers as well as the current playing track.

keegancsmith/docker-proxy 1

Proxy to expose the docker unix socket over tcp

keegancsmith/dotfiles 1

My configuration files

startedkljensen/golang-html5-sse-example

started time in 6 hours

push eventsourcegraph/sourcegraph

Keegan Carruthers-Smith

commit sha 7dd698860a4bdf865f397797b3cbee7caa678d45

Apply suggestions from code review Co-authored-by: uwedeportivo <534011+uwedeportivo@users.noreply.github.com> Co-authored-by: ᴜɴᴋɴᴡᴏɴ <joe@sourcegraph.com>

view details

push time in 13 hours

issue commentsourcegraph/sourcegraph

Search 504s

Syntect is a complex stack machine with lots of arbitrary regexp thrown into the mix, some files are going to take a really long time to highlight or get the stack machine stuck, unfortunately. This is also a known issue and is why we use a subprocess model in syntect-server to enable us to kill/escape those cases.

Don't we already have heuristics in the frontend client for syntect that prevents asking it to highlight some files? It seems completely reasonable to me to skip highlighting a file with a long line or that is large?

nicksnyder

comment created time in 15 hours

push eventsourcegraph/sourcegraph

Keegan Carruthers-Smith

commit sha 74ad60c5b67a800804279e036f8c849e14fc02c9

dev: add DEV_NO_CONFIG to prevent overwriting config (#12788) When developing and restarting the sourcegraph frontend, it will overwrite the configuration stored in the database. This setting prevents loading configuration from files, which in turn will prevent the overwriting behaviour.

view details

push time in 15 hours

delete branch sourcegraph/sourcegraph

delete branch : k/dev-no-config

delete time in 15 hours

PR merged sourcegraph/sourcegraph

Reviewers
dev: add DEV_NO_CONFIG to prevent overwriting config

When developing and restarting the sourcegraph frontend, it will overwrite the configuration stored in the database. This setting prevents loading configuration from files, which in turn will prevent the overwriting behaviour.

+16 -10

4 comments

3 changed files

keegancsmith

pr closed time in 15 hours

push eventsourcegraph/src-cli

Boyan Soubachov

commit sha 79c45f56537108b4e746353205b867edc9effca3

Fix `serve-git` documentation inconsistency (#257)

view details

push time in 15 hours

PR merged sourcegraph/src-cli

Fix `serve-git` documentation inconsistency

Fixed a documentation inconsistency for the server-git command in README.md

+1 -1

0 comment

1 changed file

boyan-soubachov

pr closed time in 15 hours

issue commentsourcegraph/sourcegraph

redis-cache segfault in single server dogfood instance

This problem also tickled me. I took a quick look after what Tomas found. I remember the on-disk formats for appendonly.aof was quite simple, and I hoped the same was true for RDB. So I looked at a hexdump in both files around the bad key. (in redis-cache-backup)

From reading rdbSaveRawString strings are just encoded as len data. It has some heuristics, but it looks like it does the verbatim encoding since we can read the string. rdbSaveLen is quite complicated, so we need to guess where the length starts but we can reverse engineer. The string has length 85 so it seems it will be encoded into [2]byte (2 bits for int type, 14 bits for length). Quoting redis src:

01|XXXXXX XXXXXXXX => 01, the len is 14 byes, 6 bits + 8 bits of next byte

So this little bit of python can decode lengths:

f = lambda a, b: ((a&0x3F)<<8)|b
# >>> f(0x40, 0x55)
# 85

So the "@U" is the length of the string! It looks like the value is appearing after that, but I have just hit my wall and will go to sleep now. I'll leave checking out the length of the value (to see if we are doing overflows) as an exercise for the reader.

The key in dump.rdb

# strings -t d dump.rdb    | grep orgs/sourcegraph 
  83514 @Uv2:http:http://127.0.0.1:3180/orgs/sourcegraph/repos?sort=created&page=3&per_page=100
2572037 @Uv2:http:http://127.0.0.1:3180/orgs/sourcegraph/repos?sort=created&page=2&per_page=100
2797503 @Uv2:http:http://127.0.0.1:3180/orgs/sourcegraph/repos?sort=created&page=5&per_page=100
3406123 @Uv2:http:http://127.0.0.1:3180/orgs/sourcegraph/repos?sort=created&page=4&per_page=100
4275305 @Uv2:http:http://127.0.0.1:3180/orgs/sourcegraph/repos?sort=created&page=1&per_page=100

# hexdump -s 4275290 -n 200 -C dump.rdb 
00413c5a  c6 0c ad cb 73 01 00 00  f8 80 00 05 02 57 00 40  |....s........W.@|
00413c6a  55 76 32 3a 68 74 74 70  3a 68 74 74 70 3a 2f 2f  |Uv2:http:http://|
00413c7a  31 32 37 2e 30 2e 30 2e  31 3a 33 31 38 30 2f 6f  |127.0.0.1:3180/o|
00413c8a  72 67 73 2f 73 6f 75 72  63 65 67 72 61 70 68 2f  |rgs/sourcegraph/|
00413c9a  72 65 70 6f 73 3f 73 6f  72 74 3d 63 72 65 61 74  |repos?sort=creat|
00413caa  65 64 26 70 61 67 65 3d  31 26 70 65 72 5f 70 61  |ed&page=1&per_pa|
00413cba  67 65 3d 31 30 30 c3 80  00 01 4c 56 80 00 08 59  |ge=100....LV...Y|
00413cca  f5 1f 48 54 54 50 2f 31  2e 31 20 32 30 30 20 4f  |..HTTP/1.1 200 O|
00413cda  4b 0d 0a 54 72 61 6e 73  66 65 72 2d 45 6e 63 6f  |K..Transfer-Enco|
00413cea  64 69 1f 6e 67 3a 20 63  68 75 6e 6b 65 64 0d 0a  |di.ng: chunked..|
00413cfa  41 63 63 65 73 73 2d 43  6f 6e 74 72 6f 6c 2d 41  |Access-Control-A|
00413d0a  6c 6c 6f 0a 77 2d 4f 72  69 67 69 6e 3a 20 2a e0  |llo.w-Origin: *.|
00413d1a  08 1f 19 45 78 70 6f 73                           |...Expos|
00413d22

The key in appendonly.aof

# strings -t d appendonly.aof     | grep orgs/sourcegraph 
 151149 @Uv2:http:http://127.0.0.1:3180/orgs/sourcegraph/repos?sort=created&page=5&per_page=100
 745433 @Uv2:http:http://127.0.0.1:3180/orgs/sourcegraph/repos?sort=created&page=4&per_page=100
 982409 @Uv2:http:http://127.0.0.1:3180/orgs/sourcegraph/repos?sort=created&page=2&per_page=100
2756815 @Uv2:http:http://127.0.0.1:3180/orgs/sourcegraph/repos?sort=created&page=1&per_page=100
3184082 @Uv2:http:http://127.0.0.1:3180/orgs/sourcegraph/repos?sort=created&page=3&per_page=100

# hexdump -s 2756790 -n 200 -C appendonly.aof 
002a10b6  73 22 3a 31 30 36 7d 5d  7d fc c6 0c ad cb 73 01  |s":106}]}.....s.|
002a10c6  00 00 f8 80 00 05 02 59  00 40 55 76 32 3a 68 74  |.......Y.@Uv2:ht|
002a10d6  74 70 3a 68 74 74 70 3a  2f 2f 31 32 37 2e 30 2e  |tp:http://127.0.|
002a10e6  30 2e 31 3a 33 31 38 30  2f 6f 72 67 73 2f 73 6f  |0.1:3180/orgs/so|
002a10f6  75 72 63 65 67 72 61 70  68 2f 72 65 70 6f 73 3f  |urcegraph/repos?|
002a1106  73 6f 72 74 3d 63 72 65  61 74 65 64 26 70 61 67  |sort=created&pag|
002a1116  65 3d 31 26 70 65 72 5f  70 61 67 65 3d 31 30 30  |e=1&per_page=100|
002a1126  c3 80 00 01 4c 56 80 00  08 59 f5 1f 48 54 54 50  |....LV...Y..HTTP|
002a1136  2f 31 2e 31 20 32 30 30  20 4f 4b 0d 0a 54 72 61  |/1.1 200 OK..Tra|
002a1146  6e 73 66 65 72 2d 45 6e  63 6f 64 69 1f 6e 67 3a  |nsfer-Encodi.ng:|
002a1156  20 63 68 75 6e 6b 65 64  0d 0a 41 63 63 65 73 73  | chunked..Access|
002a1166  2d 43 6f 6e 74 72 6f 6c  2d 41 6c 6c 6f 0a 77 2d  |-Control-Allo.w-|
002a1176  4f 72 69 67 69 6e 3a 20                           |Origin: |
002a117e
uwedeportivo

comment created time in a day

pull request commentsourcegraph/sourcegraph

dev: add DEV_NO_CONFIG to prevent overwriting config

@slimsag The latter. Personally I never want to load config from a file in dev. I prefer relying on the state in the DB. The number of times I have cursed when working on a new feature (behind a feature flag) and sourcegraph frontend starting up (due to start.sh or recompile) removing the feature flag is innumerable.

As it stands we have a "default dev" config hardcoded, we could mutate that if we wanted to change the default dev config. My guess is most devs would also prefer that behaviour?

keegancsmith

comment created time in a day

PR opened sourcegraph/sourcegraph

Reviewers
dev: add DEV_NO_CONFIG to prevent overwriting config

When developing and restarting the sourcegraph frontend, it will overwrite the configuration stored in the database. This setting prevents loading configuration from files, which in turn will prevent the overwriting behaviour.

+16 -10

0 comment

3 changed files

pr created time in a day

create barnchsourcegraph/sourcegraph

branch : k/dev-no-config

created branch time in a day

Pull request review commentsourcegraph/sourcegraph

search: facilitate simple searches if globbing is active

 func (r *searchResolver) doResults(ctx context.Context, forceOnlyResultType stri 		multiErr = nil 	} -	sortResults(results)+	if _, isAndOr := r.query.(*query.AndOrQuery); isAndOr {+		r.sortResultsAndOr(ctx, results)+	} else {+		sortResults(results)+	}

I'm not sure why you need a different sort function for and/or. Can't we just modify sortResults?

stefanhengl

comment created time in 2 days

Pull request review commentsourcegraph/sourcegraph

search: facilitate simple searches if globbing is active

 func compareSearchResults(a, b SearchResultResolver) bool { 	if arepo == brepo { 		return afile < bfile 	}- 	return arepo < brepo }  func sortResults(r []SearchResultResolver) { 	sort.Slice(r, func(i, j int) bool { return compareSearchResults(r[i], r[j]) }) } +// compareSearchResultsAndOr is like compareSearchResults, but overrides sorting in alphabetical order if+// one of the filenames is contained in exactFilePatterns, in which case exact matches are sorted by+// length of their file path and then alphabetically.+func compareSearchResultsAndOr(a, b searchResultURIGetter, exactFilePatterns map[string]struct{}) bool {+	arepo, afile := a.searchResultURIs()+	brepo, bfile := b.searchResultURIs()++	if arepo == brepo {+		if exactFilePatterns == nil || len(exactFilePatterns) == 0 {+			return afile < bfile+		}+		_, aMatch := exactFilePatterns[filepath.Base(afile)]+		_, bMatch := exactFilePatterns[filepath.Base(bfile)]+		if aMatch || bMatch {+			if aMatch && bMatch {+				if len(afile) < len(bfile) {+					return true+				}+				if len(bfile) < len(afile) {+					return false+				}

nit:

                // Prefer shorter file names (ie root files come first)
				if len(afile) != len(bfile) {
					return len(afile) < len(bfile)
				}
stefanhengl

comment created time in 2 days

Pull request review commentsourcegraph/sourcegraph

search: facilitate simple searches if globbing is active

 func compareSearchResults(a, b SearchResultResolver) bool { 	if arepo == brepo { 		return afile < bfile 	}- 	return arepo < brepo }  func sortResults(r []SearchResultResolver) { 	sort.Slice(r, func(i, j int) bool { return compareSearchResults(r[i], r[j]) }) } +// compareSearchResultsAndOr is like compareSearchResults, but overrides sorting in alphabetical order if+// one of the filenames is contained in exactFilePatterns, in which case exact matches are sorted by+// length of their file path and then alphabetically.+func compareSearchResultsAndOr(a, b searchResultURIGetter, exactFilePatterns map[string]struct{}) bool {+	arepo, afile := a.searchResultURIs()+	brepo, bfile := b.searchResultURIs()++	if arepo == brepo {+		if exactFilePatterns == nil || len(exactFilePatterns) == 0 {+			return afile < bfile+		}+		_, aMatch := exactFilePatterns[filepath.Base(afile)]+		_, bMatch := exactFilePatterns[filepath.Base(bfile)]+		if aMatch || bMatch {+			if aMatch && bMatch {+				if len(afile) < len(bfile) {+					return true+				}+				if len(bfile) < len(afile) {+					return false+				}+				return afile < bfile+			}+			if aMatch {+				return true+			}+			return false
            // prefer exact matches
			return aMatch
stefanhengl

comment created time in 2 days

Pull request review commentsourcegraph/sourcegraph

search: facilitate simple searches if globbing is active

 func compareSearchResults(a, b SearchResultResolver) bool { 	if arepo == brepo { 		return afile < bfile 	}- 	return arepo < brepo }  func sortResults(r []SearchResultResolver) { 	sort.Slice(r, func(i, j int) bool { return compareSearchResults(r[i], r[j]) }) } +// compareSearchResultsAndOr is like compareSearchResults, but overrides sorting in alphabetical order if+// one of the filenames is contained in exactFilePatterns, in which case exact matches are sorted by+// length of their file path and then alphabetically.+func compareSearchResultsAndOr(a, b searchResultURIGetter, exactFilePatterns map[string]struct{}) bool {+	arepo, afile := a.searchResultURIs()+	brepo, bfile := b.searchResultURIs()++	if arepo == brepo {+		if exactFilePatterns == nil || len(exactFilePatterns) == 0 {+			return afile < bfile+		}+		_, aMatch := exactFilePatterns[filepath.Base(afile)]

I think you should use path here, not filepath. As a benefit path.Base is more performant due to assumptions it can make. We are abstracted away from the FS (I hope :P).

stefanhengl

comment created time in 2 days

Pull request review commentsourcegraph/sourcegraph

search: facilitate simple searches if globbing is active

 func compareSearchResults(a, b SearchResultResolver) bool { 	if arepo == brepo { 		return afile < bfile 	}- 	return arepo < brepo }  func sortResults(r []SearchResultResolver) { 	sort.Slice(r, func(i, j int) bool { return compareSearchResults(r[i], r[j]) }) } +// compareSearchResultsAndOr is like compareSearchResults, but overrides sorting in alphabetical order if+// one of the filenames is contained in exactFilePatterns, in which case exact matches are sorted by+// length of their file path and then alphabetically.+func compareSearchResultsAndOr(a, b searchResultURIGetter, exactFilePatterns map[string]struct{}) bool {+	arepo, afile := a.searchResultURIs()+	brepo, bfile := b.searchResultURIs()++	if arepo == brepo {+		if exactFilePatterns == nil || len(exactFilePatterns) == 0 {

you don't need to check for nil here, you can do len on nil maps.

stefanhengl

comment created time in 2 days

issue commentsourcegraph/sourcegraph

Sourcegraph.com zoekt needs more resources

zoekt-indexserver still uses significant CPU, without seeming to affect the only other service metric available (average revision resolve duration)

Average revision resolve duration will is measuring an RPC call, so it is putting load on gitserver (via frontend). So lots of CPU use indicates it is likely indexing a lot. I would look at the recently added queue metrics.

zoekt-webserver still uses all of its memory, with frequently firing "50s+ indexed search request errors every 5m by code" alerts - this might be an issue of the alert being on a hard threshold rather than a ratio

The webserver uses a lot of memory even if it is not serving any requests. IE the memory use is dominated by the working set of indexes, not the results or number of them generated. At this scale we would need a lot of traffic for it to contribute to memory use over just holding the indexes in memory.

slimsag

comment created time in 2 days

issue commentsourcegraph/sourcegraph

Sourcegraph.com zoekt needs more resources

Here is the graph over the last 14d:

image

I'm not sure how this graph interacts with silences, but it seems that the day this issue was filed (24th) the warning alerts went up a bunch. The 29th I shipped some fixes and the big one on the 30th (and one or two more the next day or two). You can see the graph go mostly silent again. Then on the weekend we scaled up index search to be 200k instead of 100k repos and you can see on Monday (when the site has more activity) all the alerts start up again. I would suspect the root cause is that. cc @beyang

slimsag

comment created time in 2 days

issue commentsourcegraph/go-diff

MultiFileDiffReader doesn’t handle messages in diff that certain files are only available in specific version.

Thanks for the proposal. I think we would. It seems like the set of inputs we understand will only increase + the empty diff makes sense. @eseliger @mrnugget thoughts? Any concerns around the nil time param?

sofiia-tesliuk

comment created time in 2 days

issue commentsourcegraph/sourcegraph

Enable globbing for Sourcegraph org members

What if we sort exact matches by length (keeping alphabetical order for everything else)? This way exact matches on root level would always come first and we would not need to introduce a new identifier for root files.

I like this approach. In fact we implemented something like this a long time ago. IE we implemented some "ranking" on results based on input. So a match on a filename is ranked higher than matching on a dirname, etc. I think simple things here go a long way. I think our current approach of sorting is just naive, so as long as we are mostly deterministic then what we are doing is good.

rvantonder

comment created time in 2 days

push eventsourcegraph/sourcegraph

Keegan Carruthers-Smith

commit sha 9051ea5579a308fd6201e6b67e88806a1d0eaf89

add upsert external services to internal client

view details

Keegan Carruthers-Smith

commit sha 3a9ee36cb4bf72ce6321ab7e36bddad912b983b4

wip remove dbstore externalservice functions

view details

push time in 2 days

issue commentsourcegraph/sourcegraph

Search 504s

Rijnard van Tonder notifications@github.com writes:

But I would suspect the issue isn't the speed of "grep", but that we are struggling to fetch an archive of ghdump. Our searcher code doesn't do per line searching, but uses the same technique in rg of searching the whole body and hydrating in line breaks.

Whoops, guess my diagnosis is off then. I'll remember that part of our searcher now :-) The downloaded zip (at master) is only about 200MB , which is about the size of the Linux and we don't have any issues with that IIRC (git clone of ghdump is 2.6GB). So seems like there's still something else going on.

Ok then it likely is us being slow at searching lol. We have integration benchmarks which can easily plug in any repo@hash. So if you feel like playing with our oldest search code have fun :D

nicksnyder

comment created time in 2 days

push eventsourcegraph/src-cli

Keegan Carruthers-Smith

commit sha 52a40c9fb9484f340331fbaf591aaeae5854e3b9

servegit: link to documentation in output

view details

Keegan Carruthers-Smith

commit sha 58b03541d0652a08a4b3020a87e368c3b4f4c49d

document serve-git

view details

push time in 2 days

PR opened sourcegraph/sourcegraph

Reviewers
web: Add codehost entry for src serve-git

"src serve-git" is the new and streamlined version of src-expose. We add an entry for it in our frontend for code host setup to simplify use of it.

Additionally this PR updates documentation around src-expose to instead point to or be about "src serve-git". Note: we still haven't done a tagged release of src including serve-git. I will ensure that is done before merging this document so the instructions are valid.

Alternatively we considered add a new code host type. But that is quite a heavyweight approach from a backend perspective.

Fixes https://github.com/sourcegraph/sourcegraph/issues/12363

+80 -8

0 comment

6 changed files

pr created time in 2 days

push eventsourcegraph/sourcegraph

Keegan Carruthers-Smith

commit sha 87dcd78d840ab98ad049bb2ee2ca75ce5b17a487

web: Add codehost entry for src serve-git "src serve-git" is the new and streamlined version of src-expose. We add an entry for it in our frontend for code host setup to simplify use of it. Additionally this PR updates documentation around src-expose to instead point to or be about "src serve-git". Note: we still haven't done a tagged release of src including serve-git. I will ensure that is done before merging this document so the instructions are valid. Alternatively we considered add a new code host type. But that is quite a heavyweight approach from a backend perspective.

view details

push time in 2 days

create barnchsourcegraph/sourcegraph

branch : k/expose-svc

created branch time in 2 days

PR closed sourcegraph/sourcegraph

Update golang Docker tag bot

This PR contains the following updates:

Package Type Update New value Sourcegraph
golang stage patch 1.14.6-alpine code search for "golang"
golang stage minor 1.14-alpine code search for "golang"

Renovate configuration

:date: Schedule: "on the 1st through 7th day of the month" in timezone America/Los_Angeles.

:vertical_traffic_light: Automerge: Disabled by config. Please merge this manually once you are satisfied.

:recycle: Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

:ghost: Immortal: This PR will be recreated if closed unmerged. Get config help if that's undesired.


  • [ ] <!-- rebase-check -->If you want to rebase/retry this PR, check this box

This PR has been generated by WhiteSource Renovate. View repository job log here.

+3 -3

1 comment

3 changed files

renovate[bot]

pr closed time in 2 days

PR closed sourcegraph/sourcegraph

Update dependency history to v4.10.1 bot npm

This PR contains the following updates:

Package Type Update New value References Sourcegraph
history resolutions minor 4.10.1 source code search for "history"

Release Notes

<details> <summary>ReactTraining/history</summary>

v4.10.1

Compare Source

  • Fixes some issues with using hash history on a page with a <base> tag (#​577, #​578)

v4.10.0

Compare Source

  • Allows basename URL prefixes with special regex characters (#​566, #​544)

v4.9.0

Compare Source

The main change in this release is pre-bundled single-file builds for both development and production environments. This is not a breaking change, since we still support old-style require('history/createBrowserHistory') statements for now. However, these are deprecated in favor of importing/requireing from the main module.

Please see the release notes for 4.98.0-beta.0 for more info.

v4.7.2

Compare Source

v4.7.1

Compare Source

v4.7.0

Compare Source

v4.6.3

Compare Source

v4.6.2

Compare Source

v4.6.1

Compare Source

v4.6.0

Compare Source

</details>


Renovate configuration

:date: Schedule: "on the 1st through 7th day of the month" in timezone America/Los_Angeles.

:vertical_traffic_light: Automerge: Disabled by config. Please merge this manually once you are satisfied.

:recycle: Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

:no_bell: Ignore: Close this PR and you won't be reminded about this update again.


  • [ ] <!-- rebase-check -->If you want to rebase/retry this PR, check this box

This PR has been generated by WhiteSource Renovate. View repository job log here.

+19 -18

0 comment

2 changed files

renovate[bot]

pr closed time in 2 days

PR closed sourcegraph/sourcegraph

Update dependency @reach/menu-button to v0.10.5 bot npm

This PR contains the following updates:

Package Type Update New value References Sourcegraph
@reach/menu-button dependencies patch 0.10.5 source code search for "@reach/menu-button"

Release Notes

<details> <summary>reach/reach-ui</summary>

v0.10.5

Compare Source

🐛 Bugfixes

✨ Enhancements

🤜🤛 3 Committers

v0.10.4

Compare Source

🐛 Bugfixes
📝 Documentation
✨ Enhancements
  • combobox
    • Ability to pass targetRef to ComboboxPopover just like a normal Popover (@​sanfilippopablo)
🤜🤛 6 Committers

v0.10.3

Compare Source

🔨 Breaking Changes

  • The only meaningful change to packages in this release is to the API for @reach/descendants. We do not document this module and don't generally recommend it for external use, but in case you're being bold, useDescendants has been renamed to useDescendantsInit. useDescendants is still used, but it will now access the actual descendant list from the appropriate context.

📝 Documentation

🏡 Housekeeping

🤜🤛 4 Committers

</details>


Renovate configuration

:date: Schedule: "on the 1st through 7th day of the month" in timezone America/Los_Angeles.

:vertical_traffic_light: Automerge: Disabled by config. Please merge this manually once you are satisfied.

:recycle: Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

:no_bell: Ignore: Close this PR and you won't be reminded about this update again.


  • [ ] <!-- rebase-check -->If you want to rebase/retry this PR, check this box

This PR has been generated by WhiteSource Renovate. View repository job log here.

+52 -52

0 comment

2 changed files

renovate[bot]

pr closed time in 2 days

PR closed sourcegraph/sourcegraph

Update dependency @reach/listbox to ^0.10.5 bot npm

This PR contains the following updates:

Package Type Update New value References Sourcegraph
@reach/listbox dependencies patch ^0.10.5 source code search for "@reach/listbox"

Release Notes

<details> <summary>reach/reach-ui</summary>

v0.10.5

Compare Source

🐛 Bugfixes

✨ Enhancements

🤜🤛 3 Committers

v0.10.4

Compare Source

🐛 Bugfixes
📝 Documentation
✨ Enhancements
  • combobox
    • Ability to pass targetRef to ComboboxPopover just like a normal Popover (@​sanfilippopablo)
🤜🤛 6 Committers

v0.10.3

Compare Source

🔨 Breaking Changes

  • The only meaningful change to packages in this release is to the API for @reach/descendants. We do not document this module and don't generally recommend it for external use, but in case you're being bold, useDescendants has been renamed to useDescendantsInit. useDescendants is still used, but it will now access the actual descendant list from the appropriate context.

📝 Documentation

🏡 Housekeeping

🤜🤛 4 Committers

v0.10.2

Compare Source

🐛 Bugfixes

✨ Enhancements

🤜🤛 3 Committers

</details>


Renovate configuration

:date: Schedule: "on the 1st through 7th day of the month" in timezone America/Los_Angeles.

:vertical_traffic_light: Automerge: Disabled by config. Please merge this manually once you are satisfied.

:recycle: Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

:no_bell: Ignore: Close this PR and you won't be reminded about this update again.


  • [ ] <!-- rebase-check -->If you want to rebase/retry this PR, check this box

This PR has been generated by WhiteSource Renovate. View repository job log here.

+62 -62

0 comment

2 changed files

renovate[bot]

pr closed time in 2 days

PR closed sourcegraph/sourcegraph

Update dependency @reach/accordion to ^0.10.5 bot npm

This PR contains the following updates:

Package Type Update New value References Sourcegraph
@reach/accordion dependencies patch ^0.10.5 source code search for "@reach/accordion"

Release Notes

<details> <summary>reach/reach-ui</summary>

v0.10.5

Compare Source

🐛 Bugfixes

✨ Enhancements

🤜🤛 3 Committers

v0.10.4

Compare Source

🐛 Bugfixes
📝 Documentation
✨ Enhancements
  • combobox
    • Ability to pass targetRef to ComboboxPopover just like a normal Popover (@​sanfilippopablo)
🤜🤛 6 Committers

v0.10.3

Compare Source

🔨 Breaking Changes

  • The only meaningful change to packages in this release is to the API for @reach/descendants. We do not document this module and don't generally recommend it for external use, but in case you're being bold, useDescendants has been renamed to useDescendantsInit. useDescendants is still used, but it will now access the actual descendant list from the appropriate context.

📝 Documentation

🏡 Housekeeping

🤜🤛 4 Committers

</details>


Renovate configuration

:date: Schedule: "on the 1st through 7th day of the month" in timezone America/Los_Angeles.

:vertical_traffic_light: Automerge: Disabled by config. Please merge this manually once you are satisfied.

:recycle: Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

:no_bell: Ignore: Close this PR and you won't be reminded about this update again.


  • [ ] <!-- rebase-check -->If you want to rebase/retry this PR, check this box

This PR has been generated by WhiteSource Renovate. View repository job log here.

+27 -27

1 comment

2 changed files

renovate[bot]

pr closed time in 2 days

PR closed sourcegraph/sourcegraph

Update alpine Docker tag bot

This PR contains the following updates:

Package Type Update New value Sourcegraph
alpine final minor 3.12 code search for "alpine"
alpine stage minor 3.12 code search for "alpine"
alpine final minor 3.12.0 code search for "alpine"

Renovate configuration

:date: Schedule: "on the 1st through 7th day of the month" in timezone America/Los_Angeles.

:vertical_traffic_light: Automerge: Disabled by config. Please merge this manually once you are satisfied.

:recycle: Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

:ghost: Immortal: This PR will be recreated if closed unmerged. Get config help if that's undesired.


  • [ ] <!-- rebase-check -->If you want to rebase/retry this PR, check this box

This PR has been generated by WhiteSource Renovate. View repository job log here.

+13 -13

1 comment

9 changed files

renovate[bot]

pr closed time in 2 days

PR closed sourcegraph/sourcegraph

Update jest monorepo to v26 (major) bot npm

This PR contains the following updates:

Package Type Update New value References Sourcegraph
babel-jest devDependencies major ^26.1.0 source code search for "babel-jest"
jest (source) devDependencies major ^26.1.0 homepage, source code search for "jest"

Release Notes

<details> <summary>facebook/jest</summary>

v26.1.0

Compare Source

Features
  • [jest-mock] Export Mock, MockInstance, SpyInstance types (#​10138)
  • [jest-config] Support config files exporting (async) functions (#​10001)
  • [jest-cli, jest-core] Add --selectProjects CLI argument to filter test suites by project name (#​8612)
  • [jest-cli, jest-init] Add coverageProvider to jest --init prompts (#​10044)
Fixes
  • [jest-console] getConsoleOutput to receive global stack trace config and use it to format stack trace (#​10081)
  • [jest-jasmine2] Stop adding : after an error that has no message (#​9990)
  • [jest-diff] Control no diff message color with commonColor in diff options (#​9997)
  • [jest-snapshot] Fix TypeScript compilation (#​10008)
Chore & Maintenance
  • [docs] Correct confusing filename in enableAutomock example (#​10055)
  • [jest-core] 🎉🎉🎉🎉🎉🎉🎉🎉🎉🎉🎉🎉🎉🎉🎉🎉🎉🎉🎉🎉🎉🎉🎉🎉🎉 (#​10000)
  • [jest-core, jest-reporters, jest-test-result, jest-types] Cleanup displayName type (#​10049)
  • [jest-runtime] Jest-internal sandbox escape hatch (#​9907)
  • [jest-fake-timers] Update now param type to support Date in addition to number. (#​10169)
  • [docs] Add param to setSystemTime docs and remove preceding period from it and getRealSystemTime (#​10169)
  • [jest-snapshot, jest-util] Replace make-dir with fs.mkdir (#​10136)
  • [docs] Added parcel-bundler documentation inside readme.md file
Performance
  • [jest-core, jest-transform, jest-haste-map] Improve Jest startup time and test runtime, particularly when running with coverage, by caching micromatch and avoiding recreating RegExp instances (#​10131)

v26.0.1

Compare Source

Fixes
  • [jest-circus] Backward compatibility for deprecated DescribeBlock.tests to not break e.g. Detox reporter

v26.0.0

Compare Source

Features
  • [jest-environment-jsdom] [BREAKING] Upgrade jsdom to v16 (#​9606)
  • [@jest/fake-timers] Add possibility to use a modern implementation of fake timers, backed by @sinonjs/fake-timers (#​7776)
  • [jest-runtime] Add createMockFromModule as an alias for genMockFromModule (#​9962)
Fixes
  • [babel-jest] Handle null being passed to createTransformer (#​9955)
  • [jest-circus, jest-console, jest-jasmine2, jest-reporters, jest-util, pretty-format] Fix time durating formatting and consolidate time formatting code (#​9765)
  • [jest-circus] [BREAKING] Fail tests if a test takes a done callback and have return values (#​9129)
  • [jest-circus] [BREAKING] Throw a proper error if a test / hook is defined asynchronously (#​8096)
  • [jest-circus] Throw more descriptive error if hook is defined inside test (#​9957)
  • [jest-circus] [BREAKING] Align execution order of tests to match jasmine's top to bottom order (#​9965)
  • [jest-config, jest-resolve] [BREAKING] Remove support for browser field (#​9943)
  • [jest-haste-map] Stop reporting files as changed when they are only accessed (#​7347)
  • [jest-resolve] Show relative path from root dir for module not found errors (#​9963)
  • [jest-runtime] Fix absolute path moduleNameMapper + jest.mock bug (#​8727)
Chore & Maintenance
  • [*] [BREAKING] TypeScript definitions requires a minimum of TypeScript v3.8 (#​9823)
  • [*] [BREAKING] Drop support for Node 8 (#​9423)
  • [*] Upgrade to chalk@4 (#​9752)
  • [*] Remove usage of realpath-native (#​9952)
  • [docs] Fix example reference implementation to use Jest with Phabricator (#​8662)
  • [docs] Added default compiler to tranform (#​8583)
  • [docs] Updated Testing Frameworks guide with React; make it generic (#​9106)
  • [expect, jest-mock, pretty-format] [BREAKING] Remove build-es5 from package (#​9945)
  • [@jest/fake-timers, @&#8203;jest/environment] [BREAKING] Rename LolexFakeTimers to ModernFakeTimers (#​9960)
  • [jest-haste-map] [BREAKING] removed providesModuleNodeModules (#​8535)
  • [jest-runtime] [BREAKING] Remove long-deprecated require.requireActual and require.requireMock methods (#​9854)

</details>


Renovate configuration

:date: Schedule: "on the 1st through 7th day of the month" in timezone America/Los_Angeles.

:vertical_traffic_light: Automerge: Disabled by config. Please merge this manually once you are satisfied.

:recycle: Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

:no_bell: Ignore: Close this PR and you won't be reminded about these updates again.


  • [ ] <!-- rebase-check -->If you want to rebase/retry this PR, check this box

This PR has been generated by WhiteSource Renovate. View repository job log here.

+820 -80

0 comment

2 changed files

renovate[bot]

pr closed time in 2 days

PR closed sourcegraph/sourcegraph

Update golang:1.14-alpine Docker digest to 9887985 bot

This PR contains the following updates:

Package Type Update New value Sourcegraph
golang stage digest 1.14-alpine code search for "golang"

Renovate configuration

:date: Schedule: "on the 1st through 7th day of the month" in timezone America/Los_Angeles.

:vertical_traffic_light: Automerge: Disabled by config. Please merge this manually once you are satisfied.

:recycle: Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

:no_bell: Ignore: Close this PR and you won't be reminded about this update again.


  • [ ] <!-- rebase-check -->If you want to rebase/retry this PR, check this box

This PR has been generated by WhiteSource Renovate. View repository job log here.

+2 -2

2 comments

2 changed files

renovate[bot]

pr closed time in 2 days

PR closed sourcegraph/sourcegraph

Update alpine:3.11 Docker digest to 39eda93 bot

This PR contains the following updates:

Package Type Update New value Sourcegraph
alpine final digest 3.11 code search for "alpine"

Renovate configuration

:date: Schedule: "on the 1st through 7th day of the month" in timezone America/Los_Angeles.

:vertical_traffic_light: Automerge: Disabled by config. Please merge this manually once you are satisfied.

:recycle: Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

:no_bell: Ignore: Close this PR and you won't be reminded about this update again.


  • [ ] <!-- rebase-check -->If you want to rebase/retry this PR, check this box

This PR has been generated by WhiteSource Renovate. View repository job log here.

+2 -2

1 comment

2 changed files

renovate[bot]

pr closed time in 2 days

PR closed sourcegraph/sourcegraph

Update prom/prometheus:v2.16.0 Docker digest to efd99a6 bot

This PR contains the following updates:

Package Type Update New value Sourcegraph
prom/prometheus stage digest v2.16.0 code search for "prom/prometheus"

Renovate configuration

:date: Schedule: "on the 1st through 7th day of the month" in timezone America/Los_Angeles.

:vertical_traffic_light: Automerge: Disabled by config. Please merge this manually once you are satisfied.

:recycle: Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

:no_bell: Ignore: Close this PR and you won't be reminded about this update again.


  • [ ] <!-- rebase-check -->If you want to rebase/retry this PR, check this box

This PR has been generated by WhiteSource Renovate. View repository job log here.

+1 -1

1 comment

1 changed file

renovate[bot]

pr closed time in 2 days

PR closed sourcegraph/sourcegraph

Update monaco bot npm

This PR contains the following updates:

Package Type Update New value References Sourcegraph
monaco-editor dependencies minor ^0.20.0 source code search for "monaco-editor"
monaco-editor-webpack-plugin devDependencies minor ^1.9.0 source code search for "monaco-editor-webpack-plugin"

Release Notes

<details> <summary>Microsoft/monaco-editor</summary>

v0.20.0

Compare Source

New & Noteworthy
  • The editor can now be hosted inside a Shadow Root.
  • There is new API to read the editor's content width and height.
  • New editor options:
    • renderValidationDecorations - render validation decorations even in read only editors
    • wrappingStrategy - delegate wrapping points computation to the browser
    • comments.insertSpace - insert a space around comments when running commenting commands
    • foldingHighlight - highlight folded regions
    • peekWidgetDefaultFocus - focus the inline editor or the tree when opening peek view
Breaking changes
  • Renamed onCompositionStart, onCompositionEnd to onDidCompositionStart, onDidCompositionEnd
  • Changed the argument passed in to onDidPaste
  • WorkspaceEdit.edits has now changed its shape such that edits must no longer be grouped by resource.
  • The Monaco Editor no longer supports IE 11. The last version that was tested in IE 11 is 0.18.1.
Thank you

Contributions to monaco-editor:

Contributions to monaco-typescript:

Contributions to monaco-languages:

v0.19.3

Compare Source

  • brings back a way to get resolved editor options - #​1734
Thank you

Contributions to monaco-editor:

Contributions to monaco-languages:

v0.19.2

Compare Source

  • fixes issue with default value of autoIndent - #​1726

v0.19.1

Compare Source

  • fixes issue with .d.ts file in the ESM distribution - #​1729
  • adds types for global editor options (such as wordBasedSuggestions) - #​1746
  • adds support for reStructuredText.
Thank you

Contributions to monaco-editor:

Contributions to monaco-languages:

v0.19.0

Compare Source

New & Noteworthy
  • It is now possible to pass in a dimension in the editor construction options in order to avoid a synchronous layout.
  • There is new API to provide semantic tokens.
  • New options:
    • multiCursorPaste: define how to distribute paste in case of multi-cursor
    • matchBrackets: control if enclosing brackets should be highlighted
  • Fixes for tokenization in: TypeScript, JavaScript, Handlebars, Kotlin and VB.
Breaking changes
  • getConfiguration() is replaced by getRawOptions(), which returns the passed in editor options.
  • Starting with this version, the Monaco Editor no longer supports IE 11. The last version that was tested in IE 11 is 0.18.1.
Thank you

Contributions to monaco-editor:

Contributions to monaco-editor-webpack-plugin:

Contributions to monaco-languages:

Contributions to monaco-typescript:

</details>

<details> <summary>Microsoft/monaco-editor-webpack-plugin</summary>

v1.9.0

Compare Source

</details>


Renovate configuration

:date: Schedule: "on the 1st through 7th day of the month" in timezone America/Los_Angeles.

:vertical_traffic_light: Automerge: Disabled by config. Please merge this manually once you are satisfied.

:recycle: Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

:ghost: Immortal: This PR will be recreated if closed unmerged. Get config help if that's undesired.


  • [ ] <!-- rebase-check -->If you want to rebase/retry this PR, check this box

This PR has been generated by WhiteSource Renovate. View repository job log here.

+6 -11

3 comments

2 changed files

renovate[bot]

pr closed time in 2 days

PR closed sourcegraph/sourcegraph

Update dependency jsdom to v16 bot needs work npm

This PR contains the following updates:

Package Type Update New value References Sourcegraph
jsdom devDependencies major ^16.2.2 source code search for "jsdom"
@types/jsdom devDependencies major 16.2.0 source code search for "@types/jsdom"

Release Notes

<details> <summary>jsdom/jsdom</summary>

v16.2.2

Compare Source

  • Updated StyleSheetList for better spec compliance; notably it no longer inherits from Array.prototype. (ExE-Boss)
  • Fixed requestAnimationFrame() from preventing process exit. This likely regressed in v16.1.0.
  • Fixed setTimeout() to no longer leak the closures passed in to it. This likely regressed in v16.1.0. (AviVahl)
  • Fixed infinite recursion that could occur when calling click() on a <label> element, or one of its descendants.
  • Fixed getComputedStyle() to consider inline style="" attributes. (eps1lon)
  • Fixed several issues with <input type="number">'s stepUp() and stepDown() functions to be properly decimal-based, instead of floating point-based.
  • Fixed various issues where updating selectEl.value would not invalidate properties such as selectEl.selectedOptions. (ExE-Boss)
  • Fixed <input>'s src property, and <ins>/<del>'s cite property, to properly reflect as URLs.
  • Fixed window.addEventLister, window.removeEventListener, and window.dispatchEvent to properly be inherited from EventTarget, instead of being distinct functions. (ExE-Boss)
  • Fixed errors that would occur if attempting to use a DOM object, such as a custom element, as an argument to addEventListener.
  • Fixed errors that would occur when closing a window with outstanding requests to data: URLs.
  • Fixed sporadic issues with the value of <input type="month"> that could occur in some time zones and for some times.
  • Fixed document.implementation.createDocument() to return an XMLDocument, instead of a Document. (ExE-Boss)
  • Fixed running jsdom in a browser to detect globals more reliably. (ExE-Boss)

v16.2.1

Compare Source

  • Updated saxes, to bring in some BOM-related fixes.
  • Updated Acorn-related packages to squelch npm audit warnings.

v16.2.0

Compare Source

  • Added support for custom elements! Congratulations and thanks to @​pmdartus for making this happen, after ten months of hard work and lots of effort poured into the complex architectural prerequisites in jsdom and supporting packages.
  • Fixed some issues when trying to use Attr as a Node, e.g. by checking its baseURI property or calling attr.cloneNode().
  • Fixed a memory leak during parsing that was introduced in v14.0.0.
  • Fixed edge cases in number/string conversion used for certain element properties that reflected integer attributes.

v16.1.0

Compare Source

  • Added console.timeLog().
  • Changed Attr to extend Node, to align with specifications. (ExE-Boss)
  • Changed <noscript> children to be parsed as nodes, instead of as text, when runScripts is left as the default of undefined. (ACHP)
  • Upgraded cssstyle to v2.1.0, which brings along fixes to handling of rgba() and hsl() colors. (kraynel)
  • Fixed some selection-related issues when manipulating the value of <input>s and <textarea>s. (Matthew-Goldberg)
  • Fixed various issues with setTimeout(), setInterval(), and requestAnimationFrame(), particularly around window closing and recursive calls.

v16.0.1

Compare Source

  • Fixed Node v10 and v11 support when runScripts was set.
  • Fixed the behavior when changing an <input>'s type="" attribute.
  • Fixed input validation behavior for <input type="range"> when max="" is less than min="".

v16.0.0

Compare Source

For this release we'd like to welcome @​pmdartus to the core team. Among other work, he's driven the heroic effort of constructor prototype and reform in jsdom and its dependencies over the last few months, to allow us to move away from shared constructors and prototypes, and set the groundwork for custom elements support (coming soon!).

Breaking changes:

  • Node v10 is now the minimum supported version.
  • The dom.runVMScript() API has been replaced with the more general dom.getInternalVMContext() API.
  • Each jsdom Window now creates new instances of all the web platform globals. That is, our old shared constructor and prototypes caveat is no longer in play.
  • Each jsdom Window now exposes all JavaScript-spec-defined globals uniformly. When runScripts is disabled, it exposes them as aliases of the ones from the outer Node.js environment. Whereas when runScripts is enabled, it exposes fresh copies of each global from the new scripting environment. (Previously, a few typed array classes would always be aliased, and with runScripts disabled, the other classes would not be exposed at all.)

Other changes:

  • Added the AbstractRange, Range, StaticRange, Selection, and window.getSelection() APIs.
  • Added working constructors for Comment, Text, and DocumentFragment.
  • Added valueAsDate, valueAsNumber, stepUp() and stepDown() to <input> elements. (kraynel)
  • Added window.origin.
  • Removed document.origin.
  • Fixed <template> to work correctly inside XML documents.
  • Fixed some bugs which would cause jsdom to choose the wrong character encoding because it was failing to detect <meta charset> or <meta http-equiv="charset"> elements.
  • Fixed input.type to default to "text". (connormeredith)
  • Fixed incorrect validation errors for <input> with fractional values for their step="" attribute. (kontomondo)
  • Fixed incorrect validation errors on readonly <input> elements.
  • Fixed <input type="email" multiple pattern="..."> validation.
  • Fixed fileReader.readAsDataURL() to always base64-encode the result. (ytetsuro)
  • Fixed inserting <img> elements into documents without a browsing context to no longer crash when the canvas package is installed.
  • Fixed a memory leak when using window.setTimeout() or window.setInterval().
  • Improved the performance of getComputedStyle(). (eps1lon)

</details>


Renovate configuration

:date: Schedule: "on the 1st through 7th day of the month" in timezone America/Los_Angeles.

:vertical_traffic_light: Automerge: Disabled by config. Please merge this manually once you are satisfied.

:recycle: Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

:no_bell: Ignore: Close this PR and you won't be reminded about these updates again.


  • [ ] <!-- rebase-check -->If you want to rebase/retry this PR, check this box

This PR has been generated by WhiteSource Renovate. View repository job log here.

+165 -53

0 comment

2 changed files

renovate[bot]

pr closed time in 2 days

PR closed sourcegraph/sourcegraph

Reviewers
Update dependency @octokit/rest to ^16.43.1 bot npm

This PR contains the following updates:

Package Type Update New value References Sourcegraph
@octokit/rest devDependencies minor ^16.43.1 source code search for "@octokit/rest"

Release Notes

<details> <summary>octokit/rest.js</summary>

v16.43.1

Compare Source

Bug Fixes
  • TypeScript: named { Octokit } export (486eadd)

v16.43.0

Compare Source

Features

v16.42.2

Compare Source

Bug Fixes

v16.42.1

Compare Source

Bug Fixes
  • do not throw if using deprecated Octokit default export without options (dec7334)

v16.42.0

Compare Source

Features
  • deprecate const Octokit = require("@&#8203;octokit/rest") in favor of const { Octokit } = require("@&#8203;octokit/rest") (bbc3e48)

v16.41.2

Compare Source

Bug Fixes

v16.41.1

Compare Source

Bug Fixes

v16.41.0

Compare Source

Features
  • typescript: types for latest Action endpoint methods (#​1567) (0758d61)

v16.40.2

Compare Source

Bug Fixes
  • typescript: new auth strategy options (0c7b109)

v16.40.1

Compare Source

Bug Fixes

v16.40.0

Compare Source

Features
  • Actions endpoint methdos (a706b0d)

v16.39.0

Compare Source

Features
  • accessing response.data[namespacekey] in .paginate(options, mapFn) map function now logs a deprecation (3caa24a)

v16.38.3

Compare Source

Bug Fixes

v16.38.2

Compare Source

Bug Fixes
  • everest preview graduated (441c452)
  • add .endpoint() method to all deprecated endpoint methods (9701151)
  • bring back octokit.teams.removeMember() (d439ca7)
  • bring back `octokit.teams.removeMembership() (dadfe9a)
  • correct deprecation message for octokit.apps.* OAuth authorization methods' (5f5e62e)

v16.38.1

Compare Source

Bug Fixes
  • typescript: authStrategy key in Octokit constructor options (fb19f11)

v16.38.0

Compare Source

Features

  • All authentication strategies from @octokit/auth are now supported. Setting auth to an object or a basic authentication string is now deprecated and will be removed in v17 (4573ee2)

v16.37.0

Compare Source

Features
  • .apps.revokeInstallationToken() (ed31137)
  • .migrations.listReposForOrg(), .migrations.listReposForUser() (8777073)
  • .repos.listForOrg()'s type parameter can now be set to "internal" (GitHub Enterprise Cloud only) (9c71d18)
  • visibility parameter for .repos.createForAuthenticatedUser() and .repos.createInOrg() (31b094b)
  • new delete_branch_on_merge parameter for .repos.createForAuthenticatedUser(), .repos.createInOrg(), .repos.update() (c783249)
  • new parameters for .orgs.update(): members_can_create_internal_repositories, members_can_create_private_repositories, members_can_create_public_repositories (7be5174)
  • new parameters for .repos.updateBranchProtection(): allow_deletions, allow_force_pushes, required_linear_history (c5632ff)
  • rename & deprecation of endpoint methods using /teams/:team_id* path internally. The are being replaced by *ForOrg methods that use /orgs/:org/teams/:team_slug* internally. With the next breaking version, the *ForOrg suffix will be removed using another deprecation. (faecc99)
Bug Fixes
  • graduated previews removed: barred-rock, echo, hagar (53b8021)

</details>


Renovate configuration

:date: Schedule: "on the 1st through 7th day of the month" in timezone America/Los_Angeles.

:vertical_traffic_light: Automerge: Disabled by config. Please merge this manually once you are satisfied.

:recycle: Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

:no_bell: Ignore: Close this PR and you won't be reminded about this update again.


  • [ ] <!-- rebase-check -->If you want to rebase/retry this PR, check this box

This PR has been generated by WhiteSource Renovate. View repository job log here.

+48 -10

1 comment

2 changed files

renovate[bot]

pr closed time in 2 days

PR closed sourcegraph/sourcegraph

Update alpine:3.10 Docker digest to de78803 bot

This PR contains the following updates:

Package Type Update New value Sourcegraph
alpine final digest 3.10 code search for "alpine"
alpine stage digest 3.10 code search for "alpine"

Renovate configuration

:date: Schedule: "on the 1st through 7th day of the month" in timezone America/Los_Angeles.

:vertical_traffic_light: Automerge: Disabled by config. Please merge this manually once you are satisfied.

:recycle: Rebasing: Whenever PR becomes conflicted, or you tick the rebase/retry checkbox.

:no_bell: Ignore: Close this PR and you won't be reminded about these updates again.


  • [ ] <!-- rebase-check -->If you want to rebase/retry this PR, check this box

This PR has been generated by WhiteSource Renovate. View repository job log here.

+9 -9

1 comment

5 changed files

renovate[bot]

pr closed time in 2 days

PR closed sourcegraph/sourcegraph

Update wrouesnel/postgres_exporter Docker tag to v0.8.0 bot

This PR contains the following updates:

Package Type Update New value Sourcegraph
wrouesnel/postgres_exporter final minor v0.8.0 code search for "wrouesnel/postgres_exporter"

Renovate configuration

:date: Schedule: "on the 1st through 7th day of the month" in timezone America/Los_Angeles.

:vertical_traffic_light: Automerge: Disabled by config. Please merge this manually once you are satisfied.

:recycle: Rebasing: Whenever PR becomes conflicted, or if you tick the rebase/retry checkbox below.

:no_bell: Ignore: Close this PR and you won't be reminded about this update again.


  • [ ] <!-- rebase-check -->If you want to rebase/retry this PR, check this box

This PR has been generated by WhiteSource Renovate. View repository job log here.

+1 -1

1 comment

1 changed file

renovate[bot]

pr closed time in 2 days

issue closedsourcegraph/sourcegraph

Rename default branch to `main`

We are going to do this for reasons mentioned in this thread: https://twitter.com/mislav/status/1270388510684598272

More context in https://github.com/sourcegraph/sourcegraph/issues/11389#issuecomment-642877899

closed time in 2 days

nicksnyder

PR closed sourcegraph/sourcegraph

Reviewers
Update Yarn to ^1.21.0 bot

This PR contains the following updates:

Package Type Update New value References Sourcegraph
yarn engines minor ^1.21.0 source code search for "yarn"

Release Notes

<details> <summary>yarnpkg/yarn</summary>

v1.21.0

Compare Source

v1.19.2

Compare Source

  • Folders like .cache won't be pruned from the node_modules after each install.

    #​7699 - Maël Nison

  • Correctly installs workspace child dependencies when workspace child not symlinked to root.

    #​7289 - Daniel Tschinder

  • Makes running scripts with Plug'n Play possible on node 13.

    #​7650 - Sander Verweij

  • Change run command to check cwd/node_modules/.bin for commands. Fixes run in workspaces.

    #​7151 - Jeff Valore

v1.19.1

Compare Source

Important: This release contains a cache bump. It will cause the very first install following the upgrade to take slightly more time, especially if you don't use the Offline Mirror feature. After that everything will be back to normal.

v1.19.0

Compare Source

Important: This release contains a cache bump. It will cause the very first install following the upgrade to take slightly more time, especially if you don't use the Offline Mirror feature. After that everything will be back to normal.

  • Fixes a potential vulnerability regarding how the build artifacts are stored

    Reported by ChALkeR, fixed by Maël Nison

v1.18.0

Compare Source

v1.17.3

Compare Source

v1.17.2

Compare Source

v1.17.1

Compare Source

v1.17.0

Compare Source

v1.16.0

Compare Source

v1.15.2

Compare Source

The 1.15.1 doesn't exist due to a release hiccup.

  • Reverts a behavior causing boggus interactions between PowerShell and yarn global

    #​6954 - briman0094

  • Fixes a bug where non-zero exit codes were converted to a generic 1 when running yarn run

    #​6926 - Kyle Fang

  • Fixes production / development reporting when running yarn audit

    #​6970 - Adam Richardson

v1.15.1

Compare Source

v1.15.0

Compare Source

v1.14.0

Compare Source

v1.13.0

Compare Source

v1.12.3

Compare Source

Important: This release contains a cache bump. It will cause the very first install following the upgrade to take slightly more time, especially if you don't use the Offline Mirror feature. After that everything will be back to normal.

v1.12.1

Compare Source

v1.12.0

Compare Source

v1.11.1

Compare Source

v1.11.0

Compare Source

</details>


Renovate configuration

:date: Schedule: "on the 1st through 7th day of the month" in timezone America/Los_Angeles.

:vertical_traffic_light: Automerge: Disabled by config. Please merge this manually once you are satisfied.

:recycle: Rebasing: Whenever PR becomes conflicted, or if you modify the PR title to begin with "rebase!".

:no_bell: Ignore: Close this PR and you won't be reminded about this update again.


  • [x] <!-- renovate-rebase -->If you want to rebase/retry this PR, check this box

Newsflash: Renovate has joined WhiteSource, and is now free for all use. Learn more or view updated terms and privacy policies.

+370 -387

1 comment

16 changed files

renovate[bot]

pr closed time in 2 days

issue closedsourcegraph/sourcegraph

Indexing multiple branches timing out with git command `failed: exit status 128`

Reported by https://app.hubspot.com/contacts/2762526/company/768958891

See Slack thread at https://sourcegraph.slack.com/archives/CTJCMDCCS/p1595352999461200?thread_ts=1595350296.457900&cid=CTJCMDCCS

2020/07/21 17:35:51 error indexing <$REPO>@HEAD=<$SHA_BRANCH_LIST>: command [git -C /data/index/.indexserver.tmp/tmpfriend-6-554088670/<$REPO.git> -c protocol.version=2 fetch --depth=1 http://sourcegraph-frontend-internal/.internal/git/<$REPO> <$SHA LIST>] failed: exit status 128
OUT:
ERR: warning: redirecting to http://gitserver-0.gitserver:3178/git/<$REPO>/
error: RPC failed; HTTP 500 curl 22 The requested URL returned error: 500
fatal: the remote end hung up unexpectedly

When the user reduced the number of branches from double digits to just 1, it worked just fine. He then increased up to 5, and got the same error.

  • Kubernetes
  • Fetching code from Gerritt

closed time in 3 days

dadlerj

issue commentsourcegraph/sourcegraph

Indexing multiple branches timing out with git command `failed: exit status 128`

Haven't really heard back from the customer, nor have I heard this happen to others. I'm going to close this out. Feel free to re-open if we need to investigate more.

dadlerj

comment created time in 3 days

pull request commentsourcegraph/sourcegraph

dev: support switching our default branch to main

FYI this was the script I ran to update all our open PRs to merge into main. It was modified from the action to run locally + use netrc for auth.

#!/bin/sh -l

PREVIOUS_DEFAULT="master"
NEW_DEFAULT="main"
GITHUB_API_URL="https://api.github.com"
GITHUB_REPOSITORY="sourcegraph/sourcegraph"

echo "Migrating from '$PREVIOUS_DEFAULT' to '$NEW_DEFAULT'"
echo "Repo Name: $GITHUB_REPOSITORY"

# Check the API to see what the current default branch is
CURRENT_DEFAULT=$(curl -n --silent $GITHUB_API_URL/repos/$GITHUB_REPOSITORY | jq -r .default_branch)
echo "Default branch: $CURRENT_DEFAULT"

if [ "$CURRENT_DEFAULT" = "$NEW_DEFAULT" ]; then
	echo "Default has changed to $NEW_DEFAULT"
	# Check all existing PRs to see if we should change their base
	PRS=$(curl -n --silent "$GITHUB_API_URL/repos/$GITHUB_REPOSITORY/pulls?base=$PREVIOUS_DEFAULT&state=open")

	# TODO: PATCH /repos/:owner/:repo/pulls/:pull_number
	for row in $(echo $PRS | jq -r 'map(select(.locked == false)) | .[].url'); do
		echo "Attempting to update $row"
		curl -n --silent --show-error -w "Status Code: %{http_code}\n" --request PATCH --data '{"base": "'$NEW_DEFAULT'"}' "$row"
	done

	exit
fi;
keegancsmith

comment created time in 3 days

push eventsourcegraph/sourcegraph

Keegan Carruthers-Smith

commit sha b773625e95073d7b10f1ff706ade2584a0e18772

dev: support switching our default branch to main (#11453) This mostly updates our dev CI/helpers which hardcoded master to instead hardcode main. In the case of CI we treat both main and master as the default branch.

view details

push time in 3 days

push eventsourcegraph/sourcegraph

Keegan Carruthers-Smith

commit sha b773625e95073d7b10f1ff706ade2584a0e18772

dev: support switching our default branch to main (#11453) This mostly updates our dev CI/helpers which hardcoded master to instead hardcode main. In the case of CI we treat both main and master as the default branch.

view details

push time in 3 days

delete branch sourcegraph/sourcegraph

delete branch : k/main

delete time in 3 days

PR merged sourcegraph/sourcegraph

dev: support switching our default branch to main

This mostly updates our dev CI/helpers which hardcoded master to instead hardcode main. In the case of CI we treat both main and master as the default branch.

Part of https://github.com/sourcegraph/sourcegraph/issues/11464

+51 -46

10 comments

21 changed files

keegancsmith

pr closed time in 3 days

push eventsourcegraph/sourcegraph

Farhan Attamimi

commit sha 77e86700638c8e58480056c883fe68997e98034e

Fix: parens get removed from interactive mode queries (#12696)

view details

Erik Seliger

commit sha 2cc30710940e8dddc13fb798a885d34f6d631021

URL-persist changeset filters (#12662) This also refactors the filters into a separate component, which will make it easier to implement the new UI components based on this.

view details

Thorsten Ball

commit sha 4b9f66addac4b9722d7083a4858f17678709fa6c

Implement ApplyCampaign and changeset reconciler (#12435) * Implement ApplyCampaign and changeset reconciler * Undo changes to dev/start.sh * Clean up test script * Run gofmt * Return changeset.FailureMessage in resolver * Fix typo * Small fixes * Update comments * Update dev docs * Rename ChangesetSpecDescription.IsExisting to IsImportingExisting * Implement filtering by Publication/ReconcilerState * Move CheckRepoSupported back to service.go * Regenerate schema and fix frontend build * Move upper/lower conversion of ReconcilerState to store layer * Turn ChangesetSpec.Spec into a pointer * Add tests for changeset reconciler and fix tiny bugs * Test repository permissions handling in ApplyCampaign * Fix typo in variable name * Remove duplicate import * Remove empty file * Add blank line * Move reconciler test script to ./dev * Reset failure message in reconciler * Yes, CI, localhost is okay * Update snapshots * Fix another TS compiler error

view details

Farhan Attamimi

commit sha e79b5f45ce5b863ba62c6b686ee4fcda834dd7c9

Web content styles (#12387)

view details

Eric Fritz

commit sha b737ab2947f4a2d9e23c77eba42696269d4a3fda

codeintel: Index queue client (#12688)

view details

Eric Fritz

commit sha 85c685467999707afc208e77ce49efbc6b41fcf3

codeintel: Index queue API (#12657)

view details

Thorsten Ball

commit sha 2ad211edc9da0e7cf27941d4e67bd3413afc599d

Split up campaigns.Store in multiple files (#12705) While adding `ChangesetSpecs` and `CampaignSpecs` I noticed how much easier it is to work with the store and its tests when you have different files for different tables. So, this PR does _nothing_ besides splitting up the code that was in `store.go` into `store.go`, `store_campaigns.go`, `store_changeset_events.go` and `store_changesets.go`.

view details

Thorsten Ball

commit sha 67afca6be2afb716aa48d4601b6fda08dc5e030c

Various small fixes for the campaigns.Store (#12710) * Remove unused Store.TryAcquireAdvisoryLock We don't need this method anymore and if we need support for locks again we can simply use `basestore.Store` which we already use. * Remove debug print * Remove wrapper UpdateChangesets/CreateChangesets (plural) methods Before we loosened the constraints on the `changesets` table these methods did batch-insert and batch-update in a single database query. Since that's not needed (nor that easy) anymore, we can remove these methods, since they only provide syntactic sugar for a few usecases. If we decide we need the syntactic sugar back, we can re-add them. * Fix missing references

view details

Stefan Hengl

commit sha 0346923df298d344d167465e0e1eeae8773f4e95

search: treat * at pattern borders properly (#12697)

view details

Dax McDonald

commit sha 7292502dcb17365c036042db5d8be4404da0f91b

Set env var for reviewdog (#12717)

view details

Dax McDonald

commit sha 12b19de8e8752070590c8f4eba24a1b6b9e3ecba

Update Troubleshooting for nvm versions (#12687)

view details

Thorsten Ball

commit sha dd120b19c146c944ebe69cdf0f23ca2cbba25f9a

Check for publicationState/reconcilerState in permission tests (#12715)

view details

Thorsten Ball

commit sha d88a09359b6d0e62bd65d73f25306d3354a43d84

Fix 'template inlines' CI check (#12703) Looks like this has been broken for a while. Funnily enough: CI didn't break.

view details

Erik Seliger

commit sha 8ba77eedca480704eef549f352f60da0ee50128c

Update outdated note on GitLab support (#12719)

view details

Bunny

commit sha d142f1ef4b99fe2b0de892b72405641f75865a4d

Updated the logo (#12540) * Updated the logo * New-new logo from Victoria

view details

Felix Becker

commit sha 75188c68d0d68f0f958be7a611641debe4d8d449

Remove prepublish script (#12702)

view details

Adam Harvey

commit sha d474539c90befcb1851c315f3bb7032c75a37d5b

campaigns: add GitLab webhook support (#12139) Co-authored-by: Erik Seliger <erikseliger@me.com>

view details

TJ Kandala

commit sha 53339e93a17b7934abf3bc4aae3565c15a0631a9

Stop word-count extension from breaking extensions and command palette (#12573) * swallow errors for all types of extension contributions

view details

Eric Fritz

commit sha 948c70bc63dca2818242f04ee1634f7729d9ca17

codeintel: Internal API proxy (#12691)

view details

Rijnard van Tonder

commit sha 8b3e06bfb4af0b5c0881b7a490af3cd2a4c75d50

web: highlight search operators and/or/not (#12694)

view details

push time in 3 days

push eventsourcegraph/sourcegraph

Farhan Attamimi

commit sha 77e86700638c8e58480056c883fe68997e98034e

Fix: parens get removed from interactive mode queries (#12696)

view details

Erik Seliger

commit sha 2cc30710940e8dddc13fb798a885d34f6d631021

URL-persist changeset filters (#12662) This also refactors the filters into a separate component, which will make it easier to implement the new UI components based on this.

view details

Thorsten Ball

commit sha 4b9f66addac4b9722d7083a4858f17678709fa6c

Implement ApplyCampaign and changeset reconciler (#12435) * Implement ApplyCampaign and changeset reconciler * Undo changes to dev/start.sh * Clean up test script * Run gofmt * Return changeset.FailureMessage in resolver * Fix typo * Small fixes * Update comments * Update dev docs * Rename ChangesetSpecDescription.IsExisting to IsImportingExisting * Implement filtering by Publication/ReconcilerState * Move CheckRepoSupported back to service.go * Regenerate schema and fix frontend build * Move upper/lower conversion of ReconcilerState to store layer * Turn ChangesetSpec.Spec into a pointer * Add tests for changeset reconciler and fix tiny bugs * Test repository permissions handling in ApplyCampaign * Fix typo in variable name * Remove duplicate import * Remove empty file * Add blank line * Move reconciler test script to ./dev * Reset failure message in reconciler * Yes, CI, localhost is okay * Update snapshots * Fix another TS compiler error

view details

Farhan Attamimi

commit sha e79b5f45ce5b863ba62c6b686ee4fcda834dd7c9

Web content styles (#12387)

view details

Eric Fritz

commit sha b737ab2947f4a2d9e23c77eba42696269d4a3fda

codeintel: Index queue client (#12688)

view details

Eric Fritz

commit sha 85c685467999707afc208e77ce49efbc6b41fcf3

codeintel: Index queue API (#12657)

view details

Thorsten Ball

commit sha 2ad211edc9da0e7cf27941d4e67bd3413afc599d

Split up campaigns.Store in multiple files (#12705) While adding `ChangesetSpecs` and `CampaignSpecs` I noticed how much easier it is to work with the store and its tests when you have different files for different tables. So, this PR does _nothing_ besides splitting up the code that was in `store.go` into `store.go`, `store_campaigns.go`, `store_changeset_events.go` and `store_changesets.go`.

view details

Thorsten Ball

commit sha 67afca6be2afb716aa48d4601b6fda08dc5e030c

Various small fixes for the campaigns.Store (#12710) * Remove unused Store.TryAcquireAdvisoryLock We don't need this method anymore and if we need support for locks again we can simply use `basestore.Store` which we already use. * Remove debug print * Remove wrapper UpdateChangesets/CreateChangesets (plural) methods Before we loosened the constraints on the `changesets` table these methods did batch-insert and batch-update in a single database query. Since that's not needed (nor that easy) anymore, we can remove these methods, since they only provide syntactic sugar for a few usecases. If we decide we need the syntactic sugar back, we can re-add them. * Fix missing references

view details

Stefan Hengl

commit sha 0346923df298d344d167465e0e1eeae8773f4e95

search: treat * at pattern borders properly (#12697)

view details

Dax McDonald

commit sha 7292502dcb17365c036042db5d8be4404da0f91b

Set env var for reviewdog (#12717)

view details

Dax McDonald

commit sha 12b19de8e8752070590c8f4eba24a1b6b9e3ecba

Update Troubleshooting for nvm versions (#12687)

view details

Thorsten Ball

commit sha dd120b19c146c944ebe69cdf0f23ca2cbba25f9a

Check for publicationState/reconcilerState in permission tests (#12715)

view details

Thorsten Ball

commit sha d88a09359b6d0e62bd65d73f25306d3354a43d84

Fix 'template inlines' CI check (#12703) Looks like this has been broken for a while. Funnily enough: CI didn't break.

view details

Erik Seliger

commit sha 8ba77eedca480704eef549f352f60da0ee50128c

Update outdated note on GitLab support (#12719)

view details

Bunny

commit sha d142f1ef4b99fe2b0de892b72405641f75865a4d

Updated the logo (#12540) * Updated the logo * New-new logo from Victoria

view details

Felix Becker

commit sha 75188c68d0d68f0f958be7a611641debe4d8d449

Remove prepublish script (#12702)

view details

Adam Harvey

commit sha d474539c90befcb1851c315f3bb7032c75a37d5b

campaigns: add GitLab webhook support (#12139) Co-authored-by: Erik Seliger <erikseliger@me.com>

view details

TJ Kandala

commit sha 53339e93a17b7934abf3bc4aae3565c15a0631a9

Stop word-count extension from breaking extensions and command palette (#12573) * swallow errors for all types of extension contributions

view details

Eric Fritz

commit sha 948c70bc63dca2818242f04ee1634f7729d9ca17

codeintel: Internal API proxy (#12691)

view details

Rijnard van Tonder

commit sha 8b3e06bfb4af0b5c0881b7a490af3cd2a4c75d50

web: highlight search operators and/or/not (#12694)

view details

push time in 3 days

pull request commentsourcegraph/sourcegraph

search: filter out noisy inputs used in diff testing

Because it sounded like your previous suggestion was to exclude these before logging? Saves piping the noise over the wire. I plan to delete this tomorrow, if you're worried about overhead?

Oops, my bad :P Yeah whatever, its all just logs. I've been finding it useful to look at every search query. Really helps drive home how many more search queries we do for symbols (thanks to code intel) vs interactive input from users.

rvantonder

comment created time in 3 days

issue commentsourcegraph/sourcegraph

Enable globbing for Sourcegraph org members

I think the sort of magic which responds to the results makes things quite complicated to understand. Displaying different results differently makes a lot of sense to me and can be done purely in the code that collects results so that sounds much more reliable.

Another suggestion, which I think I have seen done before. If your glob doesn't contain a / or ** there is an implicit transormation to **G** where G was the original glob. If we have a way to reference root files (/ prefix?) this allows all the same power with the same convenience.

rvantonder

comment created time in 3 days

issue commentsourcegraph/sourcegraph

Search 504s

We don't seem to be getting searcher traces. But I would suspect the issue isn't the speed of "grep", but that we are struggling to fetch an archive of ghdump. Our searcher code doesn't do per line searching, but uses the same technique in rg of searching the whole body and hydrating in line breaks.

nicksnyder

comment created time in 3 days

pull request commentsourcegraph/sourcegraph

search: treat * at pattern borders properly

I'm guessing this is a bit tricky around consistency and convenience. ie I like that I can just do -file:_test.go. But I suppose it needs to be **_test.go?

@rvantonder mentioned a similar example. Right now, we try to be very predictable and consistent. IE -file:_test.go will exclude the file _test.go in the project's root folder, while -file:**_test.go will exclude all test files. Rijnard already wrote down some ideas on how to help users in #12467. I wonder whether it just takes a couple of searches to get used to or whether it is really inconvenient and an entry barrier.

Agreed, if we guide the user that makes a lot of sense. I agree that once you learn to include ** it becomes very natural and much more exact.

stefanhengl

comment created time in 3 days

pull request commentsourcegraph/sourcegraph

dev: support switching our default branch to main

I ended up working late last night so got a late start today. I'm gonna do this tomorrow morning eu-timezones rather.

keegancsmith

comment created time in 3 days

push eventsourcegraph/sourcegraph

Keegan Carruthers-Smith

commit sha 1bf678a80f01138b101810b05ac39aacb1e5caee

Revert "update links to sourcegraph on github" for browser tests This partially reverts commit a9a10c5c4c003c39535b58b2b806ac8a852c64ae.

view details

push time in 3 days

issue commentsourcegraph/sourcegraph

Search 3.19 Tracking issue

Last Week

Main focus was search indexing and performance at scale. I first focussed on a large customer, but then then morphed into search performance for sourcegraph.com. Ended up looking into it after hack sessions with Robert, Tomas, Stefan and Ryan. Biggest win was custom marshalling of the list of repo/branches to search we sent to Zoekt. Quoting my posts on #progress:

Our large scale instance can now index every repo. This makes it possible to do an exhaustive search (eg find accidental commits of passwords) which was a signing hurdle. As an extra bonus the page which tells you which repos need indexing now works (a long time complaint at scale).

Searching all indexed repos on Sourcegraph.com is now ~0.4s faster (about 25% faster overall). When doing an indexed search we transfer a list of repos to search to Zoekt (our indexed searcher). This is a cost we pay every search no matter how many results it finds. We optimized how we create the list and it now takes about 75ms to transfer 100k repos instead of 500ms on sourcegraph.com. More details here sourcegraph#12546. For the most eye melting benchmark results look here zoekt#53.

Shaved off another 350ms with Ryan. Searching on sourcegraph.com across 100k repos now takes less than half a second on average.

See sourcegraph#12557 for how we tackled that. We are nearly at the point where search time is dominated by actual source code searching instead of our infrastructure around zoekt. Stefan and I explored the final slowdown now and will likely try tackle it next week.

A lot of this work also ended required observability to find what to do. So there have been lots of improvements made to logs in zoekt and jaeger tracing around zoekt in the frontend.

Ended up not journaling daily last week, regrets. But was more transparent in #progress so could use that to track high level work.

I ended up not doing my final work for src-expose. Will spend a day on it this week.

This Week

First half of the week is smaller tasks I've been meaning to get to + src-expose. Thursday I'll start work on streaming search. My focus will be purely experimental. I've been meaning to create an alternative graphql/JS free search UI. I'll likely create a dumb UI for this and use it as a playground for streaming work. So that will be my main goal this week.

Explicit non-goal: Over the weekend we doubled the number of indexed repos on Sourcegraph.com (100k -> 200k) leading to performance + OOM issues. I'm going to let other look into that, and instead focus on other work.

Next Week

Monday is a public holiday, so 4 days. Will hopefully have a very fast streaming search PoC that we can than look into integrating into the main product. Depending on where the code takes me I imagine two paths: streaming for searcher or integrating streaming into main search path.

nicksnyder

comment created time in 3 days

push eventsourcegraph/sourcegraph

Asdine El Hrychy

commit sha 6e93c837289ea1d4a51524dec2bfa5af888d985c

Remove cloneInProgress option from the Repositories API (#12560)

view details

TJ Kandala

commit sha 5fa001d0419c82300f0ffdc27d40a67032ead9d0

web: improve performance of enabling/disabling extensions (#12637)

view details

davejrt

commit sha e7e4fa2ba50c7baddfab88458147bc15d7543773

update docs for more verbose watchman instructions (#12670)

view details

Keegan Carruthers-Smith

commit sha c8b20ab2d2bb516e8362fdb8927457252d5d63d4

search: trace zoekt when count or timeout is specified (#12683) I noticed we sometimes didn't get zoekt spans in Jaeger traces. Turns out it's due to the "no context deadline" logic we use when count or timeout fields are specified. This is because we use a new context. We now set the trace on the new context.

view details

Keegan Carruthers-Smith

commit sha 0a19e43b9ea405f2ba4aa54ce7aaf6a39a121bf4

search: specify branch for zoektSearchHEADOnlyFiles (#12674) Structural search expects to only search HEAD (currently). However, it was using a RepoSet which will return results for any branch we have indexed. I believe this would result in potential benign false positives for files to search. We switch to instead passing in the computed RepoBranches (which should be limited to HEAD).

view details

Keegan Carruthers-Smith

commit sha b624dc15923dd9bcf115718844cb844ac94ab4c6

graphqlbackend: audit uses of context.Background (#12686)

view details

Stephen Gutekanst

commit sha 6c9103e9c0028b511fbf276d69ed46c064571dd5

docker-compose: fix permissions in LSIF data migration (#12689) Fixes an issue reported by a customer with @efritz in https://sourcegraph.atlassian.net/jira/servicedesk/projects/SG/queues/custom/1/SG-295 When following this migration: - Bundle-manager is started, which creates the volume directory with `100:101` as expected - lsif-server's data was missing the USER directive _before_ and thus was owned by root (if you missed the 3.13 -> 3.14 migration, https://docs.sourcegraph.com/admin/updates/docker_compose#v3-13-3-14) -In this migration, you copy the data over, so the volume has OK permissions but the data is still owned by root This corrects that case.

view details

Robert Lin

commit sha 3be7ac29dacb1b7b7a2f671c18448a148e7c77da

monitoring: sync dot-com alerts to generator (#12581) Co-authored-by: Stephen Gutekanst <stephen.gutekanst@gmail.com>

view details

Robert Lin

commit sha 28f6e7401798bd3e4d69d9d85d3fe812be7d1e25

monitoring: remove no_results__suggest_quotes from alert (#12693)

view details

Stefan Hengl

commit sha 973a41cfbad32c069551d54d75a1d4ea1bff7e00

search: fuzzify regex patterns if globbing is active (#12659)

view details

Loïc Guychard

commit sha 08659d63a2fa946ebe5428dd13325e7dfda47165

Fix patternType addition in SearchResults (#12698) By calling `convertPlainTextToInteractiveQuery()`, SearchResults did quite a bit more than simply setting `patternType` in the URL if it didn't exist, mutating the query content needlessly.

view details

Keegan Carruthers-Smith

commit sha f4dd654cd7075fbe6e2df43be60f3495e147151c

Merge remote-tracking branch 'origin/master' into k/main

view details

push time in 3 days

push eventsourcegraph/sourcegraph

TJ Kandala

commit sha 5fa001d0419c82300f0ffdc27d40a67032ead9d0

web: improve performance of enabling/disabling extensions (#12637)

view details

davejrt

commit sha e7e4fa2ba50c7baddfab88458147bc15d7543773

update docs for more verbose watchman instructions (#12670)

view details

Keegan Carruthers-Smith

commit sha c8b20ab2d2bb516e8362fdb8927457252d5d63d4

search: trace zoekt when count or timeout is specified (#12683) I noticed we sometimes didn't get zoekt spans in Jaeger traces. Turns out it's due to the "no context deadline" logic we use when count or timeout fields are specified. This is because we use a new context. We now set the trace on the new context.

view details

Keegan Carruthers-Smith

commit sha 0a19e43b9ea405f2ba4aa54ce7aaf6a39a121bf4

search: specify branch for zoektSearchHEADOnlyFiles (#12674) Structural search expects to only search HEAD (currently). However, it was using a RepoSet which will return results for any branch we have indexed. I believe this would result in potential benign false positives for files to search. We switch to instead passing in the computed RepoBranches (which should be limited to HEAD).

view details

Keegan Carruthers-Smith

commit sha b624dc15923dd9bcf115718844cb844ac94ab4c6

graphqlbackend: audit uses of context.Background (#12686)

view details

Stephen Gutekanst

commit sha 6c9103e9c0028b511fbf276d69ed46c064571dd5

docker-compose: fix permissions in LSIF data migration (#12689) Fixes an issue reported by a customer with @efritz in https://sourcegraph.atlassian.net/jira/servicedesk/projects/SG/queues/custom/1/SG-295 When following this migration: - Bundle-manager is started, which creates the volume directory with `100:101` as expected - lsif-server's data was missing the USER directive _before_ and thus was owned by root (if you missed the 3.13 -> 3.14 migration, https://docs.sourcegraph.com/admin/updates/docker_compose#v3-13-3-14) -In this migration, you copy the data over, so the volume has OK permissions but the data is still owned by root This corrects that case.

view details

Robert Lin

commit sha 3be7ac29dacb1b7b7a2f671c18448a148e7c77da

monitoring: sync dot-com alerts to generator (#12581) Co-authored-by: Stephen Gutekanst <stephen.gutekanst@gmail.com>

view details

Robert Lin

commit sha 28f6e7401798bd3e4d69d9d85d3fe812be7d1e25

monitoring: remove no_results__suggest_quotes from alert (#12693)

view details

Stefan Hengl

commit sha 973a41cfbad32c069551d54d75a1d4ea1bff7e00

search: fuzzify regex patterns if globbing is active (#12659)

view details

Loïc Guychard

commit sha 08659d63a2fa946ebe5428dd13325e7dfda47165

Fix patternType addition in SearchResults (#12698) By calling `convertPlainTextToInteractiveQuery()`, SearchResults did quite a bit more than simply setting `patternType` in the URL if it didn't exist, mutating the query content needlessly.

view details

push time in 3 days

delete branch sourcegraph/sourcegraph

delete branch : k/wat

delete time in 4 days

push eventsourcegraph/sourcegraph

Keegan Carruthers-Smith

commit sha b624dc15923dd9bcf115718844cb844ac94ab4c6

graphqlbackend: audit uses of context.Background (#12686)

view details

push time in 4 days

PR merged sourcegraph/sourcegraph

graphqlbackend: audit uses of context.Background

Looked at the places we used context.Background in graphqlbackend. Pretty much all are invalid.

+13 -20

1 comment

4 changed files

keegancsmith

pr closed time in 4 days

pull request commentsourcegraph/sourcegraph

graphqlbackend: audit uses of context.Background

merging this now since a commit I landed earlier has an occasional race condition fixed in this PR.

keegancsmith

comment created time in 4 days

push eventsourcegraph/sourcegraph

Keegan Carruthers-Smith

commit sha b8c9cc2bf6e2f4e3ba2b176681b4998b00159b64

graphqlbackend: fix race in contextWithoutDeadline

view details

push time in 4 days

PR opened sourcegraph/sourcegraph

graphqlbackend: audit uses of context.Background

Looked at the places we used context.Background in graphqlbackend. Pretty much all are invalid.

+7 -15

0 comment

3 changed files

pr created time in 4 days

create barnchsourcegraph/sourcegraph

branch : k/wat

created branch time in 4 days

issue commentsourcegraph/sourcegraph

Sourcegraph not recognizing that GitHub Enterprise repos are archived

Seeing the output of the rest repository listing API may also help us. We don't list specific repositories, but rather list via the repositories endpoint. The customer seems to be using repositoryQuery: public, so this should be somewhat equivalent to what we actually do:

curl \
  -H "Authorization: token token3" \
  -H "Accept: application/vnd.github.jean-grey-preview+json,application/vnd.github.mercy-preview+json" \
  -H "Accept: application/vnd.github.machine-man-preview+json" \
  'https://github.example.com/api/v3/repositories?per_page=100' | jq .

Added the extra headers since we include them: https://github.com/sourcegraph/sourcegraph/blob/0a19e43b9ea405f2ba4aa54ce7aaf6a39a121bf4/internal/extsvc/github/client.go#L257

dadlerj

comment created time in 4 days

issue commentsourcegraph/sourcegraph

Sourcegraph not recognizing that GitHub Enterprise repos are archived

I took a quick look at this, I'm not sure why it isn't working. Most likely reason is to me is still syncing isn't running for some reason, but it could be something else. @unknwon do you have bandwidth to dive deeper?

In the thread the updated_at time is 2020-06-08 18:01:39.318373+00 which indicates this repo hasn't changed in a while. But we don't update unless we think it has changed. I investigated the following and came up short:

  1. We correctly unmarshal rest archived. I tested with a mini-program using the structs we unmarshal with against the user supplied example response.
  2. We detect archived changing in the syncer. https://github.com/sourcegraph/sourcegraph/blob/0a19e43b9ea405f2ba4aa54ce7aaf6a39a121bf4/cmd/repo-updater/repos/types.go#L627
  3. We handle fork the same way we handle archived in store.go

I wish we had a simple way to tell the last time a full sync ran, I don't think we log that. This is a somewhat hacky way to find when we last updated a repo when it wasn't an insert. (we special case inserting so syncing could still be broken even though new repos are added):

select max(updated_at) from repo where updated_at = created_at;

I am guessing for the archived repo including all sql columns may help debugging.

Mini-program to test unmarshalling:

<details>

package main

import (
	"encoding/json"
	"fmt"
	"io/ioutil"
	"log"
)

type restRepositoryPermissions struct {
	Admin bool `json:"admin"`
	Push  bool `json:"push"`
	Pull  bool `json:"pull"`
}

type restRepository struct {
	ID          string `json:"node_id"` // GraphQL ID
	DatabaseID  int64  `json:"id"`
	FullName    string `json:"full_name"` // same as nameWithOwner
	Description string
	HTMLURL     string `json:"html_url"` // web URL
	Private     bool
	Fork        bool
	Archived    bool
	Permissions restRepositoryPermissions `json:"permissions"`
}

func main() {
	b, err := ioutil.ReadFile("/tmp/foo.json")
	if err != nil {
		log.Fatal(err)
	}
	var r restRepository
	if err := json.Unmarshal(b, &r); err != nil {
		log.Fatal(err)
	}
	fmt.Printf("%#+v\n", r)
	fmt.Println(r.Archived)
}

</details>

dadlerj

comment created time in 4 days

push eventsourcegraph/about

Keegan Carruthers-Smith

commit sha 93a9f55faf9efc2d292bf50e2cc7f75ab3079764

engineering: initial page on honeycomb (#1332)

view details

push time in 4 days

delete branch sourcegraph/about

delete branch : honeycomb

delete time in 4 days

push eventsourcegraph/sourcegraph

Keegan Carruthers-Smith

commit sha 0a19e43b9ea405f2ba4aa54ce7aaf6a39a121bf4

search: specify branch for zoektSearchHEADOnlyFiles (#12674) Structural search expects to only search HEAD (currently). However, it was using a RepoSet which will return results for any branch we have indexed. I believe this would result in potential benign false positives for files to search. We switch to instead passing in the computed RepoBranches (which should be limited to HEAD).

view details

push time in 4 days

delete branch sourcegraph/sourcegraph

delete branch : k/structural-search

delete time in 4 days

PR merged sourcegraph/sourcegraph

search: specify branch for zoektSearchHEADOnlyFiles

Structural search expects to only search HEAD (currently). However, it was using a RepoSet which will return results for any branch we have indexed. I believe this would result in potential benign false positives for files to search. We switch to instead passing in the computed RepoBranches (which should be limited to HEAD).

Part of https://github.com/sourcegraph/sourcegraph/issues/11902

+9 -14

0 comment

2 changed files

keegancsmith

pr closed time in 4 days

push eventsourcegraph/sourcegraph

Keegan Carruthers-Smith

commit sha c8b20ab2d2bb516e8362fdb8927457252d5d63d4

search: trace zoekt when count or timeout is specified (#12683) I noticed we sometimes didn't get zoekt spans in Jaeger traces. Turns out it's due to the "no context deadline" logic we use when count or timeout fields are specified. This is because we use a new context. We now set the trace on the new context.

view details

push time in 4 days

delete branch sourcegraph/sourcegraph

delete branch : k/zoekt-trace

delete time in 4 days

PR merged sourcegraph/sourcegraph

search: trace zoekt when count or timeout is specified

I noticed we sometimes didn't get zoekt spans in Jaeger traces. Turns out it's due to the "no context deadline" logic we use when count or timeout fields are specified. This is because we use a new context. We now set the trace on the new context.

+26 -9

0 comment

1 changed file

keegancsmith

pr closed time in 4 days

push eventsourcegraph/zoekt

Keegan Carruthers-Smith

commit sha 18a8fad5570645b4c0500210d24cc957f5f27ed8

rpc: do not set timeout if MaxWallTime is specified Sourcegraph always sets MaxWallTime, so we should let Zoekt respect that. The MaxWallTime can be greater than 10s, leading to Zoekt returning at 10s instead of MaxWallTime. We now only set a timeout on ctx if MaxWallTime is unset (to protect against accidently not setting in the future). Additionally we update the default timeout to 20s (same as Sourcegraph). Change-Id: I42c4968b58482cc3df8d22337a8443b0aec927c9

view details

push time in 4 days

PR opened sourcegraph/sourcegraph

Reviewers
search: trace zoekt when count or timeout is specified

I noticed we sometimes didn't get zoekt spans in Jaeger traces. Turns out it's due to the "no context deadline" logic we use when count or timeout fields are specified. This is because we use a new context. We now set the trace on the new context.

+26 -9

0 comment

1 changed file

pr created time in 4 days

create barnchsourcegraph/sourcegraph

branch : k/zoekt-trace

created branch time in 4 days

PR opened sourcegraph/sourcegraph

Reviewers
search: specify branch for zoektSearchHEADOnlyFiles

Structural search expects to only search HEAD (currently). However, it was using a RepoSet which will return results for any branch we have indexed. I believe this would result in potential benign false positives for files to search. We switch to instead passing in the computed RepoBranches (which should be limited to HEAD).

Part of https://github.com/sourcegraph/sourcegraph/issues/11902

+9 -14

0 comment

2 changed files

pr created time in 4 days

create barnchsourcegraph/sourcegraph

branch : k/structural-search

created branch time in 4 days

PR opened sourcegraph/about

Reviewers
engineering: initial page on honeycomb
+13 -0

0 comment

2 changed files

pr created time in 4 days

create barnchsourcegraph/about

branch : honeycomb

created branch time in 4 days

pull request commentsourcegraph/sourcegraph

dev: support switching our default branch to main

Yes, I'll update the plan above to limit pushes to master to just me :) Any other admin can add themselves to that list if needs be.

keegancsmith

comment created time in 4 days

pull request commentsourcegraph/sourcegraph

dev: support switching our default branch to main

Plan for tomorrow morning:

  1. Merge this.
  2. Push latest master to main.
  3. Update GitHub settings to make main the default.
  4. Mark main branch as protected.
  5. Run this script to update open PRs to point to main.
  6. Announce in #dev-announce: Action Required: Our default branch has been renamed from master to main. Please update your local clones by running git fetch && git checkout origin/main && ./dev/2020-08-update-default-branch-to-main.sh. Report any issues you find in #dev-chat.
  7. I'll look for things that may have broke.
  8. In a day or two I'll look for any accidental commits to master and rebase and push to main.

There is no step 9. delete master. I think I'll leave it there since it doesn't hurt to keep around for old links.

keegancsmith

comment created time in 4 days

push eventsourcegraph/sourcegraph

Keegan Carruthers-Smith

commit sha e62971770f8de600d3bcbac3a960886658b81281

dev: script to update default branch to main

view details

push time in 4 days

create barnchsourcegraph/sourcegraph

branch : main

created branch time in 4 days

push eventsourcegraph/sourcegraph

Keegan Carruthers-Smith

commit sha dfc21b971c93a1252e0118adf70116e1e5871b60

update new references to master in ci pipeline

view details

push time in 4 days

push eventsourcegraph/sourcegraph

Ryan Slade

commit sha a98e5e8242c70d5bc8f3a8b95c3445fc2fba970a

github: Move rate limiting into client (#11441) * github: Move rate limiting into client * Fix imports

view details

Keegan Carruthers-Smith

commit sha 50a1633f136e3ace68fd366630aac24b80a3b65f

search: reuse commit and repo resolver (#11458) We do not need to use the repository resolver cache here since we only are searching one repository. Additionally we can use a single commit resolver.

view details

ᴜɴᴋɴᴡᴏɴ

commit sha 70ebc09f43e4f1051361a90e559a59e6d992dcfb

e2e: add tests for `auth.userOrgMap` (#11435) Co-authored-by: Ryan Slade <ryanslade@gmail.com>

view details

Thorsten Ball

commit sha 797c5f20e8cb8ca7ad3b6f141b433e4998b10382

Reject campaign update if lacking permission for any of new patches (#11407) * Reject campaign update if lacking permission for any of new patches Previously we simply skipped the patches and didn't publish them if the user lost permissions for a repo in between creating the patchset and updating the campaign. That could lead to an unfixable state where a campaign had a patch and a changeset for the same repository. We decided that the best solution is to simply return an error and reject the update. * Update enterprise/internal/campaigns/service.go Co-authored-by: Erik Seliger <erikseliger@me.com> * Update enterprise/internal/campaigns/service.go Co-authored-by: Ryan Slade <ryanslade@gmail.com> Co-authored-by: Erik Seliger <erikseliger@me.com> Co-authored-by: Ryan Slade <ryanslade@gmail.com>

view details

Ryan Slade

commit sha 59520fa2e3a484babfe2ee5afd4e20537e236617

github: Make RateLimitMonitor private (#11455) But continue to expose it via a method. This will allow us to more easily mock the client.

view details

Ryan Slade

commit sha 0efd2552c19091a9b44a1c67964f38a42bd353a2

bitbucket: Move rate limiting into client (#11459) * bitbucket: Move rate limiting into client This change causes us to fetch the rate limiter from the global registry inside of the Bitbucket client. As a side effect, we no longer need to pass rate limiters around when using ChangesetSources so a lot of that code was cleaned up. * Remove RateLimitRegistry from SyncRegistry * Cleanup imports

view details

Ryan Slade

commit sha 5463179f24dcc6d0edfe00a5ea09eaa37e06ce2a

bitbucketcloud: Add self imposed rate limiting (#11463) This change adds self imposed rate limiting to the BitbucketCloud client. It is currently only enabled in repo-updater

view details

Ryan Slade

commit sha 16c73ea4bba140252d0c4f6e25d64e89c96425f8

gitlab: Add self imposed rate limiting (#11462) * gitlab: Add self imposed rate limiting * gitlab: Make rate limit monitor private

view details

Keegan Carruthers-Smith

commit sha 07ed1a324866e8d5f14262f9a9670cee10ca385d

search: use exact branch matching for zoekt symbol search (#11445) We introduced RepoBranches to allow us to specify exact branches to search. Previously the use of Branch could match more branches since the Pattern will match any indexed branch name containing it.

view details

Rijnard van Tonder

commit sha 36d8c535d5bbe1906b06de2c5ddfd51aa6e1f8dc

search: fix caps on annotation variable (#11465)

view details

Beyang Liu

commit sha 359ded7d66424e05f71cfb94cc7640c3d7391d55

dev: fix enterprise build on Linux (#11451)

view details

Felix Becker

commit sha e02dd115495b95428b51662b631a3efa8778dfac

Upload ESLint warnings as LSIF diagnostics (#11469)

view details

uwedeportivo

commit sha 7cda181584ed9b98f3f7f334b241600a18c5909c

ghe feeder: fix NULL string in errType (#11124) * ghe feeder: fix NULL string in errType * counting bug fixes

view details

Robert Lin

commit sha 8451bb7fa365b3d406fe51c6d447085978a7c322

monitoring: configure alert notifications from site config (#11427)

view details

Robert Lin

commit sha 4e9fddd6fd0d99576532b04f54f284b072bd2139

Revert #11427 (#11479) This reverts commit 8451bb7fa365b3d406fe51c6d447085978a7c322.

view details

Robert Lin

commit sha 601a1fecc568f2818f60e30efa9f7dfc4eb0bbed

build: fix version embeds (#11484)

view details

Robert Lin

commit sha 42669604ca46090791085c4539b9e471a0a76e38

monitoring: automatically deploy grafana notifiers (round 2) (#11483) * monitoring: configure alert notifications from site config (#11427) * never contribute warning if no notifiers are set * check for grafana-wrapper status code * return specific problem for observability alerts on sourcegraph/server * rename GRAFANA_ to GRAFANA_INTERNAL_, add more docs * indicate 3370 is a useful port with EXPOSE * fix notifier schema * improve changelog wording Co-authored-by: Stephen Gutekanst <stephen.gutekanst@gmail.com>

view details

Robert Lin

commit sha 7442c3a0dd71fad0c9c0b493192e533027e154fa

dev: silence grafana build output (#11486)

view details

Ryan Slade

commit sha 57358eca66f8919b276c1676b1d45168125878ca

ratelimit: Normalize url in ratelimit package (#11487) This removes the dependency on the extsvc package in a few places and also doesn't couple extsvc.NormalizeBaseURL to the ratelimit implementation.

view details

Thorsten Ball

commit sha ecf2b5cc3ff73cb03528775844aae94f7eeab511

Skip TestReaderCacheConcurrentRequests due to race condition (#11489)

view details

push time in 4 days

delete branch sourcegraph/sourcegraph

delete branch : k/tracer-url

delete time in 4 days

push eventsourcegraph/sourcegraph

Keegan Carruthers-Smith

commit sha e1fc002ca73135cf2b7c00e7dc8fd295208e700f

tracer: set URL to trace in x-trace header (#12656) This is what we did for lightstep and it is quite convenient. We couldn't do this when we originally added Jaeger since we didn't proxy it. However, now that we do we can provide a direct link.

view details

push time in 4 days

PR merged sourcegraph/sourcegraph

Reviewers
tracer: set URL to trace in x-trace header

This is what we did for lightstep and it is quite convenient. We couldn't do this when we originally added Jaeger since we didn't proxy it. However, now that we do we can provide a direct link.

+19 -12

0 comment

1 changed file

keegancsmith

pr closed time in 4 days

push eventsourcegraph/sourcegraph

Stefan Hengl

commit sha d1103e0e3f83e89bcaf99044742f559a88fb68a0

web: include * and ] in trigger characters (#12660)

view details

Eric Fritz

commit sha ce3163cd28992aa63c5c5e8d63b96e0ee3171717

workerutil: Add DequeueWithIndependentTransactionContext to store (#12661)

view details

Eric Fritz

commit sha c2214a83c3fe75fc9bfc2a7e895741097ea15127

workerutil: Move store into own package (#12663)

view details

davejrt

commit sha 9d471d099db2b99d500957bb67ac5fa70c1ce623

don't enforce bash 5 on linux (#12592) * Change to bash 4. Update error for Darwin and Linux

view details

Keegan Carruthers-Smith

commit sha 7534968805dd883738c2c0b8a17d3a08875b1c1a

tracer: remove support for lightstep (#12654) This was deprecated in 3.15. Note that this is for sending tracing to lightstep and is unrelated to our Sourcegraph Extension which annotates lightstep tracing code.

view details

Keegan Carruthers-Smith

commit sha e842fbd75b978ada23a3d4a42904e17a95c9a02b

tracer: minor changes to initialization (#12655) - tracer: remove racey call to SetGlobalTracer - tracer: remove unneeded mutex around jaegerEnabled - tracer: just pass serviceName to initTracer - tracer: factor out jaeger opts into struct - tracer: factor out creation of tracer

view details

Keegan Carruthers-Smith

commit sha cf91a09bbe7a72c2c466b2b218f08b3ed7d71362

tracer: set URL to trace in x-trace header This is what we did for lightstep and it is quite convenient. We couldn't do this when we originally added Jaeger since we didn't proxy it. However, now that we do we can provide a direct link.

view details

push time in 4 days

delete branch sourcegraph/sourcegraph

delete branch : k/tracer-init

delete time in 4 days

push eventsourcegraph/sourcegraph

Keegan Carruthers-Smith

commit sha e842fbd75b978ada23a3d4a42904e17a95c9a02b

tracer: minor changes to initialization (#12655) - tracer: remove racey call to SetGlobalTracer - tracer: remove unneeded mutex around jaegerEnabled - tracer: just pass serviceName to initTracer - tracer: factor out jaeger opts into struct - tracer: factor out creation of tracer

view details

push time in 4 days

PR merged sourcegraph/sourcegraph

Reviewers
tracer: minor changes to initialization

Several things that I came across while trying to change this code. Includes minor fixes and refactorings to make another PR cleaner.

+63 -44

0 comment

1 changed file

keegancsmith

pr closed time in 4 days

more