From: vincent.belaiche@gmail.com (Vincent Belaïche)
To: Eli Zaretskii <eliz@gnu.org>
Cc: 27516@debbugs.gnu.org,
"Vincent Belaïche" <vincent.belaiche@gmail.com>,
npostavs@users.sourceforge.net
Subject: bug#27516: 25.2.50; Building master branch on MSW fails because of gnutls and other reasons...
Date: Tue, 11 Jul 2017 07:20:27 +0200 [thread overview]
Message-ID: <84eftnsg6c.fsf@AigleRoyal> (raw)
In-Reply-To: <831sq1r6ss.fsf@gnu.org>
Sorry for the delay in replying, feedback below. I could not get a
running Emacs 26 with GNUTLS working also.
Le 30/06/2017 à 14:36, Eli Zaretskii a écrit :
>> From: vincent.belaiche@gmail.com (Vincent Belaïche)
>> Date: Fri, 30 Jun 2017 12:44:28 +0200
>> Cc: Vincent Belaïche <vincent.belaiche@gmail.com>
>>
>>> ./configure uses pkg-config to test for gnutls, so you need to have
>>> pkg-config as well as the gnutls.pc file in the pkg-config
>>> PKG_CONFIG_PATH. Try running
>>
>> OK, I could get some gnutls.pc file with the gnutls library and I edited
>> it for the content to match my installation, so now it is as follows:
>
> You shouldn't need to edit it, it should work as-is.
It would certainly work as is if the gnutls include files where placed
in guess-estimatable directories relative to pkg-config, but I placed
them in another directory in order not to mix up my mingw directory with
files from another source (namely gnutls.org), so I had to comment out
the prefix setting for the other settings to be taken as edited in
gnutls.pc and not replaced by guess-estimates, and I also edited the
other settings to point at the right place.
I had also to make another hack : when I took the gnutls windows port
from gnutls.org, the include files are in lib/includes, I had to move
this to includes/gnutls because the Emacs source makes some
#include <gnutls/...>
directive with an explicit gnutls/... superdirectory.
So, OK, it's my fault, I should have taken the EZwinports and placed it
into my mingw.
>
>> --- a/src/w32heap.c
>> +++ b/src/w32heap.c
>> @@ -116,9 +116,9 @@ typedef NTSTATUS (NTAPI *PRTL_HEAP_COMMIT_ROUTINE) (
>> to build only the first bootstrap-emacs.exe with the large size,
>> and reset that to a lower value afterwards. */
>> #if defined _WIN64 || defined WIDE_EMACS_INT
>> -# define DUMPED_HEAP_SIZE (21*1024*1024)
>> +# define DUMPED_HEAP_SIZE (26*1024*1024)
>> #else
>> -# define DUMPED_HEAP_SIZE (12*1024*1024)
>> +# define DUMPED_HEAP_SIZE (13*1024*1024)
>> #endif
>
> It is very strange why you need this increment, while no one else
> does.
>
I can't tell why I need this. But if you explain me how to debug the
compile crash due to lack of memory, then we may know more about it.
Please note that I have been using a 32 bit compilation, while I am on a
64bit machine.
>> I note that that current sizes between narrow and wide are not
>> consistent, as 21 is not the double of 12.
>
> The size in the 64-bit build doesn't need to be a double of the size
> for the 32-bit build, since not every Lisp object is exactly twice
> large in the 64-bit build. These numbers were found by
> trial-and-error.
>
>> Anyway, now I have a running Emacs 26, however it seems that it does not
>> have any gnutls in it, because when I try a message send, I get the
>> following error message:
>>
>> --8<----8<----8<----8<----8<-- begin -->8---->8---->8---->8---->8----
>> Sending via mail...
>> smtpmail-send-it: Sending failed: Emacs does not support TLS
>> --8<----8<----8<----8<----8<-- end -->8---->8---->8---->8---->8----
>
> Like Noam said: make sure the DLL is where Emacs looks for it.
I could make it work. The issue was that only only you need the
directory (call it dir_1) where the DLL is in the PATH envvar, but also
you need that no other directory (call it dir_2) containing conflicting
DLL be in the PATH envar *before* dir_1.
I had a dir_2 containing other libgmp-10.dll and libp11-kit-0.dll and
that was sufficient to make it fail on execution. This was due to that I
had tried in the first place to use gnutls-cli.exe, so I had a
gnutls-cli.exe installed dir_2 which was using incompatible DLL compared
to that in dir_1.
Once again, OK it's my fault, I should have blindly followed the
instructions in
(info "(emacs-gnutls) Help For Users")
and placed Eli's EZwinPorts into the same bin directory as emacs.exe,
but instead I took the latest from gnutls.org and placed it in another
directory. Sorry for I am desperately prone not to follow instructions
as-is.
Vincent.
---
L'absence de virus dans ce courrier électronique a été vérifiée par le logiciel antivirus Avast.
https://www.avast.com/antivirus
next prev parent reply other threads:[~2017-07-11 5:20 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-06-28 11:31 bug#27516: 25.2.50; Building master branch on MSW fails because of gnutls and other reasons Vincent Belaïche
2017-06-28 14:09 ` Noam Postavsky
2017-06-30 10:44 ` Vincent Belaïche
2017-06-30 11:51 ` npostavs
2017-06-30 12:36 ` Eli Zaretskii
2017-07-11 5:20 ` Vincent Belaïche [this message]
2018-07-17 22:55 ` Noam Postavsky
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=84eftnsg6c.fsf@AigleRoyal \
--to=vincent.belaiche@gmail.com \
--cc=27516@debbugs.gnu.org \
--cc=eliz@gnu.org \
--cc=npostavs@users.sourceforge.net \
/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.
Code repositories for project(s) associated with this external index
https://git.savannah.gnu.org/cgit/emacs.git
https://git.savannah.gnu.org/cgit/emacs/org-mode.git
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.