-
Notifications
You must be signed in to change notification settings - Fork 33
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Upgrade to Boost 1.78 #80
Comments
(Poking issue #77 and participants @Jean-Romain @dcooley -- may make more sense over here) So I ran a reverse depends check and it is not looking great; the results summary is here and I'll just copy the results table over:
The second (filled-in) column 'badInstall' matters, a TRUE indicates compilation problem. I only had time to look at APAtree so far while it still ran and it is geometry related. As are your packages which also seem fail to build. (The following columns just quote the current CRAN status to indicate whether the package has other open issues at CRAN.) Addressing builds under BH 1.78.0 may be something simple and I would really appreciate if you had time to poke and take a look. I have my hands still a little full my an ongoing Rcpp tranistion, and one we may want to do for RcppEigen. I will try to take a look too, maybe we can compare notes here over the couple of days and weeks. This will likely drag on into January. I have prepared a release of the snapshot / branch you can install directly from the drat repo via
This corresponds to the 1.78.0 branch in the repo here. Let me know if you have any questions. |
I'll test as soon as possible with |
Tested, and I think you can't release Boost 1.78 as is. There are numerous errors from boost code. I'm not a Boost expert and I don't understand each error but they definitively do no come from my code. Here a summary of the main errors found at compile time. The first one looks particularly odd. Seems like a bad character introduced by error.
Compilation log (excluding files with no Boost includes)
|
Line 27 is a typo I made. I will remove that. Can you edit locally and try again? modified inst/include/boost/move/algo/detail/merge.hpp
@@ -24,7 +24,7 @@
#if defined(BOOST_CLANG) || (defined(BOOST_GCC) && (BOOST_GCC >= 40600))
#pragma GCC diagnostic push
-o//#pragma GCC diagnostic ignored "-Wsign-conversion"
+//#pragma GCC diagnostic ignored "-Wsign-conversion"
#endif
namespace boost { I also updated the drat repo, you should now get |
So I removed the I'm wondering how this simple error may have induced so much unrelated troubles 😮 |
Yes it did :) Lessons learned: do not fat-finger in errors. Doh. And don't ... rush! I usually inspect all failed logs but it is such a pain with Boost and CRAN's insistence to turn off diagnostitcs suppression. Anyway, much better now: Test of BH 1.77.99.1 had 11 successes, 7 failures, and 4 skipped packages.
Ran from 2021-12-13 17:45:46.04 to 2021-12-13 18:03:19.79 for 17.562 mins
Average of 47.898 secs relative to 201.386 secs using 6 runners
Failed packages: breathteststan, gfilogisreg, PReMiuM, rstpm2, secr, TDA, tmbstan
Skipped packages: hBayesDM, nlmixr, RcppRedis, starvz
None still working
None still scheduled
Error summary:
package missingPkg badInstall error fail warn note ok hasOtherIssue
1: breathteststan FALSE 1 0 0 12 0 FALSE not BH, minor (?) recurring numeric test error
2: gfilogisreg TRUE 0 0 0 11 2 FALSE error: ‘isnormal’ was not declared in this scope
3: PReMiuM TRUE 0 0 0 4 9 FALSE known flaky
4: rstpm2 FALSE 0 0 3 6 4 FALSE not BH, minor (?) recurring numeric test error
5: secr TRUE 0 0 0 13 0 FALSE error: 'isnan' was not declared in this scope
6: TDA TRUE 0 0 3 8 2 TRUE error: wrong number of template arguments (2, should be 1)
7: tmbstan TRUE 0 0 0 4 9 FALSE uses C++03, should try C++11 or later Thanks for lending an ear, and most importantly, for spotting my blunder. I'll take a more detailed look at these in a bit after dinner. |
You're welcome, You have already made so much for the community. If we can sometime help you in return it very fair 😉 |
First fix: |
Second fix:
|
Third fix for
We are looking good now, this should be ready to ship. |
And we're on CRAN! So closing this one. |
Boost 1.78 was just released so we can continue our cadence of annual releases.
With that we should look into Boost process (cf #78) and see if #77 needs work as things have by now changed (e.g. Solaris now retired from the set of test platforms)
The text was updated successfully, but these errors were encountered: