public inbox for cygwin@cygwin.com
 help / color / mirror / Atom feed
From: JonY <10walls@gmail.com>
To: cygwin@cygwin.com
Subject: Re: [BUG] Package mingw64-x86_64-icu is broken
Date: Wed, 20 Jun 2018 18:10:00 -0000	[thread overview]
Message-ID: <b1287c14-6ed3-e842-d13c-47b879942e41@gmail.com> (raw)
In-Reply-To: <7cf87ba8-c298-cacb-63d7-54ff2c50da2c@weilnetz.de>


[-- Attachment #1.1: Type: text/plain, Size: 1928 bytes --]

On 06/19/2018 10:18 AM, Stefan Weil wrote:
> Am 19.06.2018 um 11:53 schrieb JonY:
>> On 06/18/2018 12:12 PM, Stefan Weil wrote:
>>> Am 06.06.2018 um 21:13 schrieb Stefan Weil:
>>>> Both mingw64-x86_64-icu-57.1-1 and mingw64-x86_64-icu-57.1-2 are broken:
>>>>
>>>> This code always fails:
>>>>
>>>> icu::Normalizer2::getInstance(nullptr, "nfkc", UNORM2_COMPOSE, error_code);
>>>>
>>>> The problem was detected when comparing Tesseract for Windows
>>>> executables: while the 32 bit version worked fine, the 64 bit version
>>>> failed. The failure could be localized, and the mingw64-x86_64-icu
>>>> package was identified to be causing it.
>>>>
>>>> https://github.com/tesseract-ocr/tesseract/issues/1625#issuecomment-395161152
>>>> contains details and also a short test code which can be used to
>>>> reproduce the problem.
>>>>
>>>> The 32 bit package mingw64-i686-icu-57.1-2 works fine.
>>>>
>>>> Kind regards
>>>> Stefan Weil
>>>
>>> Ping. How can I help to get this issue fixed?
>>
>> I noticed cygport is using llvm version of binutils, not sure if that
>> broke things.
>>
>> Was the issue there if you built it yourself with gcc/binutils?
> 
> 
> No, a local build with x86_64-w64-mingw32-gcc works fine.
> Tested with 57.1-2, but also with recent versions of icu.
> 
> A good indicator of a broken installation is a small icudata57.dll:
> 
> $ ls -l /usr/*/sys-root/mingw/bin/icudata57.dll
> -rwxr-xr-x 1 25680896 Nov 10  2016
> /usr/i686-w64-mingw32/sys-root/mingw/bin/icudata57.dll
> -rwxr-xr-x 1    15872 Nov 10  2016
> /usr/x86_64-w64-mingw32/sys-root/mingw/bin/icudata57.dll
> 
> The (good) file for 32 bit is much larger than the (broken) file for 64 bit.
> 
The cygport post-install debug strip is probably stripping too much.
Whoever the maintainer of this package can do a test rebuild with llvm-*
binutils temporarily moved away as a test if Cygport is used.


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

      reply	other threads:[~2018-06-19 23:46 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2018-06-06 19:13 Stefan Weil
2018-06-18 12:20 ` Stefan Weil
2018-06-19 21:21   ` JonY
2018-06-19 23:46     ` Stefan Weil
2018-06-20 18:10       ` JonY [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=b1287c14-6ed3-e842-d13c-47b879942e41@gmail.com \
    --to=10walls@gmail.com \
    --cc=cygwin@cygwin.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).