Re: [e-users] build failure caused by commit 82910/82911

[Arvind R <arv...@gm...>]

On Thu, Jan 17, 2013 at 10:47 PM, Gustavo Sverzut Barbieri
<bar...@pr...> wrote:
> On Thu, Jan 17, 2013 at 3:07 PM, Arvind R <arv...@gm...> wrote:
>> On Thu, Jan 17, 2013 at 10:23 PM, Gustavo Sverzut Barbieri
>> <bar...@pr...> wrote:
>>> On Thu, Jan 17, 2013 at 12:37 AM, Cedric BAIL <ced...@fr...> wrote:
>>>> On Thu, Jan 17, 2013 at 11:17 AM, Cedric BAIL <ced...@fr...> wrote:
>>>>> On Thu, Jan 17, 2013 at 10:56 AM, Arvind R <arv...@gm...> wrote:
>>>>>> On Thu, Jan 17, 2013 at 6:18 AM, Gustavo Sverzut Barbieri
>>>>>> <bar...@pr...> wrote:
>>>>>>> Checked the commits now (cellphone) and 10/11 commits are unrelated to
>>>>>>> build. Are you sure 09 worked?
>>>>>> No
>>>>>>>
>>>>>>> Could you remove every file and have a clean tree before you run autogen
>>>>>>> again?
>>>>>>>
>>>>>> the problem is caused by using --with-crypto=gnutls flag!!! Foll.
>>>>>> patch fixes it!!!
>>>>>> Don't ask me why - but info autoconf ->macro_index does not have
>>>>>> AC_PATH_GENERIC.
>>>>>> diff -uprN a/configure.ac b/configure.ac
>>>>>> --- a/configure.ac      2013-01-17 02:35:19.000000000 +0530
>>>>>> +++ b/configure.ac      2013-01-17 07:18:40.000000000 +0530
>>>>>> @@ -275,10 +275,10 @@ case "$build_crypto" in
>>>>>>        EFL_DEPEND_PKG([crypto], [GNUTLS], [gnutls >= 2.11])
>>>>>>
>>>>>>        # libgcrypt
>>>>>> -      AC_PATH_GENERIC([libgcrypt], [], [:],
>>>>>> -         [AC_MSG_ERROR([libgcrypt required but not found])])
>>>>>> -      requirements_libs_crypto="${LIBGCRYPT_LIBS} ${requirements_libs_crypto}"
>>>>>> -      requirements_cflags_crypto="${LIBGCRYPT_CFLAGS}
>>>>>> ${requirements_cflags_crypto}"
>>>>>> +      #AC_PATH_GENERIC([libgcrypt], [], [:],
>>>>>> +      #   [AC_MSG_ERROR([libgcrypt required but not found])])
>>>>>> +      #requirements_libs_crypto="${LIBGCRYPT_LIBS} ${requirements_libs_crypto}"
>>>>>> +      #requirements_cflags_crypto="${LIBGCRYPT_CFLAGS}
>>>>>> ${requirements_cflags_crypto}"
>>>>>>        ;;
>>>>>>
>>>>>>     openssl)
>>>>>
>>>>> Will check that and revert my revert if that fix the build issue.
>>>>
>>>> Ok, reverted the revert and pushed your temporary patch. Still need to
>>>> investigate why and turn it back on once properly fixed.
>>>
>>> Hi all,
>>>
>>> I've investigated and:
>>>  - gcrypt is indeed used by gnutls part (eet, ecore_con...)
>>>  - there is no gcrypt.pc or similar
>>>  - libgcrypt ships with AM_PATH_LIBGCRYPT() that would do the "right thing"
>>>
>>> That said I did http://trac.enlightenment.org/e/changeset/82939 which
>>> hopefully fixes it. Please test, it works here (ArchLinux). The
>>> libcrypt.m4 is included in our repository so it will work for people
>>> doing autogen.sh without that package.
>>>
>>> The LIBGCRYPT_LIBS variable will be populated with `libgcrypt-config
>>> --libs` output, and that will end into USE_EET_LIBS. Then please check
>>> the output of your libgcrypt-config --libs if that fails.
>>>
>>
>> Alas, debian stable | testing libgcrypt-config --libs outputs -L/lib |
>> -L/lib/x86_...
>> before the -lgcrypt -:(
>> AM_PATH_GENERIC and AM_PATH_LIBGCRYPT both end up calling
>> libgrypt-config.  How is it OK in older, say svn82296, which also uses
>> the macro?
>
> before we were 'forgetting' those behing :-/
>
>
>> So the 82939 changeset still breaks debian -:(
>
> Could you check if you fix to output the proper path in -L fixes that?
> (I understand your libraries are NOT in /lib, but rather somewhere
> else)

Hey, the path is right - debian installs libgcrypt in /lib...  The
working version (82296) has the -L flag in as you say ECORE_CON_LIBS
and EET_LIBS only. But some recent m4 changes have it appearing
everywhere and breaking libtool, I think.

>
> If that still doesn't fix, we'll have to figure out what's wrong. One
> way to help me is to send the pastebin of your src/Makefile and also
> the last bits of:
>
>    make V=1 -C src lib/evas/libevas.la
>
> (I assumed that libevas.la was a failing case)
No. Fails on the first compile instance of tree - where it seems to
think the -L.. is a target that has to be built!

>
> And please compile with -j1 so I can get output in-order to see what's
> happening.
>
Will check - after a while.
> --
> Gustavo Sverzut Barbieri
> http://profusion.mobi embedded systems
> --------------------------------------
> MSN: bar...@gm...
> Skype: gsbarbieri
> Mobile: +55 (19) 9225-2202
>
> ------------------------------------------------------------------------------
> Master Visual Studio, SharePoint, SQL, ASP.NET, C# 2012, HTML5, CSS,
> MVC, Windows 8 Apps, JavaScript and much more. Keep your skills current
> with LearnDevNow - 3,200 step-by-step video tutorials by Microsoft
> MVPs and experts. ON SALE this month only -- learn more at:
> http://p.sf.net/sfu/learnmore_122712
> _______________________________________________
> enlightenment-users mailing list
> enl...@li...
> https://lists.sourceforge.net/lists/listinfo/enlightenment-users