Hi,
thanks.
Ludovic Courtès writes:
> Hi!
>
> ng0 skribis:
>
>> Could someone with the right privileges mark this in patchwork as
>> commited? I have no options in the below link to mark it as
>> anything.
>>
>> http://patchwork.sourceware.org/patch/13377/
>
> Done. Unfortunately I won’t have
Hi!
ng0 skribis:
> Could someone with the right privileges mark this in patchwork as
> commited? I have no options in the below link to mark it as
> anything.
>
> http://patchwork.sourceware.org/patch/13377/
Done. Unfortunately I won’t have the patience to manually mark patches
as committed.
Ludovic Courtès writes:
> Hello,
>
> Matthew Jordan skribis:
>
>>> Also, why do we need this “src” output? IIUC, it contains the source of
>>> some of the unit tests, right?
>>
>> I changed the "src" output to "tests", to clarify.
>
> OK; it’s not clear to me if users need it, but it probably do
> What is this "bash" used for? I don't see it referred to in the install
> phase.
Thanks.
I'll remove it in the next patch for go@1.4.
--
Matthew Jordan
Sent with my mu4e
> I also noticed the unusual file system layout in the main output:
>
> --8<---cut here---start->8---
> $ ls $(./pre-inst-env guix build go)
> /gnu/store/05al7hhwng4a5hs1ix2i4agqzm9cs4gy-go-1.4.3-doc:
> share
>
> /gnu/store/6f6ns5c7gfr79pl190r9xjyx6rbx1vv1-go-1.
On Sun, Jun 26, 2016 at 06:14:48PM -0400, Matthew Jordan wrote:
> Subject: [PATCH] gnu: Add go@1.4
>
> * gnu/local.mk: Modified file.
> * gnu/packages/golang.scm: New file.
> + (replace 'install
> + (lambda* (#:key outputs inputs #:allow-other-keys)
> + (let* ((outpu
Matthew Jordan skribis:
>> One last thing I forgot to mention: building with --rounds=2 shows that
>> the build is not reproducible (some differences in a few .a files.)
>>
>> If you prefer, we can address it in a later patch, though.
>
> Yes I would prefer address this issue in a later patch.
S
Hello,
Matthew Jordan skribis:
>> Also, why do we need this “src” output? IIUC, it contains the source of
>> some of the unit tests, right?
>
> I changed the "src" output to "tests", to clarify.
OK; it’s not clear to me if users need it, but it probably doesn’t hurt
to keep it.
> From 9063100
Good Day,
> One last thing I forgot to mention: building with --rounds=2 shows that
> the build is not reproducible (some differences in a few .a files.)
>
> If you prefer, we can address it in a later patch, though.
Yes I would prefer address this issue in a later patch.
Respectuflly,
--
Matt
Good Day,
> Since there are parts I don’t fully understand, I thought I’d leave it
> up to you. In particular, could you could at least replace the “why” in
> the comments below with explanations of why this is done?
I updated the patch with additional comments as suggested. Let me know
if more
l...@gnu.org (Ludovic Courtès) skribis:
> It seems we’re almost done! I was about to commit it with the changes
> below on top of the patch you posted (they are mostly cosmetic.)
>
> Since there are parts I don’t fully understand, I thought I’d leave it
> up to you. In particular, could you coul
Hi!
It seems we’re almost done! I was about to commit it with the changes
below on top of the patch you posted (they are mostly cosmetic.)
Since there are parts I don’t fully understand, I thought I’d leave it
up to you. In particular, could you could at least replace the “why” in
the comments
Good Day,
I have updated the package in accordance with the suggestions made.
>From 16ff70593c243290d0fe6288e75caeb2897d292c Mon Sep 17 00:00:00 2001
From: Matthew Jordan
Date: Thu, 26 May 2016 08:57:16 -0400
Subject: [PATCH] gnu: Add go@1.4
* gnu/local.mk: Modified file.
* gnu/packages/golang.
On Mon 06 Jun 2016 11:05, Alex Vong writes:
> I think match-lambda is not documented in the guile manual (it only
> appears in an example). While its usage is more or less the same as in
> racket, I think it should be documented. What is your idea?
Yes it should be documented :)
Andy
Hello,
I think match-lambda is not documented in the guile manual (it only
appears in an example). While its usage is more or less the same as in
racket, I think it should be documented. What is your idea?
Thanks,
Alex
On 06/06/2016, Andy Wingo wrote:
> Hi :)
>
> Looking good! I have some styl
Hi :)
Looking good! I have some style nits :)
On Fri 03 Jun 2016 19:39, Matthew Jordan
writes:
> + ;; Removing net/ tests
> + (for-each
> +(lambda (srcfile)
> + (let ((srcfile (string-append "net/" srcfile)))
> + (i
Good Day,
Patch updated as suggested.
>From debf89e7a3456862b6de77e8d948b954d88bfca6 Mon Sep 17 00:00:00 2001
From: Matthew Jordan
Date: Thu, 26 May 2016 08:57:16 -0400
Subject: [PATCH] gnu: Add go@1.4
* gnu/local.mk: Modified file.
* gnu/packages/golang.scm: New file.
Co-author: Efraim Flashn
I don't think this package needs to depend on rc or gccgo. Without rc,
guix will fail to patch the paths to rc, but those are only used on Plan
9 systems. And despite a few references to gccgo in the package
definition, removing it does not actually affect the build at all.
--
Alex Griffin
Hi,
Sounds great :)
On Thu 02 Jun 2016 18:29, Matthew Jordan
writes:
> Also do you mind if I add you name and email an Co-author of these three
> packages?
Sure, sounds fine.
Cheers,
Andy
Hey Andy and everyone,
Just a quick update, will send more details, or some patches later.
Just finished up building go@1.5.4, thanks to last bit you sent. The
runpath now validates in both go@1.4.3 and go@1.5.4. All enabled test
now pass.
>> if Linkshared {
>> seenDirs := ma
Hi,
How goes the patching?
FWIW for Go 1.5 and 1.6, if I remove this test in the linker then I can
build without the LD_LIBRARY_PATH hacks and the tests work for me:
On Sun 29 May 2016 19:49, Andy Wingo writes:
> if Linkshared {
> seenDirs := make(map[string]bool)
>
Hi Andy,
And thanks for code snippet, I'll look at and see how I can integrate it
with the current patches. I'm new to both go and guile, so there is
still much to learn.
Thanks again!
--
Matthew Jordan
Sent with my mu4e
Andy Wingo writes:
> Hi!
>
> I also took a look at this package over t
Hi!
I also took a look at this package over the weekend but was away from
the internet. It turns out that what is needed is to add "-rpath"
invocations to Go itself. The issue is that Go binaries are still
dynamically linked to libgcc_s, but of course there is no useful default
search path in wh
23 matches
Mail list logo
