Hi!
One more thing to keep in mind for the proposal: sometimes we need to merge two
contexts.
https://github.com/golang/go/issues/36503
https://github.com/golang/go/issues/57928
--
WBR, Alex.
--
You received this message because you are subscribed to the Google Groups
Hi!
On Sun, Apr 22, 2018 at 09:48:13PM -0700, nsakth...@gmail.com wrote:
> this pattern requires a language level change to avoid unwanted array copy
[...]
> but shouldn't the language provide an extension to already available syntax
The language is trying to avoid hiding performance-critical th
Hi!
On Fri, Apr 13, 2018 at 12:41:35PM -0700, Lee Painton wrote:
> The only situation in which this limit has seemed unreasonable is when a
> large number of independent cases followed a selection function (a large
> case statement), which was allowed.
Another situation is function which does val
Hi!
On Thu, Apr 12, 2018 at 08:34:28PM -0700, Nick Snyder wrote:
> For (3) though, I am a little hesitant to delete the v1 code from master
> since this will break people go getting v1.
If you wanna support go both for v1 and v2 using current import path then
you've no choice - you have to keep
Hi!
On Wed, Apr 11, 2018 at 10:19:21AM -0700, Ian Lance Taylor wrote:
> You might want to open an issue about this.
Done: https://github.com/golang/go/issues/24808
--
WBR, Alex.
--
You received this message because you are subscribed to the Google Groups
"golang-nuts"
Hi!
On Wed, Apr 11, 2018 at 07:35:50AM -0700, pierspowlesl...@gmail.com wrote:
> So I believe I have found the source of this problem. A method Temporary is
> defined on syscall.Errno, for a reason that is unclear to me ECONNRESET and
> ECONNABORTED are considered temporary. I would definitely c
Hi!
On Tue, Apr 10, 2018 at 07:58:29AM -0700, pierspowlesl...@gmail.com wrote:
> I'm trying to understand what is going on under the hood here.
SO_LINGER (see socket(7)) delay happens. Add this after Accept():
conn, err := listener.Accept()
if err == nil {
err = c
Hi!
On Tue, Apr 10, 2018 at 06:14:46AM -0700, Sun Frank wrote:
> my question is how to reuse underlying TCP connection when doing a lot of
> frequent short HTTP requests using golang?
This happens by default: https://play.golang.org/p/XnzQoGqQlno
> they say the key is to close resp.Body
Yes, b
Hi!
On Mon, Apr 09, 2018 at 08:57:04AM -0700, Paul Lalonde wrote:
> The FindReaderSubmatchIndex may need to read arbitrarily far after the end
> of the match, so if you want multiple matches you'll need a ReadSeeker and
> re-wind to the end of the previous match.
Ah, I see. So I should manually
Hi!
On Mon, Apr 09, 2018 at 04:09:55AM -0700, Tamás Gulácsi wrote:
> A small hack can help: add someting at the beginning of the Reader (with
> io.MultiReader) that won't allow the match on the beginning.
Try to add "\n" at beginning of the stream and then start regexp with \n
instead of ^. If r
Hi!
On Mon, Apr 09, 2018 at 09:25:02AM +0100, roger peppe wrote:
> While we're on the subject, ISTM that FindReaderSubmatchIndex should
> really return []int64 not []int. A bit late now though.
BTW, is there any good example for FindReaderSubmatchIndex?
I've just tried it with strings.Reader and
Hi!
When we run `go test` (without package param) it doesn't redirect
os.Stdout, so it's possible to detect is it a TTY and enable color output:
unix.IoctlGetTermios(int(os.Stdout.Fd()), ioctlReadTermios)
But when we run `go test .` (with package param) it does redirect
os.Stdout, so it's bec
Hi!
On Tue, Apr 03, 2018 at 03:25:44PM -0700, roar...@zillowgroup.com wrote:
> In an effort to make sure code that builds on my local machine will build
> on our build machine, I am looking for a build option to force the compiler
> to only look for imported code in the vendor directory & pkg.
Hi!
On Thu, Mar 29, 2018 at 04:29:09AM -0700, gary.willoug...@victoriaplumb.com
wrote:
> Thanks for the all the responses but I was after a more dynamic solution
> where you don't need to 'register' the types or select them using a switch
> but that it just works.
I believe this is just a wron
Hi!
On Wed, Mar 28, 2018 at 02:13:55PM -0400, Shawn Milochik wrote:
> Thanks for the challenge! I found some unused code in the getItem function,
> and a way to make it a bit shorter by removing the lookup map. However,
> depending on the number of struct types the map may be cleaner for OP.
>
>
Hi!
On Wed, Mar 28, 2018 at 01:38:04PM -0400, Shawn Milochik wrote:
> Yes, you can do this. Perhaps someone knows of some kind of cleaner
> shortcuts to get there, but this works and is easy to understand without
> any magic:
>
> https://play.golang.org/p/tqon7s2KxTz
Here is a bit simplified ver
Hi!
On Sat, Mar 24, 2018 at 03:41:19AM -0700, T L wrote:
> Shouldn't it be that the receive from the only buffer and fill next queuing
> value to the only buffer in one atomic operation?
AFAIR spec doesn't provide such a guarantee, so this isn't a bug.
If you need to know when last value from c
Hi!
On Fri, Mar 23, 2018 at 12:10:55PM -0700, Alex Dvoretskiy wrote:
> For example, I have a multidimensional slice, which I want to modify, but
> keep an original slice untouched.
>
> Do I need to do this all the time or there is another convenient way to do
> this?:
>
> https://play.golang.o
Hi!
On Tue, Mar 20, 2018 at 10:37:40AM -0700, st ov wrote:
> json.Unmarshal(resp.Body, &data)
This one is invalid.
> json.NewDecoder(resp.Body).Decode(&data)
>
> or
>
> b, _ := ioutil.ReadAll(resp.Body)
> json.Unmarshal(b, &data)
In the ReadAll case you'll have to allocate []byte in memory to
Hi!
On Mon, Mar 19, 2018 at 05:52:52AM -0700, Matt Mueller wrote:
> Is it possible to programmatically tell os.Stdin that we've reached EOF
> while you're trying to read from it?
As default os.Stdin is in blocking mode then only way to have it read EOF
after calling Read() on it is to actually p
Hi!
On Thu, Mar 15, 2018 at 12:49:15PM +, roger peppe wrote:
> If it helps, you can set deadlines on *os.File now, as of Go 1.10:
>
>https://play.golang.org/p/h3Pg9Ql0CMB
>
> I don't see a way to cancel without deciding the deadline in advance though.
Why? What makes os.Stdin so special
Hi!
On Thu, Mar 15, 2018 at 09:09:18AM -0700, Rio wrote:
> Yeah, there's no way to avoid the kernel buffer waste.
You can tune kernel buffer size to minimize it.
> > I did such a custom poller for a network proxy. It was a pain
> > to get right (I only did the Eww!poll version).
>
> I tried an
Hi!
Many years ago I've implemented something similar in Perl, which was later
released as https://metacpan.org/pod/DBIx::SecureCGI. Nowadays I suppose
best way to do something like this is using GraphQL (for ex.
https://github.com/graphql-go/graphql).
--
WBR, Alex.
--
Hi!
On Sun, Feb 25, 2018 at 01:51:08AM -0800, Maxim Ivanov wrote:
> > type Time struct {
> > Err error
> > *time.Time
> > }
> Problem is that it continues to be product type, nothing really enforces
> that you can't use t as a time.Time when t.E
Hi!
On Sat, Feb 24, 2018 at 02:02:37PM -0800, Bakul Shah wrote:
> r := os.Open(filename)
> if isError(r) { return r.(error) }
> f := r.(*os.File) // or better: var f *os.File; f = r
>
> Error checking being a common pattern, isError() can be added as a builtin
> or trivially in errors pkg. Likely
Hi!
On Fri, Feb 23, 2018 at 06:23:33PM -0800, Spencer Nelson wrote:
> Is it possible for me to have different minor/patch versions for different
> modules in one repository? For example, suppose I have this:
>
> github.com/my/thing/go.mod
> github.com/my/thing/blue/go.mod
>
> And suppos
26 matches
Mail list logo
