[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: libidn tst_stringprep tst_idna fail on Cygwin mingw64-x86_64 cross-b

From: Brian Inglis
Subject: Re: libidn tst_stringprep tst_idna fail on Cygwin mingw64-x86_64 cross-build
Date: Wed, 13 Jul 2022 01:25:46 -0600
User-agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:91.0) Gecko/20100101 Thunderbird/91.11.0

On 2022-07-12 11:41, Brian Inglis wrote:
On 2022-07-12 10:01, Brian Inglis wrote:
On 2022-07-12 04:05, Simon Josefsson wrote:
Brian Inglis writes:

Cygwin mingw64-x86_64 cross-build tst_stringprep tst_idna fail on
1.39-1.41, all earlier and other i686/x86 cross build and Cygwin
"native" x86_64 and i686 build tests all pass.

Tried some things with no change - details attached.

Any idea if and which gnulib changes could have affected this?
Suggestions for other things to look at, try, next steps?

Thanks for the report.  Does reproducing this require a Windows machine,
or can I use Cygwin via Wine these days?

AFAIK Cygwin under Wine should work well as the projects cooperate with bugs and patches.

The 201 error code is STRINGPREP_MALLOC_ERROR, does that help?  Also,
can you look at the report and patch from Francesco Pretto to this list
earlier?  It was about a Win64 problem, but I wasn't able to reproduce
it.  Can you try it and see if the patch helps?

That is a surprising error code but maybe not from under mingw64-x86_64 rather than Cygwin.

Thanks for that pointer; I'll check the sources against the patch and see if it builds and helps with the gcc cross-build and test:

If possible, I would like to reproduce this, so any hints on how to do
that would be appreciated.

Will let you know how things go.

Great - mingw64 x86_64/amd64 1.39-1.41 build and check successfully!

I am also pushing those changes to our Github Actions CI to check each build also succeeds in pristine environments, unlike my Cygwin build system with every devel and support package ever needed! ;^>

I'll rerun with mingw64 i686/x86 from 1.39-1.41 to check for regressions, then Cygwin i686/x86 and x86_64/amd64 1.39-1.41, and each also in CI.

Running all Cygwin "native" and mingw64 cross, x86_64 and i686, desktop and CI builds with that patch causes no (obvious) problems and builds and checks okay.

To reproduce under Cygwin, see:

select package name link takes you to source summary page:

select packaging repository link takes you to git repo:

select playground branch link takes you to:;a=shortlog;h=refs/heads/playground

select the tag entry 1.41 and you get a set of sources you can use to build under Cygwin, after installing Cygwin, download and run setup-x86_64 to install base packages, then rerun setup-x86_64 to select packages cygport (which pulls in all the Cygwin build depends) and also the DEPEND entries in the cygport (gettext-devel mingw64-x86_64-gettext mingw64-x86_64-win-iconv mingw64-x86_64-binutils mingw64-x86_64-gcc-core doxygen gengetopt gperf gtk-doc help2man) to install.

To run cygport on this (cygport is a collection of bash scripts):

        $ cygport mingw64-x86_64-libidn.cygport get all check &> log &

move the fix-...-x64.patch out of the way to reproduce the test failures in releases 1.39-41.

Take care. Thanks, Brian Inglis, Calgary, Alberta, Canada

This email may be disturbing to some readers as it contains
too much technical detail. Reader discretion is advised.
[Data in binary units and prefixes, physical quantities in SI.]

reply via email to

[Prev in Thread] Current Thread [Next in Thread]