[travis] upload packages to transfer.sh and to the Snap Store #1125

Saviq · 2019-10-15T14:35:37Z

No description provided.

codecov · 2019-10-15T15:13:54Z

Codecov Report

Merging #1125 into master will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master    #1125   +/-   ##
=======================================
  Coverage   71.04%   71.04%           
=======================================
  Files         201      201           
  Lines        7374     7374           
=======================================
  Hits         5239     5239           
  Misses       2135     2135

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 1a32e64...e73a0e7. Read the comment docs.

.travis.yml

CMakeLists.txt

Saviq · 2019-10-17T14:28:00Z

tools/report-build.py

+import requests
+import sys
+
+HEADERS = {"Authorization": "Bearer {}".format(os.environ["GITHUB_TOKEN"])}


There's very little exception handling in this script, on purpose - anything out of the ordinary should make it go away.

ricab · 2019-10-17T15:38:06Z

Thanks, this idea sounds really useful. I did not look at the code (so this is not a review) but I have some comments on the topics you raised.

Saviq · 2019-10-17T18:36:58Z

@townsend2010 care to share your opinion on the topics I raised?

- do not `snapcraft login` if not publishing - do not run covreport on failure - rename `build_id` - streamline `name`/`owner` determination - add more timeline event types Co-authored-by: Ricardo Abreu <[email protected]>

- rename report_build to snake-case - reorder imports - fix long line

On merges from a fork, credentials are unavailable.

KeyError is too big of a net.

transfer.sh, when in a bad state, causes curl to hang indefinitely.

ricab · 2019-10-23T09:36:04Z

UX detail (I should say DX here, for developer 😛): I think the messages look a bit more tangled up in appearance after the snap command. How would you feel about changing the format to something like:

macOS build available: multipass-0.10.0-dev.422.pr1125+g6bb5ecd7.mac-Darwin.pkg
Snap build available: multipass_0.10.0-dev.320.pr1125+g5c4858d_amd64.snap
- or snap refresh multipass --channel edge/pr1125

Saviq · 2019-10-23T09:44:04Z

How would you feel about changing the format to something like:

Yes, in a further PR. Note I'm matching the build names with a simple str.startswith() - anything more involved will require deeper Markdown parsing :)

ricab · 2019-10-23T10:05:06Z

tools/report_build.py

+        query GetEvents($owner: String!,
+                        $name: String!,
+                        $pull_request: Int!,
+                        $count: Int!) {


I wish our clang-format accepted this sort of formatting in C++, it often makes for much more readable code.

.travis.yml

multipass-ci-bot · 2019-10-23T11:20:55Z

Snap build available: snap refresh multipass --channel edge/pr1125

Saviq

Unfortunately I seem to have saturated transfer.sh again… It shouldn't be that bad when I stop experimenting with this, but will look around for an alternative for the future.

Saviq · 2019-10-23T11:42:53Z

.travis.yml

+      - '[[ "${TRAVIS_EVENT_TYPE}" != "pull_request" || "${TRAVIS_SECURE_ENV_VARS}" == "false" ]] && exit'
+      - BUILDS=()
+      - >-
+        [ -n "${MULTIPASS_SNAP_CHANNEL}" ]
+        && snapcraft login --with ${HOME}/snap-login
+        && snapcraft push multipass_*.snap --release ${MULTIPASS_SNAP_CHANNEL}
+        && BUILDS+=("\`snap refresh multipass --channel ${MULTIPASS_SNAP_CHANNEL}\`")
+      - CURL_OUTPUT=$( timeout 5m curl --fail --upload-file *.snap https://transfer.sh/ )
+        && BUILDS+=("[$( basename ${CURL_OUTPUT} )](${CURL_OUTPUT})");
+        echo ${CURL_OUTPUT}
+      - ./tools/report_build.py Snap "${BUILDS[@]}"


I now wonder whether we should do this on the Debug build? The snap checks and now the uploads mean that this job runs almost twice as long as the Debug one.

And maybe it'd be better for debugging to have Debug builds out of PRs?

I don't know, I don't have a strong opinion on that. Release would be closer to the actual thing, but debug would indeed be more practical. Would that affect actual releasing? And aren't the snap checks useful in the build that uploads the snap? What is the gain in having debug and release builds switching places in what concerns duration?

Would that affect actual releasing?

No, for releases (RC we'd have to decide on) we'd build RelWithDebInfo still.

And aren't the snap checks useful in the build that uploads the snap?

At worst they would fail to upload. Now I think of it maybe we can drop the review step as the store does it for us… The snap upload would have to move up to script: though, so failure to release would be a CI failure, and alternate with review-tools if not publishing.

What is the gain in having debug and release builds switching places in what concerns duration?

One would do the snap review check, another would do the upload.

Any case, something to ponder for the future.

At least the "faster" concern is moot now: #1141

ricab

bors r+

1125: [travis] upload packages to transfer.sh and to the Snap Store r=ricab a=Saviq Co-authored-by: Michał Sawicz <[email protected]>

bors · 2019-10-23T13:05:55Z

Build succeeded

continuous-integration/travis-ci/push

1125: [travis] upload packages to transfer.sh and to the Snap Store r=ricab a=Saviq Co-authored-by: Michał Sawicz <[email protected]>

Saviq force-pushed the transfer-sh-builds branch 2 times, most recently from 05f7f89 to 3253e36 Compare October 15, 2019 15:13

This comment has been minimized.

Sign in to view

Saviq force-pushed the transfer-sh-builds branch from 3253e36 to 3721738 Compare October 16, 2019 14:18

This comment has been minimized.

Sign in to view

Saviq force-pushed the transfer-sh-builds branch from 3721738 to 95d9c29 Compare October 16, 2019 16:09

Saviq added the no-merge label Oct 16, 2019

Saviq force-pushed the transfer-sh-builds branch from 3ff2bfa to 95d9c29 Compare October 16, 2019 17:58

Saviq commented Oct 17, 2019

View reviewed changes

This comment has been minimized.

Sign in to view

Saviq force-pushed the transfer-sh-builds branch from 20cca79 to 42a4d38 Compare October 18, 2019 08:57

canonical deleted a comment from multipass-ci-bot Oct 18, 2019

This comment has been minimized.

Sign in to view

Saviq force-pushed the transfer-sh-builds branch from 3e2622f to 4d8c2b5 Compare October 18, 2019 12:13

This comment has been minimized.

Sign in to view

Saviq force-pushed the transfer-sh-builds branch 2 times, most recently from feb5552 to 092f753 Compare October 18, 2019 13:57

This comment has been minimized.

Sign in to view

Saviq force-pushed the transfer-sh-builds branch 6 times, most recently from cf03038 to f059020 Compare October 18, 2019 15:38

This comment has been minimized.

Sign in to view

Saviq changed the title ~~[travis] upload packages to transfer.sh~~ [travis] upload packages to transfer.sh and to the Snap Store Oct 22, 2019

Saviq force-pushed the transfer-sh-builds branch from 04a1dc8 to 6dc6d82 Compare October 22, 2019 20:38

This comment has been minimized.

Sign in to view

Saviq and others added 4 commits October 23, 2019 09:49

[tools] updates for review comments

63d0564

- do not `snapcraft login` if not publishing - do not run covreport on failure - rename `build_id` - streamline `name`/`owner` determination - add more timeline event types Co-authored-by: Ricardo Abreu <[email protected]>

[tools] fix flake8 warnings

b8f38f7

- rename report_build to snake-case - reorder imports - fix long line

[travis] only run privileged steps when possible

888f4c4

On merges from a fork, credentials are unavailable.

[travis] fix gating on event type

3a8ef71

Saviq force-pushed the transfer-sh-builds branch from dc179d4 to 96cfca8 Compare October 23, 2019 07:53

Saviq added 3 commits October 23, 2019 10:57

[tools] use __typename to control flow

44f90f7

KeyError is too big of a net.

[travis] do not rely on curl to time out

780120d

transfer.sh, when in a bad state, causes curl to hang indefinitely.

[travis] improve variable names

267deb9

Saviq force-pushed the transfer-sh-builds branch from 96cfca8 to 267deb9 Compare October 23, 2019 09:16

This comment has been minimized.

Sign in to view

ricab reviewed Oct 23, 2019

View reviewed changes

[cmake] use MULTIPASS_BUILD_LABEL instead

e73a0e7

Saviq commented Oct 23, 2019

View reviewed changes

ricab approved these changes Oct 23, 2019

View reviewed changes

bors bot added a commit that referenced this pull request Oct 23, 2019

Merge #1125

2c2b8f7

1125: [travis] upload packages to transfer.sh and to the Snap Store r=ricab a=Saviq Co-authored-by: Michał Sawicz <[email protected]>

bors bot merged commit e73a0e7 into master Oct 23, 2019

bors bot deleted the transfer-sh-builds branch October 23, 2019 13:05

Saviq added a commit that referenced this pull request Oct 23, 2019

Merge #1125

b64e9be

1125: [travis] upload packages to transfer.sh and to the Snap Store r=ricab a=Saviq Co-authored-by: Michał Sawicz <[email protected]>

Saviq mentioned this pull request Oct 23, 2019

[CI] publish Debug builds instead of RelWithDebInfo #1146

Open

Saviq added enhancement low Low priority. Combine with "medium" to signal intermediate low-medium priority. labels Oct 23, 2019

Saviq added a commit that referenced this pull request Oct 24, 2019

Merge #1125

f9f33f2

1125: [travis] upload packages to transfer.sh and to the Snap Store r=ricab a=Saviq Co-authored-by: Michał Sawicz <[email protected]>

Saviq added a commit that referenced this pull request Oct 28, 2019

Merge #1125

d0af328

1125: [travis] upload packages to transfer.sh and to the Snap Store r=ricab a=Saviq Co-authored-by: Michał Sawicz <[email protected]>

[travis] upload packages to transfer.sh and to the Snap Store #1125

[travis] upload packages to transfer.sh and to the Snap Store #1125

Uh oh!

Conversation

Saviq commented Oct 15, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

codecov bot commented Oct 15, 2019 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

Uh oh!

Uh oh!

Uh oh!

Saviq Oct 17, 2019

Choose a reason for hiding this comment

Uh oh!

This comment has been minimized.

ricab commented Oct 17, 2019

Uh oh!

Saviq commented Oct 17, 2019

Uh oh!

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

ricab commented Oct 23, 2019

Uh oh!

Saviq commented Oct 23, 2019

Uh oh!

This comment has been minimized.

ricab Oct 23, 2019

Choose a reason for hiding this comment

Uh oh!

Uh oh!

multipass-ci-bot commented Oct 23, 2019

Uh oh!

Saviq left a comment • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Saviq Oct 23, 2019

Choose a reason for hiding this comment

Uh oh!

ricab Oct 23, 2019

Choose a reason for hiding this comment

Uh oh!

Saviq Oct 23, 2019

Choose a reason for hiding this comment

Uh oh!

ricab Oct 23, 2019

Choose a reason for hiding this comment

Uh oh!

Saviq Oct 23, 2019

Choose a reason for hiding this comment

Uh oh!

ricab left a comment

Choose a reason for hiding this comment

Uh oh!

bors bot commented Oct 23, 2019

Build succeeded

Uh oh!

Uh oh!

Saviq commented Oct 15, 2019 •

edited

Loading

codecov bot commented Oct 15, 2019 •

edited

Loading

Saviq left a comment •

edited

Loading