digitalmars.D.learn - Is there something special required to use Appender.clear
- Ellie Harper (36/36) Mar 27 2018 Sorry if this is a stupid question, but is there something
- bauss (3/41) Mar 27 2018 I would say that's a bug, because it seems like it's trying to
- Anonymouse (6/38) Mar 27 2018 Appender only implements clear for mutable types, and string is
- Simen =?UTF-8?B?S2rDpnLDpXM=?= (6/13) Mar 27 2018 Since string is immutable(char)[], clear() is simply not
- FeepingCreature (6/19) Jan 24 2019 Isn't this wrong, though? Appender controls the memory it
- =?UTF-8?Q?Ali_=c3=87ehreli?= (5/26) Jan 24 2019 I think Appender is trying to protect previous data from Appender's
- FeepingCreature (13/53) Jan 25 2019 Aren't the semantics of .clear that it's invalid to access
- Steven Schveighoffer (11/67) Jan 25 2019 Yes, because Appender doesn't know where that memory comes from. A
- FeepingCreature (25/53) Jan 28 2019 That's stupid. Why would you advise me not to, if there's no
- Steven Schveighoffer (27/64) Jan 28 2019 Just like this:
- FeepingCreature (33/60) Jan 30 2019 Fair enough.
- Steven Schveighoffer (34/80) Jan 30 2019 Wow, I take this back. I tried it, and it DOES reallocate.
- Steven Schveighoffer (22/42) Jan 24 2019 Appender is sometimes given a starting array. clear isn't callable in
Sorry if this is a stupid question, but is there something
special required to call Appender.clear? When I attempt even
just a simple use I am getting compile errors relating to
`template object.clear`.
When I try:
import std.array;
void main(string[] args){
auto foo = appender!string;
foo.clear;
}
I receive the following output:
$ ldc2 source/tmp.d
source/tmp.d(5): Error: template object.clear cannot deduce
function from argument types !()(Appender!string), candidates are:
/usr/local/include/d/ldc/object.d(2041):
object.clear(T : Value[Key], Value, Key)(T aa)
/usr/local/include/d/ldc/object.d(2046):
object.clear(T : Value[Key], Value, Key)(T* aa)
$ dmd source/tmp.d
source/tmp.d(5): Error: template object.clear cannot deduce
function from argument types !()(Appender!string), candidates are:
/usr/include/dmd/druntime/import/object.d(1983):
object.clear(T : Value[Key], Value, Key)(T aa)
/usr/include/dmd/druntime/import/object.d(1988):
object.clear(T : Value[Key], Value, Key)(T* aa)
versions:
$ dmd --version
DMD64 D Compiler v2.078.0
Copyright (c) 1999-2017 by The D Language Foundation written
by Walter Bright
$ ldc2 --version
LDC - the LLVM D compiler (1.8.0git-921bb7f):
based on DMD v2.078.3 and LLVM 4.0.1
built with DMD64 D Compiler v2.078.0
Is there something I am missing here?
Thanks
Mar 27 2018
bauss <jj_1337 live.dk>
On Tuesday, 27 March 2018 at 12:17:58 UTC, Ellie Harper wrote:
Sorry if this is a stupid question, but is there something
special required to call Appender.clear? When I attempt even
just a simple use I am getting compile errors relating to
`template object.clear`.
When I try:
import std.array;
void main(string[] args){
auto foo = appender!string;
foo.clear;
}
I receive the following output:
$ ldc2 source/tmp.d
source/tmp.d(5): Error: template object.clear cannot deduce
function from argument types !()(Appender!string), candidates
are:
/usr/local/include/d/ldc/object.d(2041):
object.clear(T : Value[Key], Value, Key)(T aa)
/usr/local/include/d/ldc/object.d(2046):
object.clear(T : Value[Key], Value, Key)(T* aa)
$ dmd source/tmp.d
source/tmp.d(5): Error: template object.clear cannot deduce
function from argument types !()(Appender!string), candidates
are:
/usr/include/dmd/druntime/import/object.d(1983):
object.clear(T : Value[Key], Value, Key)(T aa)
/usr/include/dmd/druntime/import/object.d(1988):
object.clear(T : Value[Key], Value, Key)(T* aa)
versions:
$ dmd --version
DMD64 D Compiler v2.078.0
Copyright (c) 1999-2017 by The D Language Foundation
written by Walter Bright
$ ldc2 --version
LDC - the LLVM D compiler (1.8.0git-921bb7f):
based on DMD v2.078.3 and LLVM 4.0.1
built with DMD64 D Compiler v2.078.0
Is there something I am missing here?
Thanks
I would say that's a bug, because it seems like it's trying to
get the clear() function used for associative arrays.
Mar 27 2018
On Tuesday, 27 March 2018 at 12:17:58 UTC, Ellie Harper wrote:Sorry if this is a stupid question, but is there something special required to call Appender.clear? When I attempt even just a simple use I am getting compile errors relating to `template object.clear`. When I try: import std.array; void main(string[] args){ auto foo = appender!string; foo.clear; }Appender only implements clear for mutable types, and string is an array of immutables. I'm not sure but it looks like you're hitting the associative array clear by UFCS instead? Make it Appender!(char[]) and it will work.// only allow overwriting data on non-immutable and non-const data static if (isMutable!T) { /** * Clears the managed array. This allows the elements of the array to be reused * for appending. * * Note: clear is disabled for immutable or const element types, due to the * possibility that $(D Appender) might overwrite immutable data. */ void clear() trusted pure nothrow { if (_data) { _data.arr = _data.arr.ptr[0 .. 0]; } } [...]https://github.com/dlang/phobos/blob/master/std/array.d#L3140
Mar 27 2018
On Tuesday, 27 March 2018 at 12:17:58 UTC, Ellie Harper wrote:Sorry if this is a stupid question, but is there something special required to call Appender.clear? When I attempt even just a simple use I am getting compile errors relating to `template object.clear`.From the documentation for Appender:Note clear is disabled for immutable or const element types, due to the possibility that Appender might overwrite immutable data.Since string is immutable(char)[], clear() is simply not available for appender!string. -- Simen
Mar 27 2018
On Tuesday, 27 March 2018 at 12:31:05 UTC, Simen Kjærås wrote:On Tuesday, 27 March 2018 at 12:17:58 UTC, Ellie Harper wrote:Isn't this wrong, though? Appender controls the memory it references. It could just choose to allocate non-immutable memory internally. As long as any const data put into the appender is *returned* as const, there is no chance of immutable memory being overwritten.Sorry if this is a stupid question, but is there something special required to call Appender.clear? When I attempt even just a simple use I am getting compile errors relating to `template object.clear`.From the documentation for Appender:Note clear is disabled for immutable or const element types, due to the possibility that Appender might overwrite immutable data.Since string is immutable(char)[], clear() is simply not available for appender!string. -- Simen
Jan 24 2019
On 01/24/2019 04:35 AM, FeepingCreature wrote:On Tuesday, 27 March 2018 at 12:31:05 UTC, Simen Kjærås wrote:I think Appender is trying to protect previous data from Appender's later use. If it handed out immutable data, the user is expecting it to not change. So, Appender cannot clear it for later use. AliOn Tuesday, 27 March 2018 at 12:17:58 UTC, Ellie Harper wrote:Isn't this wrong, though? Appender controls the memory it references. It could just choose to allocate non-immutable memory internally. As long as any const data put into the appender is *returned* as const, there is no chance of immutable memory being overwritten.Sorry if this is a stupid question, but is there something special required to call Appender.clear? When I attempt even just a simple use I am getting compile errors relating to `template object.clear`.From the documentation for Appender:Note clear is disabled for immutable or const element types, due to the possibility that Appender might overwrite immutable data.Since string is immutable(char)[], clear() is simply not available for appender!string. -- Simen
Jan 24 2019
On Thursday, 24 January 2019 at 17:49:34 UTC, Ali Çehreli wrote:On 01/24/2019 04:35 AM, FeepingCreature wrote:Aren't the semantics of .clear that it's invalid to access references to .data after calling .clear, period? And if not, then shouldn't they be? Consider if Appender managed its own memory and held on to previously-allocated arrays while you were appending, only to free them on .clear. That seems like something Appender should be allowed to do. If you don't want it, just reinitialize Appender instead of calling .clear.On Tuesday, 27 March 2018 at 12:31:05 UTC, Simen Kjærås wrote:wrote:On Tuesday, 27 March 2018 at 12:17:58 UTC, Ellie HarperspecialSorry if this is a stupid question, but is there somethinga simplerequired to call Appender.clear? When I attempt even justobject.clear`.use I am getting compile errors relating to `templateto theFrom the documentation for Appender:Note clear is disabled for immutable or const element types, dueavailable forpossibility that Appender might overwrite immutable data.Since string is immutable(char)[], clear() is simply notreferences. Itappender!string. -- SimenIsn't this wrong, though? Appender controls the memory itcould just choose to allocate non-immutable memoryinternally. As longas any const data put into the appender is *returned* asconst, there isno chance of immutable memory being overwritten.I think Appender is trying to protect previous data from Appender's later use. If it handed out immutable data, the user is expecting it to not change. So, Appender cannot clear it for later use. AliAppender is sometimes given a starting array. clear isn't callable in that case, and we don't distinguish the difference in the type or at runtime.Any reason that the semantics of .clear should be different for a starting array? Anyway if so, I'd prefer to just make that a runtime error. I'm mostly fishing around if anyone has an objection to a PR to change this.
Jan 25 2019
On 1/25/19 3:20 AM, FeepingCreature wrote:On Thursday, 24 January 2019 at 17:49:34 UTC, Ali Çehreli wrote:You are advised not to. But it's not unsafe, as the memory is still there.On 01/24/2019 04:35 AM, FeepingCreature wrote:Aren't the semantics of .clear that it's invalid to access references to .data after calling .clear, period? And if not, then shouldn't they be? Consider if Appender managed its own memory and held on to previously-allocated arrays while you were appending, only to free them on .clear. That seems like something Appender should be allowed to do. If you don't want it, just reinitialize Appender instead of calling .clear.On Tuesday, 27 March 2018 at 12:31:05 UTC, Simen Kjærås wrote:wrote:On Tuesday, 27 March 2018 at 12:17:58 UTC, Ellie HarperspecialSorry if this is a stupid question, but is there somethinga simplerequired to call Appender.clear? When I attempt even justobject.clear`.use I am getting compile errors relating to `templateto theFrom the documentation for Appender:Note clear is disabled for immutable or const element types, dueavailable forpossibility that Appender might overwrite immutable data.Since string is immutable(char)[], clear() is simply notreferences. Itappender!string. -- SimenIsn't this wrong, though? Appender controls the memory itcould just choose to allocate non-immutable memoryinternally. As longas any const data put into the appender is *returned* asconst, there isno chance of immutable memory being overwritten.I think Appender is trying to protect previous data from Appender's later use. If it handed out immutable data, the user is expecting it to not change. So, Appender cannot clear it for later use. AliYes, because Appender doesn't know where that memory comes from. A string could be, for instance, passed to another thread and being used, you wouldn't want to overwrite it. Generally speaking, overwriting immutable data is UB in D anyway. Making it a runtime error would be possible, but there has to be a good reason to make it that way.Appender is sometimes given a starting array. clear isn't callable in that case, and we don't distinguish the difference in the type or at runtime.Any reason that the semantics of .clear should be different for a starting array? Anyway if so, I'd prefer to just make that a runtime error.I'm mostly fishing around if anyone has an objection to a PR to change this.Without good reasons to change, I don't see why it would be accepted. Maybe you can describe your use case? -Steve
Jan 25 2019
On Friday, 25 January 2019 at 14:33:16 UTC, Steven Schveighoffer wrote:On 1/25/19 3:20 AM, FeepingCreature wrote:That's stupid. Why would you advise me not to, if there's no problem with it? Either it should be accepted or it should be forbidden, just like warnings.On Thursday, 24 January 2019 at 17:49:34 UTC, Ali Çehreli wrote: Aren't the semantics of .clear that it's invalid to access references to .data after calling .clear, period? And if not, then shouldn't they be? Consider if Appender managed its own memory and held on to previously-allocated arrays while you were appending, only to free them on .clear. That seems like something Appender should be allowed to do. If you don't want it, just reinitialize Appender instead of calling .clear.You are advised not to. But it's not unsafe, as the memory is still there.In which case, you can't reuse the Appender and should not call `clear`! It seems like we're gimping functionality of a very basic tool for memory efficiency for the sake of not confusing users. This rarely pays off.Any reason that the semantics of .clear should be different for a starting array? Anyway if so, I'd prefer to just make that a runtime error.Yes, because Appender doesn't know where that memory comes from. A string could be, for instance, passed to another thread and being used, you wouldn't want to overwrite it.Generally speaking, overwriting immutable data is UB in D anyway. Making it a runtime error would be possible, but there has to be a good reason to make it that way.My use case is simply the standard usecase of Appender: I want to build up an array in a way that reduces GC churn. Maybe it's an array of structs that contain const members that I'll serialize to JSON and send on a socket. In that case, I know for a fact that no references will hang around past the serialization. That's what clear _is for._ I don't see why this would be different with const or immutable data; if you hold references past .clear being called you're in trouble *anyways.* I consider initializing Appender with an array referencing immutable data a borderline error anyways. The entire point of Appender is that it caches capacity data of GC managed memory, which is never immutable. On the first append to an immutable-memory array, it has to reallocate *anyways*. There is no benefit to initializing an Appender with immutable memory over just appending it first thing, unless you never plan to append to it ever.I'm mostly fishing around if anyone has an objection to a PR to change this.Without good reasons to change, I don't see why it would be accepted. Maybe you can describe your use case?
Jan 28 2019
On 1/28/19 7:39 AM, FeepingCreature wrote:On Friday, 25 January 2019 at 14:33:16 UTC, Steven Schveighoffer wrote:Just like this: int[] arr = [1,2,3,4,5]; auto slice = arr[0 .. 2]; slice.assumeSafeAppend; Now, you can use arr still, it's got elements in it beyond what slice has. You might have things change underneath you without expecting it. That's why it's advised not to do that. But if arr is immutable(int)[], then you are running into undefined behavior, it's a completely different problem.On 1/25/19 3:20 AM, FeepingCreature wrote:That's stupid. Why would you advise me not to, if there's no problem with it? Either it should be accepted or it should be forbidden, just like warnings.On Thursday, 24 January 2019 at 17:49:34 UTC, Ali Çehreli wrote: Aren't the semantics of .clear that it's invalid to access references to .data after calling .clear, period? And if not, then shouldn't they be? Consider if Appender managed its own memory and held on to previously-allocated arrays while you were appending, only to free them on .clear. That seems like something Appender should be allowed to do. If you don't want it, just reinitialize Appender instead of calling .clear.You are advised not to. But it's not unsafe, as the memory is still there.Right, this does seem like a big limitation. Keeping with the spirit of how slices don't own the memory in question, Appender is being conservative with what it doesn't know. I wonder if it may be more appropriate to instead of preventing clear() on immutable/const arrays, to make it system. Or maybe call it something different "dangerousClear" or something ;) There definitely should be some way to fail if clear is called on an array that was passed into the constructor. But I'm still not sure we should allow overwriting immutable memory without a cast, even in system code.Without good reasons to change, I don't see why it would be accepted. Maybe you can describe your use case?My use case is simply the standard usecase of Appender: I want to build up an array in a way that reduces GC churn. Maybe it's an array of structs that contain const members that I'll serialize to JSON and send on a socket. In that case, I know for a fact that no references will hang around past the serialization. That's what clear _is for._ I don't see why this would be different with const or immutable data; if you hold references past .clear being called you're in trouble *anyways.*I consider initializing Appender with an array referencing immutable data a borderline error anyways. The entire point of Appender is that it caches capacity data of GC managed memory, which is never immutable. On the first append to an immutable-memory array, it has to reallocate *anyways*. There is no benefit to initializing an Appender with immutable memory over just appending it first thing, unless you never plan to append to it ever.It will inspect the allocated length from the GC if the array is appendable from the beginning. So it's not always going to reallocate. e.g.: string x = "abc".idup; auto app = x.appender; app ~= "xyz"; // does not reallocate. -Steve
Jan 28 2019
On Monday, 28 January 2019 at 15:16:54 UTC, Steven Schveighoffer wrote:It will inspect the allocated length from the GC if the array is appendable from the beginning. So it's not always going to reallocate. e.g.: string x = "abc".idup; auto app = x.appender; app ~= "xyz"; // does not reallocate. -SteveFair enough.My problem is this. const and immutable are *not* well supported in the standard library at the moment. What I want is that I can use the idioms of the stdlib, semantically, in a way that lets me *not care* about const or immutable, that lets me express the patterns I want without having to worry about whether some type deep inside my code, a dub library, or phobos itself decided to declare a field immutable. Appender covers two usecases: "capacity caching" and "memory reuse." Both of those usecases have two const variations: "possibly immutable" and "mutable". mutable capacity caching | mutable memory reuse ---------------------------+--------------------- immutable capacity caching | immutable memory reuse But instead of cutting between capacity caching and memory reuse, down the middle, Appender cuts between mutable and immutable, left to right. I think this is a symptom of a broad negligence of constness as a first-class property - constness works sometimes, and it's nice if it does, but it can't be *relied* on to be well supported. This makes using const in D an eternal uphill struggle. Why even go to all the trouble to make a new major version of the language just to introduce constness, if it's not going to be treated as a first-class concern? I don't want to have to sprinkle static if(isAssignable!T) everytime I want to use a library type. If Phobos offers me as generic a data structure as Appender, parameterized on T, it should work for T, regardless of what T is doing, and *certainly* regardless of what fields in T are marked const. Especially considering that if a field is not marked as const, that hardly means it's not going to lead to bugs if its memory is reused while you're referencing it!My use case is simply the standard usecase of Appender: I want to build up an array in a way that reduces GC churn. Maybe it's an array of structs that contain const members that I'll serialize to JSON and send on a socket. In that case, I know for a fact that no references will hang around past the serialization. That's what clear _is for._ I don't see why this would be different with const or immutable data; if you hold references past .clear being called you're in trouble *anyways.*Right, this does seem like a big limitation. Keeping with the spirit of how slices don't own the memory in question, Appender is being conservative with what it doesn't know. I wonder if it may be more appropriate to instead of preventing clear() on immutable/const arrays, to make it system. Or maybe call it something different "dangerousClear" or something ;) There definitely should be some way to fail if clear is called on an array that was passed into the constructor. But I'm still not sure we should allow overwriting immutable memory without a cast, even in system code.
Jan 30 2019
On 1/30/19 7:32 AM, FeepingCreature wrote:On Monday, 28 January 2019 at 15:16:54 UTC, Steven Schveighoffer wrote:Wow, I take this back. I tried it, and it DOES reallocate. In fact, appender when given data that's not mutable avoids the check for appendability, and extension into the block. So indeed, using appender on a string that is appendable reallocates on first append. I agree with you now, Appender on an existing string is near useless. Looks like there were some changes that fix one problem, but may have inadvertently created this problem. And I'm rereading my comments in this PR, finding that I was not exactly correct on my analysis: https://github.com/dlang/phobos/pull/2046 In short, there's a question of purity, which was already on the Appender's functions, causing issue with immutable data. I think this won't be strong pure anyway, because appender's functions are not strong-pure (there's always mutable data inside it). I think we should avoid the mutability check in that call.It will inspect the allocated length from the GC if the array is appendable from the beginning. So it's not always going to reallocate. e.g.: string x = "abc".idup; auto app = x.appender; app ~= "xyz"; // does not reallocate.Fair enough.My problem is this. const and immutable are *not* well supported in the standard library at the moment. What I want is that I can use the idioms of the stdlib, semantically, in a way that lets me *not care* about const or immutable, that lets me express the patterns I want without having to worry about whether some type deep inside my code, a dub library, or phobos itself decided to declare a field immutable. Appender covers two usecases: "capacity caching" and "memory reuse." Both of those usecases have two const variations: "possibly immutable" and "mutable". mutable capacity caching | mutable memory reuse ---------------------------+--------------------- immutable capacity caching | immutable memory reuse But instead of cutting between capacity caching and memory reuse, down the middle, Appender cuts between mutable and immutable, left to right. I think this is a symptom of a broad negligence of constness as a first-class property - constness works sometimes, and it's nice if it does, but it can't be *relied* on to be well supported. This makes using const in D an eternal uphill struggle. Why even go to all the trouble to make a new major version of the language just to introduce constness, if it's not going to be treated as a first-class concern? I don't want to have to sprinkle static if(isAssignable!T) everytime I want to use a library type. If Phobos offers me as generic a data structure as Appender, parameterized on T, it should work for T, regardless of what T is doing, and *certainly* regardless of what fields in T are marked const. Especially considering that if a field is not marked as const, that hardly means it's not going to lead to bugs if its memory is reused while you're referencing it!Appender covers appending, which works for both mutable and immutable data. The clear function, if called on immutable data, would allow for immutable data to be overwritten, which is undefined behavior. This should be harder to do than just calling a member function. Note that if we fix the above problem, there is a workaround: app = app.data[0 .. 0].assumeSafeAppend.appender; Which includes the greppable assumeSafeAppend property. The biggest problem I see is not necessarily that immutable arrays should be clearable, but that arrays of types with const items should be clearable. I can't say I agree with immutable pieces, as again, that leads to potential UB. I know it's a PITA, but we have to consider as a library type all possible usages, not just focused ones. I can see a PR that allows clear for types that are not assignable, but only have const data, being accepted. Const doesn't guarantee no mutability, so it's not UB to modify data that in some other alias has a const reference. I don't know if we have a way to determine this separate from immutability. -Steve
Jan 30 2019
On 1/24/19 7:35 AM, FeepingCreature wrote:On Tuesday, 27 March 2018 at 12:31:05 UTC, Simen Kjærås wrote:Appender is sometimes given a starting array. clear isn't callable in that case, and we don't distinguish the difference in the type or at runtime. I did a lot of work on std.array.Appender when I redid the array runtime to fix array appending back in 2010 to make it safe. e.g.: https://github.com/schveiguy/phobos/commit/7e89201cdae96e71e550afe72b59ef3db145916f Disabling clear from appender for non-mutable element types was one of the changes: https://github.com/schveiguy/phobos/commit/6636569318d26545bf7c36ddac029830a6d52531 This is before we were on github, so I don't know if there's any discussion of the issues, perhaps on the mailing list, but looking at my emails around that time, there are some emails that aren't in the mailing list. A funny thing I wrote to Walter/Sean around that time, when I was just getting druntime commit access: "I don't mind helping out in druntime, but I'm not really any kind of expert on the runtime, and I doubt I'll make many contributions besides this one without assignment." ;) FYI, this thread is almost a year old. -SteveOn Tuesday, 27 March 2018 at 12:17:58 UTC, Ellie Harper wrote:Isn't this wrong, though? Appender controls the memory it references. It could just choose to allocate non-immutable memory internally. As long as any const data put into the appender is *returned* as const, there is no chance of immutable memory being overwritten.Sorry if this is a stupid question, but is there something special required to call Appender.clear? When I attempt even just a simple use I am getting compile errors relating to `template object.clear`.From the documentation for Appender:Note clear is disabled for immutable or const element types, due to the possibility that Appender might overwrite immutable data.Since string is immutable(char)[], clear() is simply not available for appender!string.
Jan 24 2019