Add --disable-icu flag for iconv backend

[daravi]

Specify library name and version: boost/all

• [✓] I've read the guidelines for contributing.
• [✓] I've followed the PEP8 style guides for Python code in the recipes.
• [✓] I've used the latest Conan client version.
• [✓] I've tried at least one configuration locally with the
  conan-center hook activated.

I'm not sure if this is a bug in Boost's documentation, its build system, or the recipe. But building without this flag I see that my boost binaries are linked with my system icu.

[ghost]

I detected other pull requests that are modifying boost/all recipe:

• add boost/1.65.1 #4354
• [boost] b2 user-config.jam needs CXX binary to be quoted as path may contain spaces #4149
• [boost] debug_level=0 should emit -d0 #4121
• Boost/1.72.0/mpi #3959
• boost: patch serialization sources #3918
• Boost: Change assertions to warnings #3900

This message is automatically generated by https://github.com/ericLemanissier/conan-center-conflicting-prs so don't hesitate to report issues/improvements there.

[madebr]

@daravi
Is this problem still present with the new boost recipe, pushed to master today?

[conan-center-bot]

All green in build 3 (ed49a80faf07185691267f3e15957a64b504c3e2)! 😊

• boost/1.69.0: Generated 261 packages. All logs here
• boost/1.70.0: Generated 261 packages. All logs here
• boost/1.71.0: Generated 261 packages. All logs here
• boost/1.72.0: Generated 261 packages. All logs here
• boost/1.73.0: Generated 261 packages. All logs here
• boost/1.74.0: Generated 243 packages. All logs here
• boost/1.75.0: Generated 131 packages. All logs here

[conan-center-bot]

All green in build 4 (ed49a80faf07185691267f3e15957a64b504c3e2)! 😊

• boost/1.69.0: Generated 261 packages. All logs here
• boost/1.70.0: Generated 261 packages. All logs here
• boost/1.71.0: Generated 261 packages. All logs here
• boost/1.72.0: Generated 261 packages. All logs here
• boost/1.73.0: Generated 261 packages. All logs here
• boost/1.74.0: Generated 243 packages. All logs here
• boost/1.75.0: Generated 131 packages. All logs here

[prince-chrismc]

It's another broken CI state 😖

[danimtb]

Something happened in this PR that generated more packages than usual. I am blocking the merge until we figure it out

[danimtb]

The CI generated 2 different recipe revisions due to different job executions that took a different master state as the target branch. I have pushed an empty commit to regenerate the packages

[conan-center-bot]

All green in build 5 (a4a6b050678e6214ed79586b568b95fe0eedeb9d)! 😊

• boost/1.69.0: Generated 131 packages. All logs here
• boost/1.70.0: Generated 131 packages. All logs here
• boost/1.71.0: Generated 131 packages. All logs here
• boost/1.72.0: Generated 131 packages. All logs here
• boost/1.73.0: Generated 131 packages. All logs here
• boost/1.74.0: Generated 131 packages. All logs here
• boost/1.75.0: Generated 131 packages. All logs here

[prince-chrismc]

The bot (mine too) seems to be incorrectly counting the reviews ... some were not dismissed but are not counted

[SSE4]

@danimtb @jgsogo what is preventing it from merging?

[jgsogo]

Hi! The bot says that @danimtb is blocking it... but he has approved the PR (after the previous block) and I can see the right (expected) data via API (https://api.github.com/repos/conan-io/conan-center-index/pulls/3855/reviews). I'll debug the algorithm to see what is happening here.

Thanks for raising the hand! 👍

[jgsogo]

Sorry, I was wrong. The problem here is that the latest commit (a4a6b050678e6214ed79586b568b95fe0eedeb9d) has only be approved by @danimtb and @prince-chrismc, we need at least one more positive review.

Algorithm is working fine, but we are not taking into account empty commits (wish it was a way to get this information using Github API).

[prince-chrismc]

@jgsogo I fixed the issue in my bot https://github.com/prince-chrismc/conan-center-index-pending-review/blob/main/pending_review/pull_request.go#L106-L118

I count valid approvals regardless if they are on the head commit or not... this way it matches the github API more closely

It seems the API handles the DISMISSED case well

[jgsogo]

Hi, @prince-chrismc! Thanks for pointing it out. I'm not confident enough (1) and that is the reason why we are checking that the reviews are associated with the latest commit SHA.

To implement our logic we would need:

• every APPROVED to be transformed into DISMISSED if a new commit modifies the sources.
• but CHANGES_REQUESTED state should remain even if there are new commits

---

(1) it clearly means that we haven't tested it enough. 🙈

[prince-chrismc]

The UI still shows a ✔️ #3855 (review)

Though the API reports it as a comment now

{
    "id": 550510346,
    "node_id": "MDE3OlB1bGxSZXF1ZXN0UmV2aWV3NTUwNTEwMzQ2",
    "user": {
      "login": "prince-chrismc",
      "id": 16867443,
      // ...
    },
    "body": "",
    "state": "COMMENTED",
    "html_url": "https://github.com/conan-io/conan-center-index/pull/3855#pullrequestreview-550510346",
    "pull_request_url": "https://api.github.com/repos/conan-io/conan-center-index/pulls/3855",
    "author_association": "CONTRIBUTOR",
    "_links": {
      "html": {
        "href": "https://github.com/conan-io/conan-center-index/pull/3855#pullrequestreview-550510346"
      },
      "pull_request": {
        "href": "https://api.github.com/repos/conan-io/conan-center-index/pulls/3855"
      }
    },
    "submitted_at": "2020-12-11T19:17:24Z",
    "commit_id": "ed49a80faf07185691267f3e15957a64b504c3e2"
  },