Re: [go-nuts] generics question
Thank you both!
This works, see my code below. Followup question: is there a way to refer
to the new type without having to list both the element type and the
pointer type separately? Below I have to write C[A, *A], which looks
slightly ugly. And in my real application something like Creator
OrderedArray[ProperName] would look much more readable than Creator
OrderedArray[ProperName, *ProperName] .
Many thanks,
Jochen
type Type interface{}
type Settable[E Type] interface {
*E
Set(int)
}
type C[E any, P Settable[E]] struct {
Val []E
}
func (c *C[E, P]) Set(v int) {
for i := range c.Val {
var ptr P = &(c.Val[i])
ptr.Set(v)
}
}
var c = C[A, *A]{Val: []A{1, 2, 3}}
On Tuesday 23 April 2024 at 03:36:25 UTC+1 Nagaev Boris wrote:
> On Mon, Apr 22, 2024 at 9:54 PM Ian Lance Taylor wrote:
> >
> > On Mon, Apr 22, 2024 at 2:25 PM Jochen Voss wrote:
> > >
> > > Using generics, can I somehow write a constraint which says that *T
> (instead of T) implements a certain interface? The following code
> illustrated what I'm trying to do:
> > >
> > > type A int
> > >
> > > func (a *A) Set(x int) {
> > > *a = A(x)
> > > }
> > >
> > > type B string
> > >
> > > func (b *B) Set(x int) {
> > > *b = B(strconv.Itoa(x))
> > > }
> > >
> > > type C1 struct {
> > > Val []A
> > > }
> > >
> > > func (c *C1) Set(v int) {
> > > for i := range c.Val {
> > > c.Val[i].Set(v)
> > > }
> > > }
> > >
> > > type C2 struct {
> > > Val []B
> > > }
> > >
> > > func (c *C2) Set(v int) {
> > > for i := range c.Val {
> > > c.Val[i].Set(v)
> > > }
> > > }
> > >
> > > I would like to use generics to use a single definition for the
> methods which here are func (c *C1) Set(v int) and func (c *C2) Set(v int).
> (My real code has many base types, instead of just A and B.) How can I do
> this?
> > >
> > > I tried the naive approach:
> > >
> > > type C[T interface{ Set(int) }] struct {
> > > Val []T
> > > }
> > >
> > > but when I try to use the type C[A] now, I get the error message "A
> does not satisfy interface{Set(int)} (method Set has pointer receiver)".
> >
> >
> > type C[P interface {
> > *E
> > Set(int)
> > }, E any] struct {
> > Val []P
> > }
> >
> > Ian
> >
>
> I think it should be this (s/Val []P/Val []E/):
>
> type C[P interface {
> *E
> Set(int)
> }, E any] struct {
> Val []E
> }
>
>
>
> --
> Best regards,
> Boris Nagaev
>
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/c5571d6c-d4d1-4cb9-9617-d5405f68d66an%40googlegroups.com.
[go-nuts] generics question
Using generics, can I somehow write a constraint which says that *T
(instead of T) implements a certain interface? The following code
illustrated what I'm trying to do:
type A int
func (a *A) Set(x int) {
*a = A(x)
}
type B string
func (b *B) Set(x int) {
*b = B(strconv.Itoa(x))
}
type C1 struct {
Val []A
}
func (c *C1) Set(v int) {
for i := range c.Val {
c.Val[i].Set(v)
}
}
type C2 struct {
Val []B
}
func (c *C2) Set(v int) {
for i := range c.Val {
c.Val[i].Set(v)
}
}
I would like to use generics to use a single definition for the methods
which here are func (c *C1) Set(v int) and func (c *C2) Set(v int). (My
real code has many base types, instead of just A and B.) How can I do this?
I tried the naive approach:
type C[T interface{ Set(int) }] struct {
Val []T
}
but when I try to use the type C[A] now, I get the error message "A does
not satisfy interface{Set(int)} (method Set has pointer receiver)".
Many thanks,
Jochen
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/e6c1647a-9ade-496e-85b5-0e96563b5e22n%40googlegroups.com.
[go-nuts] Re: heap profiling does not work on M2 MacBook air?
I've now filed https://github.com/golang/go/issues/65328 to report this.
On Friday 3 March 2023 at 09:47:26 UTC Jochen Voss wrote:
> Hi Peter,
>
> Thanks again for looking into this, and for your help!
>
> You mention that the blog post was long ago. I wonder whether the code to
> write profile information is covered by the compatibility promise. If so,
> probably the legacy WriteHeapProfile function could be made to implement
> the required code to keep this working. Should a bug report be filed to
> this effect?
>
> Similarly, should a bug report be filed about the mistake you found in the
> pprof package documentation?
>
> All the best,
> Jochen
> On Thursday, 2 March 2023 at 15:36:24 UTC peterGo wrote:
>
>> On Thursday, March 2, 2023 at 2:03:22 AM UTC-5 Jochen Voss wrote:
>>
>> Hi Peter,
>>
>> Thanks a lot, giving the "-alloc_space" option makes all the difference!
>> With this option, it also works for me.
>>
>> I wonder whether it meant to be the way that you have to give this
>> option. Maybe something broke? In the blog entry
>> https://go.dev/blog/pprof they didn't need this option to get the
>> profile output.
>>
>>
>> The Go blog post at https://go.dev/blog/pprof was written in 2011; it is
>> now 2023. For up-to-date information, see the latest pprof package
>> documentation at https://pkg.go.dev/runtime/pprof@latest.
>>
>> The pprof package documentation is incorrect:
>>
>> To add equivalent [go test] profiling support to a standalone program,
>> add code like the following to your main function:
>>
>> // ...
>> runtime.GC() // get up-to-date statistics
>> if err := pprof.WriteHeapProfile(f); err != nil {
>> log.Fatal("could not write memory profile: ", err)
>> }
>> // ...
>>
>> To obtain equivalent results, replace the above code with:
>>
>> // ...
>> runtime.GC() // get up-to-date statistics
>> if allocs := pprof.Lookup("allocs"); allocs == nil {
>> log.Fatal("could not lookup memory profile: ")
>> } else if err := allocs.WriteTo(f, 0); err != nil {
>>
>> log.Fatal("could not write memory profile: ", err)
>> }
>> // ...
>>
>> For your example,
>>
>> yyy.go: https://go.dev/play/p/epy4c3et1Io
>>
>> $ go build yyy.go && ./yyy
>> $ go tool pprof mem.prof
>> File: yyy
>> Type: alloc_space
>> Time: Mar 2, 2023 at 10:02am (EST)
>>
>> Entering interactive mode (type "help" for commands, "o" for options)
>> (pprof) top
>> Showing nodes accounting for 6975.42MB, 98.93% of 7050.77MB total
>> Dropped 62 nodes (cum <= 35.25MB)
>>
>> Showing top 10 nodes out of 18
>> flat flat% sum%cum cum%
>> 3564.88MB 50.56% 50.56% 3564.88MB 50.56%
>> golang.org/x/exp/slices.Insert[..
>> <http://golang.org/x/exp/slices.Insert%5B..>.] (inline)
>> 3122.14MB 44.28% 94.84% 3122.14MB 44.28%
>> seehuhn.de/go/layout.(*Skip).Minus
>> 117.01MB 1.66% 96.50% 117.01MB 1.66%
>> seehuhn.de/go/layout.(*Skip).Clone (inline)
>> 115.51MB 1.64% 98.14% 6919.54MB 98.14%
>> seehuhn.de/go/layout.(*knuthPlassLineBreaker).Run
>>52.89MB 0.75% 98.89%65.63MB 0.93% compress/flate.NewWriter
>> 2.50MB 0.035% 98.92% 6926.54MB 98.24%
>> seehuhn.de/go/layout.(*Engine).EndParagraph
>> 0.50MB 0.0071% 98.93% 7050.27MB 100% main.main
>> 0 0% 98.93%65.63MB 0.93% compress/flate.NewWriterDict
>> 0 0% 98.93%65.63MB 0.93% compress/zlib.(*Writer).Write
>> 0 0% 98.93%65.63MB 0.93%
>> compress/zlib.(*Writer).writeHeader
>> (pprof) quit
>> $
>>
>> Peter
>>
>>
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/52a65cdb-afee-4101-9602-42c0d17a3088n%40googlegroups.com.
Re: [go-nuts] weblist in go tool pprof does not work for me
On 26 Sep 2023, at 6:20, Sergey Gorlov wrote: > Did you find a solution to this problem? I have the same problem with > weblist. No, I didn't. I think it's just broken. If I find time, I'll submit a bug report (unless somebody beats me to it). All the best, Jochen -- http://seehuhn.de/ -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/C56812A9-7A5C-44DB-9061-985F18B3029C%40gmail.com.
[go-nuts] (minor) broken link in the Go 1.21 release notes.
Hello, Not really important, but there is a somewhat broken link in the Go 1.21 release notes. The introduction says "See “Go versions” in the “Go Toolchains” documentation for details", and this links to https://go.dev/doc/toolchain#versions The link goes to the beginning of the page, not to the intended section, because "versions" is a typo. The correct link is https://go.dev/doc/toolchain#version (without the trailing "s"). Not sure if this is worth fixing, but if anybody feels so inclined, the extra "s" could be removed to make the link work. All the best, Jochen -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/5ccd2caa-1e26-4573-8c83-55c541266ea4n%40googlegroups.com.
[go-nuts] failed fuzz tests not reproducible when re-run
Dear all, Recently I find quite often that after "go test -fuzz" reported an error, the command shown to re-run the test does not reproduce the failure. For example, just now I got the following: voss@dumpling [..nt/tounicode2] go test -fuzz FuzzToUnicode fuzz: elapsed: 0s, gathering baseline coverage: 0/2 completed fuzz: elapsed: 0s, gathering baseline coverage: 2/2 completed, now fuzzing with 10 workers fuzz: elapsed: 3s, execs: 280461 (93481/sec), new interesting: 24 (total: 26) fuzz: elapsed: 5s, execs: 410320 (72348/sec), new interesting: 60 (total: 62) --- FAIL: FuzzToUnicode (4.80s) --- FAIL: FuzzToUnicode (0.00s) tounicode_test.go:129: template: tounicode:26:2: executing "tounicode" at : error calling Single: runtime error: integer divide by zero Failing input written to testdata/fuzz/FuzzToUnicode/2566873a28045c1b To re-run: go test -run=FuzzToUnicode/2566873a28045c1b FAIL exit status 1 FAIL seehuhn.de/go/pdf/font/tounicode2 4.957s voss@dumpling [..nt/tounicode2] go test -run=FuzzToUnicode/2566873a28045c1b PASS ok seehuhn.de/go/pdf/font/tounicode2 0.122s The fuzzer reported a failure, and when I re-run the test I get a "PASS". How to debug such a problem? I understand that my fuzz function may somehow be non-deterministic, but so far I have not found any cause of non-determinism and re-running the test 10 times gives me 10 passes. Also, for a bug the fuzzer reported earlier, I found it very hard to believe that the given input could make the fuzz function reach the site of the error message. Could it be that the fuzzer sometimes looses track of which input belongs to which fuzzing run, and then reports the wrong input? If somebody wants to play with this, here is how to reproduce (up to randomness) the fuzzing run shown above: git clone https://github.com/seehuhn/go-pdf.git cd go-pdf/ git reset --hard 26c235ad cd font/tounicode2/ go test -fuzz FuzzToUnicode Any suggestions on how to debug such errors would be most welcome. Many thanks, Jochen -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/3fe3725b-e59e-4b1f-a2e9-0f28c15ea3d4n%40googlegroups.com.
[go-nuts] Re: testing whether two maps are the same object
Thanks Jason, I get your point about DeepEqual now :) On Tuesday, 18 July 2023 at 16:52:51 UTC+1 Jason Phillips wrote: > Also note: reflect.DeepEqual doesn't *just* compare the contents of the > map. It only compares contents if the maps aren't "the same map object". > From the documentation: > > > Map values are deeply equal when all of the following are true: they are > both nil or both non-nil, they have the same length, and either they are > the same map object or their corresponding keys (matched using Go equality) > map to deeply equal values. > > Comparing if they're "the same map object" is done in the way described > earlier, which seems exactly what you're looking for. > > On Tuesday, July 18, 2023 at 11:44:53 AM UTC-4 Jason Phillips wrote: > >> reflect.Value.UnsafePointer() is probably not safer than using unsafe >> directly, assuming you're using unsafe in a way that doesn't break the >> rules. Reflect is doing effectively the same unsafe.Pointer conversion >> under the hood [1]. It's certainly easier on the eyes, in my opinion, >> though. >> >> [1] - >> https://cs.opensource.google/go/go/+/refs/tags/go1.20.6:src/reflect/value.go;l=103 >> >> On Tuesday, July 18, 2023 at 11:25:42 AM UTC-4 Stephen Illingworth wrote: >> >>> >>> I like that. I think it's is quite a smart way of doing it, >>> >>> I don't think you need to check both maps when choosing a key. Once >>> you've found a candidate key in one map, you can test the other map and if >>> it *does* exist then the two maps aren't equal. You've then saved the >>> insertion and deletion step. I would also prefer to use rand.Int() rather >>> than a linear search. >>> >>> https://go.dev/play/p/ftBGgQMuvvC >>> >>> >>> On Tuesday, 18 July 2023 at 15:35:38 UTC+1 Jochen Voss wrote: >>> >>>> Dear all, >>>> >>>> To implement the "eq" operator in a simple PostScript interpreter, I >>>> need to determine whether two maps are the same object. >>>> >>>> This can be done by adding a new element to one map, and checking >>>> whether the new entry also appears in the second map, but as you can >>>> imagine, the resulting code is quite ugly. See >>>> https://go.dev/play/p/AfembYDt3en for an implementation of this idea. >>>> >>>> Is there a better way? >>>> >>>> Many thanks, >>>> Jochen >>>> >>>> -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/c8b9aac5-6a53-472f-9c5d-497a23575a79n%40googlegroups.com.
Re: [go-nuts] testing whether two maps are the same object
Dear Jason, Thanks for the idea of using reflect. Do you know whether `reflect.ValueOf(a).UnsafePointer()` is somehow "safer" than `*(*uintptr)(unsafe.Pointer())`? reflect.DeepEqual will not work, since it just compares contents of the maps: https://go.dev/play/p/tE_qZI2cKEd All the best, Jochen On Tuesday, 18 July 2023 at 15:52:24 UTC+1 Jason Phillips wrote: > You can also use the reflect package rather than (directly) reaching for > unsafe. The reflect.DeepEqual function does something along the lines of: > https://go.dev/play/p/IVt0Z-mxugh > > On Tuesday, July 18, 2023 at 10:45:18 AM UTC-4 Jan Mercl wrote: > >> On Tue, Jul 18, 2023 at 4:35 PM Jochen Voss wrote: >> >> > Is there a better way? >> >> I have never been here and please don't do this: >> https://go.dev/play/p/x4QYJubXMnQ >> > -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/30202b52-554e-462e-aaa4-dfe7f4d23554n%40googlegroups.com.
[go-nuts] testing whether two maps are the same object
Dear all, To implement the "eq" operator in a simple PostScript interpreter, I need to determine whether two maps are the same object. This can be done by adding a new element to one map, and checking whether the new entry also appears in the second map, but as you can imagine, the resulting code is quite ugly. See https://go.dev/play/p/AfembYDt3en for an implementation of this idea. Is there a better way? Many thanks, Jochen -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/9f5e5eb6-faee-4918-9a7f-e441ac15ecadn%40googlegroups.com.
[go-nuts] Re: finding the call graph of a method
Ah, found it! If I write types.NewPointer(reader.Type()) instead of just
reader.Type(), things work better.
On Tuesday, 18 April 2023 at 20:34:30 UTC+1 Jochen Voss wrote:
> Dear all,
>
> I'm trying to find the callgraph of a method, using the following code:
>
> package main
>
> import (
> "log"
>
> "golang.org/x/tools/go/callgraph"
> "golang.org/x/tools/go/packages"
> "golang.org/x/tools/go/ssa/ssautil"
> )
>
> func main() {
> cfg := {Mode: packages.LoadSyntax}
> initial, err := packages.Load(cfg, "seehuhn.de/go/pdf")
> if err != nil {
> log.Fatal(err)
> } else if packages.PrintErrors(initial) > 0 {
> log.Fatalf("packages contain errors")
> }
>
> prog, pkgs := ssautil.Packages(initial, 0)
> if len(pkgs) != 1 {
> log.Fatalf("expected 1 package, got %d", len(pkgs))
> }
> pkg := pkgs[0]
> pkg.Build()
>
> reader := pkg.Members["Reader"]
>
> f := prog.LookupMethod(reader.Type(), pkg.Pkg, "readXRef")
> g := callgraph.New(f)
> _ = g
> }
>
> (Also on the playgound at https://go.dev/play/p/_S-oKAJ6YNO, but
> resulting in some fancy error messages there.)
>
> This fails with the following error message: panic:
> seehuhn.de/go/pdf.Reader has no method seehuhn.de/go/pdf.readXRef
>
> What am I doing wrong?
>
> All the best,
> Jochen
>
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/c128c8fb-bb40-4cbe-86a1-f6da0c88a78bn%40googlegroups.com.
[go-nuts] finding the call graph of a method
Dear all,
I'm trying to find the callgraph of a method, using the following code:
package main
import (
"log"
"golang.org/x/tools/go/callgraph"
"golang.org/x/tools/go/packages"
"golang.org/x/tools/go/ssa/ssautil"
)
func main() {
cfg := {Mode: packages.LoadSyntax}
initial, err := packages.Load(cfg, "seehuhn.de/go/pdf")
if err != nil {
log.Fatal(err)
} else if packages.PrintErrors(initial) > 0 {
log.Fatalf("packages contain errors")
}
prog, pkgs := ssautil.Packages(initial, 0)
if len(pkgs) != 1 {
log.Fatalf("expected 1 package, got %d", len(pkgs))
}
pkg := pkgs[0]
pkg.Build()
reader := pkg.Members["Reader"]
f := prog.LookupMethod(reader.Type(), pkg.Pkg, "readXRef")
g := callgraph.New(f)
_ = g
}
(Also on the playgound at https://go.dev/play/p/_S-oKAJ6YNO, but resulting
in some fancy error messages there.)
This fails with the following error message: panic:
seehuhn.de/go/pdf.Reader has no method seehuhn.de/go/pdf.readXRef
What am I doing wrong?
All the best,
Jochen
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/68b8c937-7bce-44d4-81fb-2507b55fb0bcn%40googlegroups.com.
Re: [go-nuts] getting the size of an io.ReaderAt?
Dear Bruno,
Thanks for your answer. Originally I had a separate size argument for the
function which opens a new reader, something like this:
func NewReader(data io.ReaderAt, size int64, opt *ReaderOptions)
(*Reader, error)
But this makes it a bit annoying to create a reader for a file, because you
always need the additional Stat() call before, and also it looks a bit
weird, I though. So my idea was to move the code for checking the size
into the function, and to simplify the API:
func NewReader(data io.ReaderAt, opt *ReaderOptions) (*Reader, error)
This way the library does the work once, rather than every caller having to
do this separately.
All the best,
Jochen
On Friday, 31 March 2023 at 21:38:04 UTC+1 Bruno Albuquerque wrote:
> Not a direct answer to your question (your code looks like a reasonable
> implementation modulo any bugs I did not notice) but: Aren't you over
> engineering things? If the source is a PDF file, why not also pass the size
> to the function instead of doing all this work to get it? At some point you
> have to open the file, right? And checking the size of a file is trivial
> (it is just a stat call) and then you do not need to do that.
>
>
> On Fri, Mar 31, 2023 at 4:30 PM Jochen Voss wrote:
>
>> Dear all,
>>
>> I am trying to get the size of an io.ReaderAt, i.e. the offset after
>> which no more data can be read. I have some working (?) code (
>> https://go.dev/play/p/wTouYbaJ7RG , also reproduced below), but I am not
>> sure whether what I do is correct and the best way to do this. Some
>> questions:
>>
>>- Does the standard library provide a way to get the size of an
>>io,ReaderAt?
>>- Is my code correct?
>>- Is there a better way to do this?
>>- If a type provides not only a ReadAt() method, but also a Size()
>>method, would it be save to assume that Size() returns how many bytes are
>>accessible via ReadAt()? This seems to work for bytes.Reader and
>>strings.Reader, but there may be types out there where Size() does
>>something different?
>>- If ReadAt(p, x) returns io.EOF for an offset x, is it then
>>guaranteed that then ReadAt(p, y) also returns io.EOF for all y > x?
>>Or could there be different error messages, or "files with holes", or
>>whatnot?
>>- Which types in the standard library provide ReadAt methods? I know
>>of os.File, strings.Reader, and bytes.Reader. Any others?
>>
>> For context: this is for reading the cross reference table of PDF files,
>> which have to be located by following some convoluted route starting *from
>> the end* of the PDF file.
>>
>> Many thanks,
>> Jochen
>>
>> func getSize(r io.ReaderAt) (int64, error) {
>> if f, ok := r.(*os.File); ok {
>> fi, err := f.Stat()
>> if err != nil {
>> return 0, err
>> }
>> return fi.Size(), nil
>> }
>> if b, ok := r.(*bytes.Reader); ok {
>> return int64(b.Size()), nil
>> }
>> if s, ok := r.(*strings.Reader); ok {
>> return int64(s.Size()), nil
>> }
>>
>> buf := make([]byte, 1024)
>> n, err := r.ReadAt(buf, 0)
>> if err == io.EOF {
>> return int64(n), nil
>> } else if err != nil {
>> return 0, err
>> }
>>
>> lowerBound := int64(n) // all bytes before lowerBound are known to be
>> present
>> var upperBound int64 // at least one byte before upperBound is known to
>> be missing
>> for {
>> test := 2 * lowerBound
>> _, err := r.ReadAt(buf[:1], test-1)
>> if err == io.EOF {
>> upperBound = test
>> break
>> } else if err != nil {
>> return 0, err
>> }
>> lowerBound = test
>> }
>>
>> for lowerBound+1 < upperBound {
>> test := (lowerBound + upperBound + 1) / 2
>> _, err := r.ReadAt(buf[:1], test-1)
>> if err == io.EOF {
>> upperBound = test
>> } else if err != nil {
>> return 0, err
>> } else {
>> lowerBound = test
>> }
>> }
>> return lowerBound, nil
>> }
>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "golang-nuts" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to golang-nuts...@googlegroups.com.
>> To view this discussion on the web visit
>> https://groups.google.com/d/msgid/golang-nuts/aa604ac0-3542-4a9e-adef-a9eaecfd8170n%40googlegroups.com
>>
>> <https://groups.google.com/d/msgid/golang-nuts/aa604ac0-3542-4a9e-adef-a9eaecfd8170n%40googlegroups.com?utm_medium=email_source=footer>
>> .
>>
>
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/1be784af-63e5-4238-bd96-ef3123318d68n%40googlegroups.com.
[go-nuts] getting the size of an io.ReaderAt?
Dear all,
I am trying to get the size of an io.ReaderAt, i.e. the offset after which
no more data can be read. I have some working (?) code
(https://go.dev/play/p/wTouYbaJ7RG , also reproduced below), but I am not
sure whether what I do is correct and the best way to do this. Some
questions:
- Does the standard library provide a way to get the size of an
io,ReaderAt?
- Is my code correct?
- Is there a better way to do this?
- If a type provides not only a ReadAt() method, but also a Size()
method, would it be save to assume that Size() returns how many bytes are
accessible via ReadAt()? This seems to work for bytes.Reader and
strings.Reader, but there may be types out there where Size() does
something different?
- If ReadAt(p, x) returns io.EOF for an offset x, is it then guaranteed
that then ReadAt(p, y) also returns io.EOF for all y > x? Or could
there be different error messages, or "files with holes", or whatnot?
- Which types in the standard library provide ReadAt methods? I know of
os.File, strings.Reader, and bytes.Reader. Any others?
For context: this is for reading the cross reference table of PDF files,
which have to be located by following some convoluted route starting *from
the end* of the PDF file.
Many thanks,
Jochen
func getSize(r io.ReaderAt) (int64, error) {
if f, ok := r.(*os.File); ok {
fi, err := f.Stat()
if err != nil {
return 0, err
}
return fi.Size(), nil
}
if b, ok := r.(*bytes.Reader); ok {
return int64(b.Size()), nil
}
if s, ok := r.(*strings.Reader); ok {
return int64(s.Size()), nil
}
buf := make([]byte, 1024)
n, err := r.ReadAt(buf, 0)
if err == io.EOF {
return int64(n), nil
} else if err != nil {
return 0, err
}
lowerBound := int64(n) // all bytes before lowerBound are known to be
present
var upperBound int64 // at least one byte before upperBound is known to
be missing
for {
test := 2 * lowerBound
_, err := r.ReadAt(buf[:1], test-1)
if err == io.EOF {
upperBound = test
break
} else if err != nil {
return 0, err
}
lowerBound = test
}
for lowerBound+1 < upperBound {
test := (lowerBound + upperBound + 1) / 2
_, err := r.ReadAt(buf[:1], test-1)
if err == io.EOF {
upperBound = test
} else if err != nil {
return 0, err
} else {
lowerBound = test
}
}
return lowerBound, nil
}
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/aa604ac0-3542-4a9e-adef-a9eaecfd8170n%40googlegroups.com.
[go-nuts] Re: heap profiling does not work on M2 MacBook air?
Hi Peter,
Thanks again for looking into this, and for your help!
You mention that the blog post was long ago. I wonder whether the code to
write profile information is covered by the compatibility promise. If so,
probably the legacy WriteHeapProfile function could be made to implement
the required code to keep this working. Should a bug report be filed to
this effect?
Similarly, should a bug report be filed about the mistake you found in the
pprof package documentation?
All the best,
Jochen
On Thursday, 2 March 2023 at 15:36:24 UTC peterGo wrote:
> On Thursday, March 2, 2023 at 2:03:22 AM UTC-5 Jochen Voss wrote:
>
> Hi Peter,
>
> Thanks a lot, giving the "-alloc_space" option makes all the difference!
> With this option, it also works for me.
>
> I wonder whether it meant to be the way that you have to give this
> option. Maybe something broke? In the blog entry
> https://go.dev/blog/pprof they didn't need this option to get the profile
> output.
>
>
> The Go blog post at https://go.dev/blog/pprof was written in 2011; it is
> now 2023. For up-to-date information, see the latest pprof package
> documentation at https://pkg.go.dev/runtime/pprof@latest.
>
> The pprof package documentation is incorrect:
>
> To add equivalent [go test] profiling support to a standalone program, add
> code like the following to your main function:
>
> // ...
> runtime.GC() // get up-to-date statistics
> if err := pprof.WriteHeapProfile(f); err != nil {
> log.Fatal("could not write memory profile: ", err)
> }
> // ...
>
> To obtain equivalent results, replace the above code with:
>
> // ...
> runtime.GC() // get up-to-date statistics
> if allocs := pprof.Lookup("allocs"); allocs == nil {
> log.Fatal("could not lookup memory profile: ")
> } else if err := allocs.WriteTo(f, 0); err != nil {
>
> log.Fatal("could not write memory profile: ", err)
> }
> // ...
>
> For your example,
>
> yyy.go: https://go.dev/play/p/epy4c3et1Io
>
> $ go build yyy.go && ./yyy
> $ go tool pprof mem.prof
> File: yyy
> Type: alloc_space
> Time: Mar 2, 2023 at 10:02am (EST)
>
> Entering interactive mode (type "help" for commands, "o" for options)
> (pprof) top
> Showing nodes accounting for 6975.42MB, 98.93% of 7050.77MB total
> Dropped 62 nodes (cum <= 35.25MB)
>
> Showing top 10 nodes out of 18
> flat flat% sum%cum cum%
> 3564.88MB 50.56% 50.56% 3564.88MB 50.56%
> golang.org/x/exp/slices.Insert[..
> <http://golang.org/x/exp/slices.Insert%5B..>.] (inline)
> 3122.14MB 44.28% 94.84% 3122.14MB 44.28%
> seehuhn.de/go/layout.(*Skip).Minus
> 117.01MB 1.66% 96.50% 117.01MB 1.66%
> seehuhn.de/go/layout.(*Skip).Clone (inline)
> 115.51MB 1.64% 98.14% 6919.54MB 98.14%
> seehuhn.de/go/layout.(*knuthPlassLineBreaker).Run
>52.89MB 0.75% 98.89%65.63MB 0.93% compress/flate.NewWriter
> 2.50MB 0.035% 98.92% 6926.54MB 98.24%
> seehuhn.de/go/layout.(*Engine).EndParagraph
> 0.50MB 0.0071% 98.93% 7050.27MB 100% main.main
> 0 0% 98.93%65.63MB 0.93% compress/flate.NewWriterDict
> 0 0% 98.93%65.63MB 0.93% compress/zlib.(*Writer).Write
> 0 0% 98.93%65.63MB 0.93%
> compress/zlib.(*Writer).writeHeader
> (pprof) quit
> $
>
> Peter
>
>
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/a06065fb-ae57-4c52-b182-b444e9e32bcdn%40googlegroups.com.
[go-nuts] Re: heap profiling does not work on M2 MacBook air?
Hi Peter,
Thanks a lot, giving the "-alloc_space" option makes all the difference!
With this option, it also works for me.
I wonder whether it meant to be the way that you have to give this option.
Maybe something broke? In the blog entry https://go.dev/blog/pprof they
didn't need this option to get the profile output.
All the best,
Jochen
On Thursday, 2 March 2023 at 04:12:24 UTC peterGo wrote:
> Jochen Voss,
>
> On linux/amd64
>
> xxx.go: https://go.dev/play/p/Wq_OU49LVQZ
>
> $ go build xxx.go && ./xxx
>
> $ go tool pprof xxx mem.prof
> File: xxx
> Type: inuse_space
> Time: Mar 1, 2023 at 11:03pm (EST)
>
> No samples were found with the default sample value type.
> Try "sample_index" command to analyze different sample values.
> Entering interactive mode (type "help" for commands, "o" for options)
> (pprof) quit
>
> $ go tool pprof -alloc_space xxx mem.prof
> File: xxx
> Type: alloc_space
> Time: Mar 1, 2023 at 11:03pm (EST)
>
> Entering interactive mode (type "help" for commands, "o" for options)
> (pprof) top
> Showing nodes accounting for 6919.54MB, 98.93% of 6994.32MB total
> Dropped 77 nodes (cum <= 34.97MB)
> Showing top 10 nodes out of 18
>
> flat flat% sum%cum cum%
> 3564.91MB 50.97% 50.97% 3564.91MB 50.97%
> golang.org/x/exp/slices.Insert[..
> <http://golang.org/x/exp/slices.Insert%5B..>.] (inline)
> 3073.64MB 43.94% 94.91% 3073.64MB 43.94%
> seehuhn.de/go/layout.(*Skip).Minus
> 123.51MB 1.77% 96.68% 123.51MB 1.77%
> seehuhn.de/go/layout.(*Skip).Clone (inline)
> 104.50MB 1.49% 98.17% 6866.56MB 98.17%
> seehuhn.de/go/layout.(*knuthPlassLineBreaker).Run
>48.48MB 0.69% 98.87%60.68MB 0.87% compress/flate.NewWriter
> 4.50MB 0.064% 98.93% 6878.06MB 98.34%
> seehuhn.de/go/layout.(*Engine).EndParagraph
> 0 0% 98.93%60.68MB 0.87% compress/flate.NewWriterDict
> 0 0% 98.93%60.68MB 0.87% compress/zlib.(*Writer).Write
> 0 0% 98.93%60.68MB 0.87%
> compress/zlib.(*Writer).writeHeader
> 0 0% 98.93%56.12MB 0.8% fmt.Fprintln
> (pprof) quit
> $
>
> Peter
>
> On Wednesday, March 1, 2023 at 8:55:30 AM UTC-5 Jochen Voss wrote:
>
>> Dear all,
>>
>> I'm trying to profile memory use of a program, following the instructions
>> at https://go.dev/blog/pprof , but I can't get memory profiling to
>> work. Am I doing things wrong, or is this broken?
>>
>> Simplified code is at https://go.dev/play/p/Wq_OU49LVQZ . (The code
>> doesn't run on the playground, but you can download it and run it locally.)
>>
>> Following the advice from https://pkg.go.dev/runtime/pprof I added the
>> following code to the end of my main() function:
>>
>> f, err := os.Create("mem.prof")
>> if err != nil {
>> log.Fatal("could not create memory profile: ", err)
>> }
>> runtime.GC() // get up-to-date statistics
>> if err := pprof.WriteHeapProfile(f); err != nil {
>> log.Fatal("could not write memory profile: ", err)
>> }
>> err = f.Close()
>> if err != nil {
>> log.Fatal(err)
>> }
>>
>> When I run the code, this gives me a "mem.prof" file (5084 bytes). But
>> when I start "go tool pprof" on this file, I get
>>
>> >>> go tool pprof xxx mem.prof
>> File: xxx
>> Type: inuse_space
>> Time: Mar 1, 2023 at 1:15pm (GMT)
>> No samples were found with the default sample value type.
>> Try "sample_index" command to analyze different sample values.
>> Entering interactive mode (type "help" for commands, "o" for options)
>> (pprof) top10
>> Showing nodes accounting for 0, 0% of 0 total
>> flat flat% sum%cum cum%
>>
>> There seem to be no samples in this file.
>>
>> What am I doing wrong?
>>
>> All the best,
>> Jochen
>>
>>
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/99230d35-8f0a-41b4-9fb5-b0adb365b8c4n%40googlegroups.com.
Re: [go-nuts] Re: heap profiling does not work on M2 MacBook air?
Dear Sean,
Thanks for trying this out. Your result is different from mine, but I
don't think this can be right. According to the output, the only
allocation would be inside the regexp module. But there should be many
more allocations (for example in line 56 of the code)!
All the best,
Jochen
On Wednesday, 1 March 2023 at 17:44:30 UTC Sean Liao wrote:
> fwiw, that code works on my machine :tm:
>
> 1:40:47 ~/tmp/testrepo0120 0:00:27
> main » go run .
> 1:40:58 ~/tmp/testrepo0120 0:00:08
> main » go tool pprof mem.prof
> File: testrepo0120
> Type: inuse_space
> Time: Mar 2, 2023 at 1:40am (CST)
>
> Entering interactive mode (type "help" for commands, "o" for options)
> (pprof) top10
> Showing nodes accounting for 513.31kB, 100% of 513.31kB total
>
> flat flat% sum%cum cum%
> 513.31kB 100% 100% 513.31kB 100% regexp/syntax.(*compiler).inst
> (inline)
> 0 0% 100% 513.31kB 100% regexp.Compile (inline)
> 0 0% 100% 513.31kB 100% regexp.MustCompile
> 0 0% 100% 513.31kB 100% regexp.compile
> 0 0% 100% 513.31kB 100%
> regexp/syntax.(*compiler).compile
> 0 0% 100% 513.31kB 100% regexp/syntax.(*compiler).rune
> 0 0% 100% 513.31kB 100% regexp/syntax.Compile
> 0 0% 100% 513.31kB 100% runtime.doInit
> 0 0% 100% 513.31kB 100% runtime.main
> 0 0% 100% 513.31kB 100%
> seehuhn.de/go/pdf/font/tounicode.init
>
> - sean
>
>
> On Wed, Mar 1, 2023 at 11:25 PM Jochen Voss wrote:
>
>> The problem also occurs on AMD64 Linux, so it's not architecture
>> specific. Hints would be most welcome!
>>
>> On Wednesday, 1 March 2023 at 13:55:30 UTC Jochen Voss wrote:
>>
>>> Dear all,
>>>
>>> I'm trying to profile memory use of a program, following the
>>> instructions at https://go.dev/blog/pprof , but I can't get memory
>>> profiling to work. Am I doing things wrong, or is this broken?
>>>
>>> Simplified code is at https://go.dev/play/p/Wq_OU49LVQZ . (The code
>>> doesn't run on the playground, but you can download it and run it locally.)
>>>
>>> Following the advice from https://pkg.go.dev/runtime/pprof I added the
>>> following code to the end of my main() function:
>>>
>>> f, err := os.Create("mem.prof")
>>> if err != nil {
>>> log.Fatal("could not create memory profile: ", err)
>>> }
>>> runtime.GC() // get up-to-date statistics
>>> if err := pprof.WriteHeapProfile(f); err != nil {
>>> log.Fatal("could not write memory profile: ", err)
>>> }
>>> err = f.Close()
>>> if err != nil {
>>> log.Fatal(err)
>>> }
>>>
>>> When I run the code, this gives me a "mem.prof" file (5084 bytes). But
>>> when I start "go tool pprof" on this file, I get
>>>
>>> >>> go tool pprof xxx mem.prof
>>> File: xxx
>>> Type: inuse_space
>>> Time: Mar 1, 2023 at 1:15pm (GMT)
>>> No samples were found with the default sample value type.
>>> Try "sample_index" command to analyze different sample values.
>>> Entering interactive mode (type "help" for commands, "o" for options)
>>> (pprof) top10
>>> Showing nodes accounting for 0, 0% of 0 total
>>> flat flat% sum%cum cum%
>>>
>>> There seem to be no samples in this file.
>>>
>>> What am I doing wrong?
>>>
>>> All the best,
>>> Jochen
>>>
>>> --
>> You received this message because you are subscribed to the Google Groups
>> "golang-nuts" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to golang-nuts...@googlegroups.com.
>> To view this discussion on the web visit
>> https://groups.google.com/d/msgid/golang-nuts/4e85c396-ef29-41d5-9bb6-5b950648287fn%40googlegroups.com
>>
>> <https://groups.google.com/d/msgid/golang-nuts/4e85c396-ef29-41d5-9bb6-5b950648287fn%40googlegroups.com?utm_medium=email_source=footer>
>> .
>>
>
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/32579fb7-773f-49e3-bfe1-a9af4a292dd9n%40googlegroups.com.
[go-nuts] Re: heap profiling does not work on M2 MacBook air?
The problem also occurs on AMD64 Linux, so it's not architecture specific.
Hints would be most welcome!
On Wednesday, 1 March 2023 at 13:55:30 UTC Jochen Voss wrote:
> Dear all,
>
> I'm trying to profile memory use of a program, following the instructions
> at https://go.dev/blog/pprof , but I can't get memory profiling to work.
> Am I doing things wrong, or is this broken?
>
> Simplified code is at https://go.dev/play/p/Wq_OU49LVQZ . (The code
> doesn't run on the playground, but you can download it and run it locally.)
>
> Following the advice from https://pkg.go.dev/runtime/pprof I added the
> following code to the end of my main() function:
>
> f, err := os.Create("mem.prof")
> if err != nil {
> log.Fatal("could not create memory profile: ", err)
> }
> runtime.GC() // get up-to-date statistics
> if err := pprof.WriteHeapProfile(f); err != nil {
> log.Fatal("could not write memory profile: ", err)
> }
> err = f.Close()
> if err != nil {
> log.Fatal(err)
> }
>
> When I run the code, this gives me a "mem.prof" file (5084 bytes). But
> when I start "go tool pprof" on this file, I get
>
> >>> go tool pprof xxx mem.prof
> File: xxx
> Type: inuse_space
> Time: Mar 1, 2023 at 1:15pm (GMT)
> No samples were found with the default sample value type.
> Try "sample_index" command to analyze different sample values.
> Entering interactive mode (type "help" for commands, "o" for options)
> (pprof) top10
> Showing nodes accounting for 0, 0% of 0 total
> flat flat% sum%cum cum%
>
> There seem to be no samples in this file.
>
> What am I doing wrong?
>
> All the best,
> Jochen
>
>
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/4e85c396-ef29-41d5-9bb6-5b950648287fn%40googlegroups.com.
[go-nuts] heap profiling does not work on M2 MacBook air?
Dear all,
I'm trying to profile memory use of a program, following the instructions
at https://go.dev/blog/pprof , but I can't get memory profiling to work.
Am I doing things wrong, or is this broken?
Simplified code is at https://go.dev/play/p/Wq_OU49LVQZ . (The code
doesn't run on the playground, but you can download it and run it locally.)
Following the advice from https://pkg.go.dev/runtime/pprof I added the
following code to the end of my main() function:
f, err := os.Create("mem.prof")
if err != nil {
log.Fatal("could not create memory profile: ", err)
}
runtime.GC() // get up-to-date statistics
if err := pprof.WriteHeapProfile(f); err != nil {
log.Fatal("could not write memory profile: ", err)
}
err = f.Close()
if err != nil {
log.Fatal(err)
}
When I run the code, this gives me a "mem.prof" file (5084 bytes). But
when I start "go tool pprof" on this file, I get
>>> go tool pprof xxx mem.prof
File: xxx
Type: inuse_space
Time: Mar 1, 2023 at 1:15pm (GMT)
No samples were found with the default sample value type.
Try "sample_index" command to analyze different sample values.
Entering interactive mode (type "help" for commands, "o" for options)
(pprof) top10
Showing nodes accounting for 0, 0% of 0 total
flat flat% sum%cum cum%
There seem to be no samples in this file.
What am I doing wrong?
All the best,
Jochen
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/b6b7456b-16ad-4651-81f9-754184ad52c2n%40googlegroups.com.
Re: [go-nuts] Another generics question
Thanks a lot!
As an aside, the playgound is great. It made this conversation so much
easier!
All the best,
Jochen
On Saturday, 18 February 2023 at 01:31:16 UTC Ian Lance Taylor wrote:
> On Fri, Feb 17, 2023 at 4:47 PM Jochen Voss wrote:,
> >
> > I would like to write a type constraint which matches types like the
> following:
> >
> > type v1 int
> >
> > func (obj v1) Add(other v1) v1 {
> > return obj + other
> > }
> >
> > type v2 string
> >
> > func (obj v2) Add(other v2) v2 {
> > return obj + other
> > }
> >
> > This should match anything with an Add() method which can add another
> item of the same type. I want to be able to write a function like the
> following:
> >
> > func Double[Adder ...???...](x Adder) Adder {
> > return x.Add(x)
> > }
> >
> > My best attempts is https://go.dev/play/p/R0WLqSq9dys , but this
> doesn't quite work. What is the correct type constraint to use?
>
> This is one way to do it: https://go.dev/play/p/uUxkvds5HLB .
>
> Ian
>
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/7c9f4ac8-6951-44cb-b94f-2cee8751a049n%40googlegroups.com.
[go-nuts] Another generics question
Hello,
I would like to write a type constraint which matches types like the
following:
type v1 int
func (obj v1) Add(other v1) v1 {
return obj + other
}
type v2 string
func (obj v2) Add(other v2) v2 {
return obj + other
}
This should match anything with an Add() method which can add another item
of the same type. I want to be able to write a function like the following:
func Double[Adder ...???...](x Adder) Adder {
return x.Add(x)
}
My best attempts is https://go.dev/play/p/R0WLqSq9dys , but this doesn't
quite work. What is the correct type constraint to use?
Many thanks,
Jochen
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/5d8fcc25-2174-4c1d-a325-29111932fe6dn%40googlegroups.com.
Re: [go-nuts] generics: can I require "comparable" together with some methods?
Thanks again, reading the blog post indeed solved the problem!
On Friday, 17 February 2023 at 23:47:09 UTC Ian Lance Taylor wrote:
> On Fri, Feb 17, 2023 at 3:34 PM Jochen Voss wrote:
> >
> > I'm trying to write a generic function like the following (simplified)
> one:
> >
> > type V interface {
> > comparable // (1)
> > SomeMethod(V)
> > }
> >
> > func Test[val V](x val) {
> > m := make(map[val]int) // (2)
> > // uses a x as key in a map and calls x.SomeMethod()
> > }
> >
> > When I try this, see https://go.dev/play/p/cJoH5YwEHaK , I get an error
> message for (1): "cannot use type V outside a type constraint: interface is
> (or embeds) comparable". If I remove the line (1) I get an error for (2)
> instead: "invalid map key type val (missing comparable constraint)".
> >
> > Is there a way to make something like this work?
>
> Are you using Go 1.20? See https://go.dev/blog/comparable .
>
> Ian
>
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/9833388c-f9af-45bd-8f34-22d3edabdae5n%40googlegroups.com.
Re: [go-nuts] generics: can I require "comparable" together with some methods?
Dear Ian,
Yes, I'm using 1.20.1.
I hadn't seen the blog post yet, will read this next. Thanks for pointing
me at this. Judging from the first paragraph of the blog post alone, the
change described there may not solve my problem, though.
All the best,
Jochen
On Friday, 17 February 2023 at 23:47:09 UTC Ian Lance Taylor wrote:
> On Fri, Feb 17, 2023 at 3:34 PM Jochen Voss wrote:
> >
> > I'm trying to write a generic function like the following (simplified)
> one:
> >
> > type V interface {
> > comparable // (1)
> > SomeMethod(V)
> > }
> >
> > func Test[val V](x val) {
> > m := make(map[val]int) // (2)
> > // uses a x as key in a map and calls x.SomeMethod()
> > }
> >
> > When I try this, see https://go.dev/play/p/cJoH5YwEHaK , I get an error
> message for (1): "cannot use type V outside a type constraint: interface is
> (or embeds) comparable". If I remove the line (1) I get an error for (2)
> instead: "invalid map key type val (missing comparable constraint)".
> >
> > Is there a way to make something like this work?
>
> Are you using Go 1.20? See https://go.dev/blog/comparable .
>
> Ian
>
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/5d6fbfe1-be87-4c8f-ae65-3216af73202an%40googlegroups.com.
[go-nuts] generics: can I require "comparable" together with some methods?
Hello,
I'm trying to write a generic function like the following (simplified) one:
type V interface {
comparable // (1)
SomeMethod(V)
}
func Test[val V](x val) {
m := make(map[val]int) // (2)
// uses a x as key in a map and calls x.SomeMethod()
}
When I try this, see https://go.dev/play/p/cJoH5YwEHaK , I get an error
message for (1): "cannot use type V outside a type constraint: interface is
(or embeds) comparable". If I remove the line (1) I get an error for (2)
instead: "invalid map key type val (missing comparable constraint)".
Is there a way to make something like this work?
Many thanks,
Jochen
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/71d60f95-73e4-40cc-833d-dda84ef36708n%40googlegroups.com.
[go-nuts] Re: regexp question
Hi Peter, I don't think this is the solution: https://go.dev/play/p/Ti7UEGd1BVE . I think "All" means that it can deal with several copies of the regexp, rather than returning all matches of the group. All the best, Jochen On Friday, 10 February 2023 at 16:09:04 UTC Peter Galbavy wrote: > FindAllStringSubmatch() ? > https://pkg.go.dev/regexp#Regexp.FindAllStringSubmatch > > On Friday, 10 February 2023 at 12:52:34 UTC Jochen Voss wrote: > >> Dear all, >> >> What happens if a group in a regular expression matches repeatedly, via >> the * operator? Experimentally I found that FindStringSubmatch then >> returns the text of the last match. Is this guaranteed somewhere (I didn't >> find anything about this on https://pkg.go.dev/regexp )? Also, is there >> an easy way to get the other matched strings? >> >> Example: https://go.dev/play/p/h5UDongaz7u >> >> regexp: `A\s*(?:([a-z]+)\s*)*B` >> applied to: "one A two three four B five" >> the inner group gives me "four" >> >> Is FindStringSubmatch guaranteed to return the "four"? >> Is there a way to also get the matched "two" and "three"? >> >> Many thanks, >> Jochen >> >> -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/de25ad20-02bc-4fb6-ab3c-4adc30702c74n%40googlegroups.com.
[go-nuts] regexp question
Dear all, What happens if a group in a regular expression matches repeatedly, via the * operator? Experimentally I found that FindStringSubmatch then returns the text of the last match. Is this guaranteed somewhere (I didn't find anything about this on https://pkg.go.dev/regexp )? Also, is there an easy way to get the other matched strings? Example: https://go.dev/play/p/h5UDongaz7u regexp: `A\s*(?:([a-z]+)\s*)*B` applied to: "one A two three four B five" the inner group gives me "four" Is FindStringSubmatch guaranteed to return the "four"? Is there a way to also get the matched "two" and "three"? Many thanks, Jochen -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/2b8c9ec4-7967-46d3-b6d7-29730eb0f6c3n%40googlegroups.com.
[go-nuts] maps.Keys() does not work when keys are of type language.Tag
Hello,
Using "golang.org/x/text/language" I have a map of
type map[language.Tag]int. I would like to get the keys present in my
map. Why does the following command not work?
import "golang.org/x/exp/maps"
...
var x map[language.Tag]int
...
fmt.Println(maps.Keys(x))
The error message is "Tag does not implement comparable". Code on the
playground: https://go.dev/play/p/dsyEt0ClDBH .
The following function does work as expected, so this is easy to work
around:
func myKeys(m map[language.Tag]int) []language.Tag {
var res []language.Tag
for key := range m {
res = append(res, key)
}
return res
}
But I wonder now whether it is unwise to use language.Tag for the keys in a
map, and why maps.Keys() requires the keys to implement "comparable" in
addition to the constraint "M ~map[K]V".
Many thanks,
Jochen
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/fc3f94ac-e4ae-4964-8c38-ece6492ccd00n%40googlegroups.com.
[go-nuts] how to organise code with channels and error handling?
Dear all,
I am writing a program which processes data in stages. To structure the
code cleanly, I would like implement each stage as an individual function
and use channels to pass data between stages.
My questions:
- How to structure the code? Is my approach described below reasonable?
- How to deal with errors?
Here is a sketch of I currently have (runnable version
at https://go.dev/play/p/7Rrq-OLARl_R ):
func getInputs() <-chan t1 {
c := make(chan t1)
go func() {
for ... {
...
c <- x
}
close(c)
}()
return c
}
func process(in <-chan t1) <-chan t2 {
c := make(chan t2)
go func() {
for i := range in {
x, err := ...
_ = err // what to do with the error?
c <- x
}
close(c)
}()
return c
}
func main() {
c1 := getInputs()
c2 := process(c1)
summarise(c2)
}
There are some things I like about this approach:
- the main function looks nice and tidy
- the stages of work are cleanly separated
But there are also things I don't like:
- I can't see an obvious way to handle errors in the stages. In my case
the first stage is locating input files on the file system, and there may
be read errors etc. The second stage processes the files, and files may
have invalid contents.
- The code for the individual stages looks a bit strange with the double
indentation from the outer func(){} and the go func() {}().
So my questions are:
- What would be a good way to incorporate error handling into my code?
I assume this will require one or more additional channels, just for
errors. Or is there a better way?
- Is there a better way to structure my code?
Many thanks,
Jochen
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/cd4e086c-4f1c-4a48-9162-c46ab45df66fn%40googlegroups.com.
[go-nuts] returning struct vs. returning interface
Dear all, Why does the function NewReader [1] in compress/lzw return an io.ReadCloser instead of an *lzw.Reader? The docs say "It is guaranteed that the underlying type of the returned io.ReadCloser is a *Reader", so why not return a *Reader without wrapping? [1] https://pkg.go.dev/compress/lzw#NewReader Many thanks, Jochen -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/7ee4994a-a9e2-44d7-a939-9ca9fa7529d4n%40googlegroups.com.
[go-nuts] weblist in go tool pprof does not work for me
Hello,
For a while, the "weblist" command in go tool pprof has been broken for me
("list" still works). Is this a known issue? What can I do to fix this?
What I am doing:
- I am using "go version go1.18.3 darwin/arm64" on a MacBook Pro (M1 Pro
processor).
- I am profiling one of my unit tests using the following command:
go test -run '^TestAll$' -cpuprofile cpu.out -memprofile mem.out
- I am calling pprof using the command
go tool pprof sfnt.test cpu.out
- most commands seem to work, for example:
(pprof) top10
Showing nodes accounting for 20130ms, 86.54% of 23260ms total
Dropped 162 nodes (cum <= 116.30ms)
Showing top 10 nodes out of 124
flat flat% sum%cum cum%
6790ms 29.19% 29.19% 6790ms 29.19% runtime.madvise
4840ms 20.81% 50.00% 4840ms 20.81% syscall.syscall6
1910ms 8.21% 58.21% 1910ms 8.21% runtime.usleep
1740ms 7.48% 65.69% 1740ms 7.48% runtime.pthread_kill
1710ms 7.35% 73.04% 1710ms 7.35% runtime.pthread_cond_wait
990ms 4.26% 77.30% 990ms 4.26% runtime.kevent
770ms 3.31% 80.61% 2020ms 8.68%
seehuhn.de/go/pdf/font/cff.decodeCharString
540ms 2.32% 82.93% 1440ms 6.19% runtime.scanobject
530ms 2.28% 85.21% 530ms 2.28% runtime.pthread_cond_signal
310ms 1.33% 86.54% 1040ms 4.47% runtime.mallocgc
- "list" works, for example I get the expected output for list
cff.decodeCharString
- "weblist" does not work. If I call weblist cff.decodeCharString the web
browser opens but shows a page containing only the following five lines of
text, and no listing:
File: sfnt.test
Type: cpu
Time: Jun 24, 2022 at 1:09pm (BST)
Duration: 19.03s, Total samples = 23.26s (122.20%)
Total: 23.26s
Am I doing anything wrong? I didn't find any reports of similar problems
on Google, so maybe this is just some oddity of my setup or me doing
something silly?
All the best,
Jochen
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/02c205c7-e491-4940-9d9d-2421555ff5ecn%40googlegroups.com.
Re: [go-nuts] evaluation of range expression in for-range loop
Thanks Axel,
Now this makes sense to me.
All the best,
Jochen
On Friday, 18 March 2022 at 12:04:04 UTC axel.wa...@googlemail.com wrote:
> https://go.dev/ref/spec#Length_and_capacity says:
>
>> The expression len(s) is constant if s is a string constant. The
>> expressions len(s) and cap(s) are constants if the type of s is an array or
>> pointer to an array and the expression s does not contain channel receives
>> or (non-constant) function calls; in this case s is not evaluated.
>> Otherwise, invocations of len and cap are not constant and s is evaluated.
>
> So, in your case, the array contains a function call, so `len(…)` is not
> constant.
>
> Here is a case where the clause from the spec makes an observable
> difference: https://go.dev/play/p/jeCk6vHzf9u
> Without that clause, the range loop would have to iterate over the element
> of the pointee, causing a nil-dereference panic. The clause allows this
> code to work.
>
> On Fri, Mar 18, 2022 at 12:43 PM Jochen Voss wrote:
>
>> Dear all,
>>
>> The spec at https://go.dev/ref/spec#For_range says
>>
>> "The range expression x is evaluated once before beginning the loop, with
>> one exception: if at most one iteration variable is present and len(x) is
>> constant, the range expression is not evaluated."
>>
>> What does the second half of this sentence mean?
>>
>> My guess was that this says that in "for i := range [3]int{1, 2, f()}
>> ..." the function f is not called, but I quick experiment shows that this
>> guess is wrong, see https://go.dev/play/p/MXqH_C7mllx .
>>
>> All the best,
>> Jochen
>>
>> --
>> You received this message because you are subscribed to the Google Groups
>> "golang-nuts" group.
>> To unsubscribe from this group and stop receiving emails from it, send an
>> email to golang-nuts...@googlegroups.com.
>> To view this discussion on the web visit
>> https://groups.google.com/d/msgid/golang-nuts/db6d865e-233e-4120-8ce5-5feb832732b6n%40googlegroups.com
>>
>> <https://groups.google.com/d/msgid/golang-nuts/db6d865e-233e-4120-8ce5-5feb832732b6n%40googlegroups.com?utm_medium=email_source=footer>
>> .
>>
>
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/589d8a21-931c-4c74-95e8-62ef83a501a9n%40googlegroups.com.
[go-nuts] evaluation of range expression in for-range loop
Dear all,
The spec at https://go.dev/ref/spec#For_range says
"The range expression x is evaluated once before beginning the loop, with
one exception: if at most one iteration variable is present and len(x) is
constant, the range expression is not evaluated."
What does the second half of this sentence mean?
My guess was that this says that in "for i := range [3]int{1, 2, f()} ..."
the function f is not called, but I quick experiment shows that this guess
is wrong, see https://go.dev/play/p/MXqH_C7mllx .
All the best,
Jochen
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/db6d865e-233e-4120-8ce5-5feb832732b6n%40googlegroups.com.
Re: [go-nuts] Pointer to a pointer
Thanks for the pointer to peano.go, this is fun! It took me a while to locate the file. In case others are interested: peano.go is at https://github.com/golang/website/blob/master/_content/doc/play/peano.go and can also be found by choosing "Peano Integers" from the top-right menu in the Go playground. All the best, Jochen On Thursday, 10 March 2022 at 09:31:58 UTC Rob 'Commander' Pike wrote: > On Thu, Mar 10, 2022 at 5:08 PM shan...@gmail.com > wrote: > > > > Is this really how you want to be known? > > Sure, why not? It's a more interesting program than one might think. > > For a richer example of the foundational idea here, see the peano.go > program in the test directory in the repo. > > -rob > -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/efbb4589-ff23-4e57-ae6f-0ab4e82ca1cfn%40googlegroups.com.
Re: [go-nuts] is url->String->Parse != url a bug?
Hi Axel, You write "in general, there are multiple valid escaped forms of any URL". This would still allow for all escaped forms to result in the same URL object when parsed. What I hoped for is a way to go URL->string->URL and get back the same result. I understand that this is how is works for every URL. Now I wonder, if I do url1 := ... s2 := url1.String() url2, _ := url1.Parse(s2) s3 := url2.String() url3,_ := url2.Parse(s3) will I then have reflect.DeepEqual(url2, url3)? If this works, then I could use the url1->url2 step to "normalize" my URL, and url2 would have the property I want. Or is there an easier way to obtain this "normalized" version? All the best, Jochen On Friday, 4 March 2022 at 13:31:50 UTC axel.wa...@googlemail.com wrote: > On Fri, Mar 4, 2022 at 2:12 PM Jochen Voss wrote: > >> If I convert an url.URL to a string and then parse the result, am I meant >> to get back the original url? >> > > I don't think so. String() returns an escaped URL and in general, there > are multiple valid escaped forms of any URL. So, by necessity, there are > different URLs which return the same String() result. > The documentation points that out: > https://pkg.go.dev/net/url#URL.EscapedPath > > >> While playing with the new fuzzer I found that in practise this is not >> the case, >> but I'm not sure whether this is due to bugs, or whether this is expected. >> >> The affected URLs are of course all silly urls, chosen by the fuzzer. >> Example: https://go.dev/play/p/VFe0MbBty3r >> > > For this specific example, there is also this sentence in the docs: > > Trying to parse a hostname and path without a scheme is invalid but may >> not necessarily return an error, due to parsing ambiguities. > > https://pkg.go.dev/net/url#Parse > > I don't know terribly much about the actual URL format, but this sounds to > me, that the parsing result for that string is not actually well-defined > anyway. > > >> All the best, >> Jochen >> >> -- >> You received this message because you are subscribed to the Google Groups >> "golang-nuts" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to golang-nuts...@googlegroups.com. >> To view this discussion on the web visit >> https://groups.google.com/d/msgid/golang-nuts/7a45513b-c0b3-457e-9f3b-3360204a713en%40googlegroups.com >> >> <https://groups.google.com/d/msgid/golang-nuts/7a45513b-c0b3-457e-9f3b-3360204a713en%40googlegroups.com?utm_medium=email_source=footer> >> . >> > -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/288ac10a-2dab-42ad-a134-0e160006d187n%40googlegroups.com.
[go-nuts] is url->String->Parse != url a bug?
Hello, If I convert an url.URL to a string and then parse the result, am I meant to get back the original url? While playing with the new fuzzer I found that in practise this is not the case, but I'm not sure whether this is due to bugs, or whether this is expected. The affected URLs are of course all silly urls, chosen by the fuzzer. Example: https://go.dev/play/p/VFe0MbBty3r All the best, Jochen -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/7a45513b-c0b3-457e-9f3b-3360204a713en%40googlegroups.com.
[go-nuts] feedback from my first experiment with generic go code
Dear all,
Today I tried out the new generics support in Go for the first time by
implementing Dijkstra's algorithm for the shortest path in a directed
graph. In case it's interesting to somebody, here are my impressions, code
is at https://github.com/seehuhn/go-dijkstra .
The key parts of my code are:
type Vertex[edge any] interface {
comparable
Edges() []edge
}
type Edge[vertex any, length constraints.Ordered] interface {
From() vertex
To() vertex
Length() length
}
func ShortestPath[edge Edge[vertex, length], vertex Vertex[edge], length
constraints.Ordered](start, end vertex) ([]edge, error) { ...}
My comments/questions:
- I found generics reasonably easy to use. This was my first experience
with generics, and I went wrong only once. Overall, writing the code was
straightforward.
- Since I am using the beta release, I have to type "go1.18beta2" to
call go. Sometimes I typed "go" accidentally, and I found the resulting
error messages often confusing and unhelpful. For example, "go test" says
"package constraints is not in GOROOT". I wonder whether in case of errors
the go command should show a message if the go version in go.mod is newer
than the current version.
- "go doc" is still sometimes confused about generics. When I type
"go1.18beta2 doc seehuhn.de/go/dijkstra Vertex" the output claims that
Vertex "has unexported methods". The message is misleading, it is cause by
Vertex embedding "comparable". Is this worth reporting as a bug?
- To me it is not intuitive why "comparable" is built intro the language
but "Ordered" is in an external package.
- I have not managed to make it possible for ShortestPath() to
automatically infer the type parameters. Is there a trick I have missed?
- Currently I am using constraints.Ordered for the edge length. This
currently works, but I'm slightly cheating here since I am also using "+"
on the edge lengths. If in the future new ordered types were added to
constraints.Ordered, my code might break. Is there a way to say "I need <=
and +" (short of enumerating all types)?
- When checking for negative edge lengths, I need the zero value for my
instance of constraints.Ordered. Currently I am using a separate,
uninitialised variable for this. Is there a better way to say "x <= 0"
when x is of a type which matches constraints.Ordered?
- I needed to reverse a slice in-place. Would a function to do this
make a good addition to golang.org/x/exp?
- The "func ShortestPath..." line is hard to read and quite long. I
didn't find a good place for a line break either. Any ideas for making
this more readable?
I'd be happy to receive feedback about my code.
All the best,
Jochen
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/ec227a9a-4c8b-4db0-b515-63997af429b1n%40googlegroups.com.
[go-nuts] type parameter names
Dear all,
I stumbled across a difference between type names and type parameter names
(in the go1.18 beta): I can re-use a type name as a variable name, but the
same is not possible with type parameter names. Is this difference
intentional?
For illustration, the following code is
valid https://go.dev/play/p/r2NAddZmoE3 :
type other int
func test(x other) other {
other := x + 1
return other
}
But this doesn't work https://go.dev/play/p/Dsh9IgvKDDY?v=gotip :
type Numeric interface {
~int | ~float64
}
func test[other Numeric](x other) other {
other := x + 1
return other
}
All the best,
Jochen
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/4d0066c3-315d-4746-a0cf-3749bfbcb45dn%40googlegroups.com.
Re: [go-nuts] Re: golang-announce is blocked in Google Groups?
Same here, the problem seems to have gone away. On Thursday, 20 January 2022 at 19:37:54 UTC seank...@gmail.com wrote: > For me it did show as blocked a few hours earlier but works now > >> > -- >> > You received this message because you are subscribed to the Google >> Groups "golang-nuts" group. >> > To unsubscribe from this group and stop receiving emails from it, send >> an email to golang-nuts...@googlegroups.com. >> > To view this discussion on the web visit >> https://groups.google.com/d/msgid/golang-nuts/113efe51-64c0-4a53-9631-8177ca8ed8fcn%40googlegroups.com. >> >> >> > -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/b08cd338-d908-4d92-9823-cf8920b74541n%40googlegroups.com.
[go-nuts] golang-announce is blocked in Google Groups?
Hello, It seems that golang-announce is blocked in Google Groups. When I try to access the group, I only get the following message: Banned content warning golang-announce has been identified as containing spam, malware or other malicious content. For more information about content policies on Google Groups, see ... Anybody knows what this is about? All the best, Jochen -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/110a3919-e8f6-4d43-8e10-007e6dcbbab7n%40googlegroups.com.
[go-nuts] Re: 410 Gone response from sum.golang.org for new module version?
Never mind, waiting fixed it. The package showed up around 40 minutes after I published it on github. All the best, Jochen On Saturday, 2 October 2021 at 11:11:23 UTC+1 Jochen Voss wrote: > Hello, > > I have just published a new version of my websocket library, but I cannot > get it to load into the client. The error message is as follows: > > > go get seehuhn.de/go/webs...@v1.1.0 > <http://seehuhn.de/go/websocket@v1.1.0> > go: downloading seehuhn.de/go/websocket v1.1.0 > go get: seehuhn.de/go/webs...@v1.1.0 > <http://seehuhn.de/go/websocket@v1.1.0>: verifying module: > seehuhn.de/go/webs...@v1.1.0 <http://seehuhn.de/go/websocket@v1.1.0>: > reading https://sum.golang.org/lookup/seehuhn.de/go/webs...@v1.1.0 > <https://sum.golang.org/lookup/seehuhn.de/go/websocket@v1.1.0>: 410 Gone > server response: not found: seehuhn.de/go/webs...@v1.1.0 > <http://seehuhn.de/go/websocket@v1.1.0>: invalid version: unknown > revision v1.1.0 > > The previous version works: go get seehuhn.de/go/webs...@v1.0.1 > <http://seehuhn.de/go/websocket@v1.0.1> and I believe I have correctly > tagged the new version, since the tag shows up here: > https://github.com/seehuhn/go-websocket/tags . > > What am I doing wrong? > > Many thanks, > Jochen > > -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/e7698410-047e-49e5-81e0-6ffa5fc4af8en%40googlegroups.com.
[go-nuts] 410 Gone response from sum.golang.org for new module version?
Hello, I have just published a new version of my websocket library, but I cannot get it to load into the client. The error message is as follows: > go get seehuhn.de/go/websocket@v1.1.0 go: downloading seehuhn.de/go/websocket v1.1.0 go get: seehuhn.de/go/websocket@v1.1.0: verifying module: seehuhn.de/go/websocket@v1.1.0: reading https://sum.golang.org/lookup/seehuhn.de/go/websocket@v1.1.0: 410 Gone server response: not found: seehuhn.de/go/websocket@v1.1.0: invalid version: unknown revision v1.1.0 The previous version works: go get seehuhn.de/go/websocket@v1.0.1 and I believe I have correctly tagged the new version, since the tag shows up here: https://github.com/seehuhn/go-websocket/tags . What am I doing wrong? Many thanks, Jochen -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/fc6d5a31-3197-440e-b9ae-9cc6422ee63en%40googlegroups.com.
Re: [go-nuts] ParseInt confusion
Oh, yes, thank you and sorry about the noise!
On Tuesday, 9 February 2021 at 17:02:41 UTC ja...@kastelo.net wrote:
> An int8 has a range of -128 to +127. You may be looking for
> strconv.ParseUint if you want an 8 bit unsigned int?
>
> //jb
>
> On 9 Feb 2021, at 17:51, Jochen Voss wrote:
>
> Dear all,
>
> The command strconv.ParseInt("256", 8, 8) should not return an error,
> because octal 256 is decimal 174 which fits into 8 bits. Right?
>
> I'm asking, because actually running this command gives 'strconv.ParseInt:
> parsing "256": value out of range'! https://play.golang.org/p/gNKN0kKLfDd
>
> What am I overlooking here?
>
> All the best,
> Jochen
>
> --
> You received this message because you are subscribed to the Google Groups
> "golang-nuts" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to golang-nuts...@googlegroups.com.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/golang-nuts/73e9d2a5-d8fb-468d-aebe-95f171bbff1dn%40googlegroups.com
>
> <https://groups.google.com/d/msgid/golang-nuts/73e9d2a5-d8fb-468d-aebe-95f171bbff1dn%40googlegroups.com?utm_medium=email_source=footer>
> .
>
>
>
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/d1dcd8dd-a31f-48e8-993d-005f05e24d91n%40googlegroups.com.
[go-nuts] ParseInt confusion
Dear all,
The command strconv.ParseInt("256", 8, 8) should not return an error,
because octal 256 is decimal 174 which fits into 8 bits. Right?
I'm asking, because actually running this command gives 'strconv.ParseInt:
parsing "256": value out of range'! https://play.golang.org/p/gNKN0kKLfDd
What am I overlooking here?
All the best,
Jochen
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/73e9d2a5-d8fb-468d-aebe-95f171bbff1dn%40googlegroups.com.
Re: [go-nuts] printing an empty slice with fmt.Printf()
Dear all,
Thanks for your help. The solution is indeed to use "%x" instead of "%02x".
My confusion was caused by the following sentence in the docs: "For
compound operands such as slices and structs, the format applies to the
elements of each operand". I never spotted the relevant exception
mentioned a few lines later: "However, when printing a byte slice with a
string-like verb (%s %q %x %X), it is treated ... as a single item." So
the "%02" applies to the whole byte slice, and not to each byte as I had
wrongly thought.
All the best,
Jochen
On Saturday, 19 December 2020 at 00:24:07 UTC kortschak wrote:
> I think that's the question. Here's a simpler example,
> https://play.golang.org/p/9Kv3PhlM-OF
>
> That is, is 00 an expected %02x representation of a zero-length byte
> slice?
>
> The answer to that is yes; the 02 forces leading zeros. The %x verb
> essentially renders bit strings as hex, so a zero-length bit string
> with mandated two leading zeros is 00.
>
> If you leave out the 02, you get the expected empty string.
>
> https://play.golang.org/p/uVFt3lecKxf
>
> On Fri, 2020-12-18 at 15:38 -0800, Marcin Romaszewicz wrote:
> > It's expected behavior.
> >
> > Your for loop runs once for l=0, since your condition is <=0 because
> > len([]byte{}) is 0.
> >
> > -- Marcin
>
>
>
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/bd781856-6e04-4978-b23f-8b7d98611e56n%40googlegroups.com.
[go-nuts] printing an empty slice with fmt.Printf()
Hello,
I can print slices of bytes as hex strings, using code like the following:
x := []byte{0, 1, 2, 3}
fmt.Printf("%02x", x[:l])
This gives the output "00010203" as expected. But this fails for the empty
slice: running
x := []byte{}
fmt.Printf("%02x", x[:l])
gives "00" instead of the empty string. See
https://play.golang.org/p/rvLLqydDDE6 for a demonstration.
Is this a bug, or is this expected behaviour?
All the best,
Jochen
--
You received this message because you are subscribed to the Google Groups
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to golang-nuts+unsubscr...@googlegroups.com.
To view this discussion on the web visit
https://groups.google.com/d/msgid/golang-nuts/bb8774f3-aa17-4044-8435-ed1cd162976cn%40googlegroups.com.
[go-nuts] Re: request for feedback: another websocket server implementation
Dear Ilia, Thank you for your comment. In my mind, Read and Write are made for streaming data, whereas a websocket connection transports messages. How would I handle message types and message boundaries when using the io.Reader and io.Writer interfaces? (I do try to implement these interfaces for the *contents* of the messages themselves, see my SendMessage and ReceiveMessage functions.) Many thanks, Jochen On Wednesday, 18 September 2019 14:37:49 UTC+1, Ilia Choly wrote: > > You should try to make Conn implement io.Reader and io.Writer > > On Tuesday, September 17, 2019 at 6:29:08 PM UTC-4, Jochen Voss wrote: >> >> Dear all, >> >> Because I wanted to learn about the websocket protocol, I have >> implemented a websocket server in Go: >> >> code: https://github.com/seehuhn/go-websocket >> documentation: https://godoc.org/seehuhn.de/go/websocket >> >> I would be happy to receive feedback about my code. >> >> Specific questions: >> >> - Does the API look reasonable? Anything to improve? >> >> - I use the function readMultiplexer (at the end of reader.go) to read >> all frame headers from the network connection, and to make sure that the >> readers of the frame bodies don't step on each other's feet. Similarly, >> the function writeMultiplexer (at the end of writer.go) coordinates all >> writers and performs all writes to the network connection. This system is >> held together by a tangle of go channels. The code is much messier than I >> would like it to be. Is there a way to structure such code so that it >> is easier to reason about, so that I can be more confident that it is >> correct? The bits a struggled most with are the various ways the >> connection can be closed (closed by server explicitly or because of a >> protocol error, closed by client with/without close frame). >> >> - I attempted to profile the code, to find whether there are any >> performance bottlenecks. My code for this is in websocket_test.go >> (function BenchmarkEcho), but the profiles then only ever show the code >> waiting for data on go channels or for data from the network. My feeling >> is that I was really benchmarking the network connection and not my code. >> Is there a way to improve BenchmarkEcho to get more meaningful results? >> >> Many thanks, >> Jochen >> >> -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/de7cd4f5-e2dd-4a17-876d-c4ac61da2c3f%40googlegroups.com.
[go-nuts] Re: request for feedback: another websocket server implementation
Dear T L, Thanks for this. I was aware of implementations number 1 and 4 (and I think I like my own better than these), but I didn't spot numbers 2 and 3. I'll have a look at these, too. Many thanks, Jochen On Wednesday, 18 September 2019 14:27:17 UTC+1, T L wrote: > > Just for reference. Here are other Go websocket implementations: > * Gorilla WebSocket https://github.com/gorilla/websocket > * https://github.com/nhooyr/websocket > * https://github.com/gobwas/ws > * x/net/websocket https://godoc.org/golang.org/x/net/websocket > > On Tuesday, September 17, 2019 at 6:29:08 PM UTC-4, Jochen Voss wrote: >> >> Dear all, >> >> Because I wanted to learn about the websocket protocol, I have >> implemented a websocket server in Go: >> >> code: https://github.com/seehuhn/go-websocket >> documentation: https://godoc.org/seehuhn.de/go/websocket >> >> I would be happy to receive feedback about my code. >> >> Specific questions: >> >> - Does the API look reasonable? Anything to improve? >> >> - I use the function readMultiplexer (at the end of reader.go) to read >> all frame headers from the network connection, and to make sure that the >> readers of the frame bodies don't step on each other's feet. Similarly, >> the function writeMultiplexer (at the end of writer.go) coordinates all >> writers and performs all writes to the network connection. This system is >> held together by a tangle of go channels. The code is much messier than I >> would like it to be. Is there a way to structure such code so that it >> is easier to reason about, so that I can be more confident that it is >> correct? The bits a struggled most with are the various ways the >> connection can be closed (closed by server explicitly or because of a >> protocol error, closed by client with/without close frame). >> >> - I attempted to profile the code, to find whether there are any >> performance bottlenecks. My code for this is in websocket_test.go >> (function BenchmarkEcho), but the profiles then only ever show the code >> waiting for data on go channels or for data from the network. My feeling >> is that I was really benchmarking the network connection and not my code. >> Is there a way to improve BenchmarkEcho to get more meaningful results? >> >> Many thanks, >> Jochen >> >> -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/9637cdf0-c438-4659-9ce8-35ddff219bfb%40googlegroups.com.
[go-nuts] request for feedback: another websocket server implementation
Dear all, Because I wanted to learn about the websocket protocol, I have implemented a websocket server in Go: code: https://github.com/seehuhn/go-websocket documentation: https://godoc.org/seehuhn.de/go/websocket I would be happy to receive feedback about my code. Specific questions: - Does the API look reasonable? Anything to improve? - I use the function readMultiplexer (at the end of reader.go) to read all frame headers from the network connection, and to make sure that the readers of the frame bodies don't step on each other's feet. Similarly, the function writeMultiplexer (at the end of writer.go) coordinates all writers and performs all writes to the network connection. This system is held together by a tangle of go channels. The code is much messier than I would like it to be. Is there a way to structure such code so that it is easier to reason about, so that I can be more confident that it is correct? The bits a struggled most with are the various ways the connection can be closed (closed by server explicitly or because of a protocol error, closed by client with/without close frame). - I attempted to profile the code, to find whether there are any performance bottlenecks. My code for this is in websocket_test.go (function BenchmarkEcho), but the profiles then only ever show the code waiting for data on go channels or for data from the network. My feeling is that I was really benchmarking the network connection and not my code. Is there a way to improve BenchmarkEcho to get more meaningful results? Many thanks, Jochen -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/1826e9b6-03db-4d93-92e2-ba0c0fd0b379%40googlegroups.com.
Re: [go-nuts] panic: interface conversion: interface is nil, not encoding.BinaryUnmarshaler
Hi Axel, Thanks a lot for looking into this. Your post makes things much clearer for me. In particular, I now think that I probably should not have (Un)MarshalBinary() methods on the interface type, but have them only on the concrete type instead. For example, this works: https://play.golang.org/p/oBgHqN74ZeM . On Thursday, 15 August 2019 11:16:45 UTC+1, Axel Wagner wrote: > > The panic, I think, comes from the fact that you pass a *Duck, which then > gets dereferenced, pointing at a nil Duck. Even a nil-Duck *does* have a > DecodeBinary method though, so gob thinks it should use that. It then > type-asserts to BinaryDecoder, which panics, because it's a nil-interface, > just like this example: https://play.golang.org/p/EEwOt0FQunh. > I had forgotten that type assertions don't allow nil values, but indeed they don't. Thank you for reminding me. Many thanks, Jochen -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/a707a33a-ea3a-4103-84f8-24d56a91fdae%40googlegroups.com.
Re: [go-nuts] panic: interface conversion: interface is nil, not encoding.BinaryUnmarshaler
Hi Axel, On Thursday, 15 August 2019 09:13:17 UTC+1, Axel Wagner wrote: > > I haven't really used gob much, so unfortunately I can't *really* help > you. But one thing to note is that you can dump the content of the buffer > and see that it doesn't actually contain the name of the type you are > encoding: https://play.golang.org/p/R8HB6RP8kS0 > You are right. But the type is somehow represented by some index or so, instead of the name? Since the type needs to be registered using gob.Register() beforehand, this seems not totally impossible to me. But then, when I change your code to let gob do the encoding itself, the name is suddenly there: https://play.golang.org/p/Qz7zDKL047z, so probably the absence of the type name is part of the problem. > And I agree that this probably shouldn't panic, but instead return an > error. But given that the encoded bytes don't contain the name of the > concrete struct, there is no way the decoder can know where you want to put > it, without giving it the correct type. > Yes, an informative error instead of the hard to interpret panic would be nice. I wonder whether this is indeed intended behaviour, or whether this is some kind of bug in the encoding/gob package. -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/32c27653-c50b-42cd-a525-8c9e9298bf60%40googlegroups.com.
Re: [go-nuts] panic: interface conversion: interface is nil, not encoding.BinaryUnmarshaler
Dear Marcin, My aim is to unmarshal into an interface variable, without having to know in advance which concrete type I'm receiving (the actual interface has several possible implementations, and gob data comes in over a network connection). So, while your code avoids the panic, it does not solve my problem. Many thanks, Jochen On Wednesday, 14 August 2019 22:12:54 UTC+1, Marcin Romaszewicz wrote: > > Here you go: > https://play.golang.org/p/_cPmaSxRatC > > You want to unmarshal into , not into > > This means: > var duck2 A > > not: > var duck2 Duck > > On Wed, Aug 14, 2019 at 8:46 AM Jochen Voss > wrote: > >> Hello, >> >> I'm trying to read gob-encoded data into a variable of interface type. >> In simple cases, this works for me, but when I use a custom encoding >> via MarshalBinary() and UnmarshalBinary() methods, my code keeps crashing, >> with the error message >> >> panic: interface conversion: interface is nil, not >> encoding.BinaryUnmarshaler >> >> Example: >> >> - The code at https://play.golang.org/p/y8nWNhObUwb, letting gob do the >> en-/decoding works. >> >> - The very similar code at https://play.golang.org/p/-zS7QjEJg9x, which >> uses MarshalBinary() and UnmarshalBinary() crashes with the panic shown >> above. >> >> What am I doing wrong? >> >> [I asked the same question at >> https://stackoverflow.com/questions/57485698/, with no answers so far] >> >> Many thanks, >> Jochen >> >> -- >> You received this message because you are subscribed to the Google Groups >> "golang-nuts" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to golan...@googlegroups.com . >> To view this discussion on the web visit >> https://groups.google.com/d/msgid/golang-nuts/b19c51bc-524c-4292-915b-fc00e9289052%40googlegroups.com >> >> <https://groups.google.com/d/msgid/golang-nuts/b19c51bc-524c-4292-915b-fc00e9289052%40googlegroups.com?utm_medium=email_source=footer> >> . >> > -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/042b28c0-d768-4cba-94cf-159cf4231818%40googlegroups.com.
[go-nuts] panic: interface conversion: interface is nil, not encoding.BinaryUnmarshaler
Hello, I'm trying to read gob-encoded data into a variable of interface type. In simple cases, this works for me, but when I use a custom encoding via MarshalBinary() and UnmarshalBinary() methods, my code keeps crashing, with the error message panic: interface conversion: interface is nil, not encoding.BinaryUnmarshaler Example: - The code at https://play.golang.org/p/y8nWNhObUwb, letting gob do the en-/decoding works. - The very similar code at https://play.golang.org/p/-zS7QjEJg9x, which uses MarshalBinary() and UnmarshalBinary() crashes with the panic shown above. What am I doing wrong? [I asked the same question at https://stackoverflow.com/questions/57485698/, with no answers so far] Many thanks, Jochen -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/golang-nuts/b19c51bc-524c-4292-915b-fc00e9289052%40googlegroups.com.
[go-nuts] Re: HTML template behaviour changed in Go 1.8?
Since this is a change in behaviour since Go 1.7.5, I have now reported this as a bug: https://github.com/golang/go/issues/19294 -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
[go-nuts] HTML template behaviour changed in Go 1.8?
Hello, I am using Go templates to generated the files inside an EPUB container. Since I switched to Go version 1.8 the output of my rendered templates seems to be no longer deterministic and often bit which used to be always there are now missing. Minimal working (well, broken) example: https://play.golang.org/p/BnwKHNrnom Full, messy code: https://github.com/seehuhn/epublatex When I run the code from the "minimal working example", on my machine the output differs between runs: voss@flammeri [..uhn/epublatex] go run bug.go http://www.w3.org/1999/xhtml;> voss@flammeri [..uhn/epublatex] go run bug.go http://www.w3.org/1999/xhtml;> The second run has the stylesheet line included, while it is missing in the first run. On play.golang.org the code seems deterministic, but I suspect that may be due to caching? Am I misusing Go templates here? If so, what should I be doing differently? Or is there a problem with templates in Go version 1.8? Many thanks, Jochen -- You received this message because you are subscribed to the Google Groups "golang-nuts" group. To unsubscribe from this group and stop receiving emails from it, send an email to golang-nuts+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
