Re: about std.csv and derived format
Le jeudi 01 mars 2012 à 04:36 +0100, Jesse Phillips a écrit : On Thursday, 1 March 2012 at 02:07:44 UTC, bioinfornatics wrote: It is ok i have found a way maybe is not an efficient way but it works: https://gist.github.com/1946669 a minor bug exist for parse track line will be fixed tomorrow. time to bed Big thanks to all You can edit a gist instead of creating a new. This seems like a very fragile implementation, and hard to follow. My quick untested code: auto str = readText(filePath); // Ignoring first three lines. str = array(str.util(newline).until(newline).until(newline)); auto bedInstances = csvReader!(BedData11,Malformed.ignore)(str,'\t'); But if you must keep the separate structs, I don't have any better suggestions. and how convert bedInstances input array to BedData11[] ? Add a constructo to BedData11 and use std.algorithm.map? map!BedData11(a.filed1, a.filed2...)(bedInstances);
Re: about std.csv and derived format
On Thursday, 1 March 2012 at 10:09:55 UTC, bioinfornatics wrote: and how convert bedInstances input array to BedData11[] ? std.array.array()
Should uniform(-real.max, real.max) be inf?
What's the preferred way of generating a random floating-point number in the range of a given floating-point type? We have uniform!T() for integral types, but nothing similar for floats? And uniform(-real.max, real.max) (possibly tweaking the limits) seems to only return inf, which isn't terribly helpful. What's the standard thing to do here? I could just use uniform(cast(T) -1, cast(T) 1)*T.max I guess (for some floating-point type T). Seems to work fine, at least. Am I missing the obvious way to do it? -- Magnus Lie Hetland http://hetland.org
Re: Should uniform(-real.max, real.max) be inf?
On 2012-03-01 10:52:49 +, Magnus Lie Hetland said: I could just use uniform(cast(T) -1, cast(T) 1)*T.max I guess (for some floating-point type T). Seems to work fine, at least. Aaactually, not so much. The output here seems to get about the same exponent as T.max. Which isn't all that surprising, I guess. (Then again, most floating-point numbers *are* pretty large ;-) So ... any suggestions? -- Magnus Lie Hetland http://hetland.org
SysTime in a Struct
Hi,
I have defined this struct
struct preEv {
string edate; //010112
string etime; //00:00:00
string etext; //
SysTime esystime;
this (this) {
SysTime esystime = SysTime(DateTime(
Clock.currTime.year,
to!int(this.edate[2..4]),
to!int(this.edate[0..2]),
to!int(etime[0..2]),
to!int(etime[3..5]),
to!int(etime[6..8])));
}
}
If I write to the sctruct and then print it I'm able to see the
SysTime variable with a value.
writeln(preEv) //previousEvents(140212, 05:13:26, 9 140212
05:13:26 d, 2012-Feb-14 05:13:26)
but if trying to get the value from the SysTime variable I get a
Segmentation fault. Trying to read any other variable inside this
struct will not be a problem.
writeln (preEv.esystime.day) // will compile but segfaults
On DMD32 D Compiler v2.058
Any correct way to do this?
Thank you.
Dumb question about git
OK, so I'm new to git, and I ran into this problem: - I forked druntime on github and made some changes in a branch - Pushed the changes to the fork - Pulled upstream commits to master - Merged master with branch - Ran git rebase master, so that my changes appear on top of the latest upstream master. - Tried to push branch to my fork, but now it complains that I have non-fast-forward changes and rejects the push. What's the right thing to do here? Looks like I screwed up my branch history. How do I fix it? Thanks! T -- Real Programmers use cat a.out.
Re: Dumb question about git
When people say git encourages rewriting history. Don't listen. Once you have pushed your changes to the world they are immutable. This is because git uses cryptography internally and changing the history messes everything up. If you haven't pushed you can change all of your history and it will all be fine. But if someone else (github) has the old hisory bad things happen. If you are sure nobody has pulled from github you can use --force when pushing (I think). It will work no matter what but you will piss off people if they have pulled from you. Please note that this kind of history modifying is considered bad practice. On Mar 1, 2012 10:10 AM, H. S. Teoh hst...@quickfur.ath.cx wrote: OK, so I'm new to git, and I ran into this problem: - I forked druntime on github and made some changes in a branch - Pushed the changes to the fork - Pulled upstream commits to master - Merged master with branch - Ran git rebase master, so that my changes appear on top of the latest upstream master. - Tried to push branch to my fork, but now it complains that I have non-fast-forward changes and rejects the push. What's the right thing to do here? Looks like I screwed up my branch history. How do I fix it? Thanks! T -- Real Programmers use cat a.out.
Re: SysTime in a Struct
On 03/01/2012 06:15 AM, albatroz wrote:
Hi,
I have defined this struct
struct preEv {
string edate; //010112
string etime; //00:00:00
string etext; //
SysTime esystime;
That is a member of this type.
this (this) {
SysTime esystime = SysTime(DateTime(
That is a separate local variable within this(this). Also, this(this) is
the postblit (similar to a copy constructor). Is that what you want to
define?
I think you want to simply do this anyway:
esystime =
or this:
this.esystime =
Clock.currTime.year,
to!int(this.edate[2..4]),
to!int(this.edate[0..2]),
to!int(etime[0..2]),
to!int(etime[3..5]),
to!int(etime[6..8])));
}
}
If I write to the sctruct and then print it I'm able to see the
SysTime variable with a value.
writeln(preEv) //previousEvents(140212, 05:13:26, 9 140212
05:13:26 d, 2012-Feb-14 05:13:26)
but if trying to get the value from the SysTime variable I get a
Segmentation fault.
Probably because esystime member is no initialized.
Trying to read any other variable inside this
struct will not be a problem.
writeln (preEv.esystime.day) // will compile but segfaults
On DMD32 D Compiler v2.058
Any correct way to do this?
Thank you.
Ali
Re: Should uniform(-real.max, real.max) be inf?
On 03/01/2012 02:52 AM, Magnus Lie Hetland wrote: What's the preferred way of generating a random floating-point number in the range of a given floating-point type? We have uniform!T() for integral types, but nothing similar for floats? And uniform(-real.max, real.max) (possibly tweaking the limits) seems to only return inf, which isn't terribly helpful. What's the standard thing to do here? I could just use uniform(cast(T) -1, cast(T) 1)*T.max I guess (for some floating-point type T). Seems to work fine, at least. Am I missing the obvious way to do it? I recommend reading this page: http://dlang.org/d-floating-point.html Especially the ASCII graph there is very interesting. The number of distinct values between T.min_normal and 1 are equal to the distinct values between 1 and T.max. Since there are also sub-normal values between 0 and T.min_normal, it may make sense to use the range [T.min_normal, 1) and scale the result from there. But I haven't tested whether the distinct values in that range are equally distributed. Ali
Re: Regarding std.array.Appender
On 03/01/2012 03:40 AM, Jonathan M Davis wrote: On Wednesday, February 29, 2012 21:23:54 bearophile wrote: Jonathan M Davis: put is a function on output ranges, and Appender is an output range. Also, given that it doesn't define ~ (and it wouldn't really make sense for it to), it would be very weird IMHO to define ~=. I don't understand why that's weird. In Java you can't overload an append operator, so using a method is right. But for me it's weird that Appender doesn't use the D operator to _append_. I sometimes use add instead of put by mistake, forgetting the right method name, because I find it quite unnatural. If Appender needs a put, then I suggest to give it both put method and ~= operator. Would you define += without defining +? Or *= without defining *? It strikes me as a misuse of operator overloading if you have an opOpAssign without its corresponding opBinary. - Jonathan M Davis It is not the same thing. a=a~b has different semantics from a~=b;
Re: Dumb question about git
On 01.03.2012 19:11, H. S. Teoh wrote: OK, so I'm new to git, and I ran into this problem: - I forked druntime on github and made some changes in a branch - Pushed the changes to the fork I use the magic pull --rebase how-ever-you-call-dlang master instead of these 3 if I have changes but want to sync with upstream. - Pulled upstream commits to master - Merged master with branch - Ran git rebase master, so that my changes appear on top of the latest upstream master. That's pretty much all. - Tried to push branch to my fork, but now it complains that I have non-fast-forward changes and rejects the push. I think push --force should do the trick What's the right thing to do here? Looks like I screwed up my branch history. How do I fix it? Thanks! T -- Dmitry Olshansky
Re: SysTime in a Struct
Have fixed the segfault by using DateTime instead of SysTime.
That is a separate local variable within this(this). Also,
this(this) is the postblit (similar to a copy constructor). Is
that what you want to define?
No, but not using this(this) will fail to build with:
static variable _initialized cannot be read at compile time
Making the struct similar to what you suggest will work.
struct prevEv {
string edate;
string etime;
string etext;
DateTime edatetime;
this (this)
{
edatetime = DateTime(
Clock.currTime.year,
to!int(this.edate[2..4]),
to!int(this.edate[0..2]),
to!int(etime[0..2]),
to!int(etime[3..5]),
to!int(etime[6..8]));
}
}
Just not has expected, do you have any suggestion how to do it
properly?
writeln ( prevEv.edatetime ); //0001-Jan-01 00:00:00
writeln ( prevEv ); // preEv(140212, 05:13:26, 9 140212
05:13:26 d, 2012-Feb-14 05:13:26)
Ali
Thanks
Re: Dumb question about git
On Thu, Mar 01, 2012 at 10:22:33AM -0500, Kevin Cox wrote: When people say git encourages rewriting history. Don't listen. Once you have pushed your changes to the world they are immutable. This is because git uses cryptography internally and changing the history messes everything up. If you haven't pushed you can change all of your history and it will all be fine. But if someone else (github) has the old hisory bad things happen. If you are sure nobody has pulled from github you can use --force when pushing (I think). It will work no matter what but you will piss off people if they have pulled from you. Please note that this kind of history modifying is considered bad practice. [...] OK, so what's the right way to do it then? I have some changes in a branch, but master has been updated since, so I want to merge in the latest updates so that the branch changes are compatible with the latest code. If I just pull from master, then my changes get buried underneath the newest changes. I guess I still don't quite understand how things are supposed to work in situations like this. T -- Music critic: That's an imitation fugue!
Re: Dumb question about git
On Mar 1, 2012 12:15 PM, H. S. Teoh hst...@quickfur.ath.cx wrote: On Thu, Mar 01, 2012 at 10:22:33AM -0500, Kevin Cox wrote: When people say git encourages rewriting history. Don't listen. Once you have pushed your changes to the world they are immutable. This is because git uses cryptography internally and changing the history messes everything up. If you haven't pushed you can change all of your history and it will all be fine. But if someone else (github) has the old hisory bad things happen. If you are sure nobody has pulled from github you can use --force when pushing (I think). It will work no matter what but you will piss off people if they have pulled from you. Please note that this kind of history modifying is considered bad practice. [...] OK, so what's the right way to do it then? I have some changes in a branch, but master has been updated since, so I want to merge in the latest updates so that the branch changes are compatible with the latest code. If I just pull from master, then my changes get buried underneath the newest changes. I guess I still don't quite understand how things are supposed to work in situations like this. T -- Music critic: That's an imitation fugue! This time I would just --force. In the future the idea is to only push changes once you are happy with them. What I would do next time is: - work - commit - work - commit - pull - rebase (make what you have done look pretty) - push
Re: SysTime in a Struct
On 03/01/2012 09:14 AM, albatroz wrote:
Have fixed the segfault by using DateTime instead of SysTime.
That is a separate local variable within this(this). Also, this(this)
is the postblit (similar to a copy constructor). Is that what you want
to define?
No, but not using this(this) will fail to build with:
static variable _initialized cannot be read at compile time
So you are trying to initialize the member with a default initializer,
like this:
struct prevEv {
// ...
DateTime edatetime = DateTime(/* ... */);
}
For that to work, the initial value must be a compile-time value. I am
not sure that you want that.
Also, this(this) is needed in very rare cases. Are you trying to create
objects of prevEv? Then you should use this(). Unfortunately, structs
cannot have default constructors.
A solution might be to use a function that makes and returns a prevEv:
prevEv make_prevEv()
{
return prevEv(/* ... */, DateTime(/* ... */));
}
Making the struct similar to what you suggest will work.
struct prevEv {
string edate;
string etime;
string etext;
DateTime edatetime;
this (this)
{
edatetime = DateTime(
Clock.currTime.year,
to!int(this.edate[2..4]),
to!int(this.edate[0..2]),
to!int(etime[0..2]),
to!int(etime[3..5]),
to!int(etime[6..8]));
}
}
Just not has expected, do you have any suggestion how to do it properly?
Are you trying to record the time when a prevEv is copied from another
one? If not, I suggest not defining this(this). It is the postblit, to
make things right for rare structs and only when the compiler generated
copying is wrong for a that type.
Or, are you just trying to define a type that contains time information?
writeln ( prevEv.edatetime ); //0001-Jan-01 00:00:00
writeln ( prevEv ); // preEv(140212, 05:13:26, 9 140212 05:13:26
d, 2012-Feb-14 05:13:26)
Ali
Thanks
Ali
Re: Dumb question about git
OK, so what's the right way to do it then? I have some changes in a branch, but master has been updated since, so I want to merge in the latest updates so that the branch changes are compatible with the latest code. I use a quite crappy way to rebase my feature branch: git stash git checkout master git pull -v --rebase git rebase master myworkingbranch There's probably some redundancy or whatever here, but at least it works ^^
Using lazily ?
dear,
Noob question for know if D provide a shorter way i explain
we have a struct S:
struct S{
string member1;
string member2;
string member3;
}
we parse a file:
File f = File(a path, r);
S s;
sise_t tokenLength = member1.length;
foreach( char[] line; f.byLine() )
mixin(s. ~ line[0 .. tokenLength] ~ = ~ line[tokenLength ..
$] ); // do not work because lien is not kno at compile time
I know this do not works i.e comment but it will save some line by
checking if is member1 2 or 3
They are a shorter way to do this use lazy ?
Re: Using lazily ?
On 03/01/2012 10:50 PM, bioinfornatics wrote:
dear,
Noob question for know if D provide a shorter way i explain
we have a struct S:
struct S{
string member1;
string member2;
string member3;
}
we parse a file:
File f = File(a path, r);
S s;
sise_t tokenLength = member1.length;
foreach( char[] line; f.byLine() )
mixin(s. ~ line[0 .. tokenLength] ~ = ~ line[tokenLength ..
$] ); // do not work because lien is not kno at compile time
I know this do not works i.e comment but it will save some line by
checking if is member1 2 or 3
They are a shorter way to do this use lazy ?
struct S{
string member1;
string member2;
string member3;
}
S s;
size_t tokenLength = member1.length;
void main(){
foreach(char[] line; stdin.byLine()) foreach(m;__traits(allMembers,S)){
if(line[0..tokenLength] == m) mixin(s.~m) = line[tokenLength
.. $].idup;
}
}
Re: Dumb question about git
On Thursday, March 01, 2012 09:17:18 H. S. Teoh wrote: On Thu, Mar 01, 2012 at 10:22:33AM -0500, Kevin Cox wrote: When people say git encourages rewriting history. Don't listen. Once you have pushed your changes to the world they are immutable. This is because git uses cryptography internally and changing the history messes everything up. If you haven't pushed you can change all of your history and it will all be fine. But if someone else (github) has the old hisory bad things happen. If you are sure nobody has pulled from github you can use --force when pushing (I think). It will work no matter what but you will piss off people if they have pulled from you. Please note that this kind of history modifying is considered bad practice. [...] OK, so what's the right way to do it then? I have some changes in a branch, but master has been updated since, so I want to merge in the latest updates so that the branch changes are compatible with the latest code. If I just pull from master, then my changes get buried underneath the newest changes. I guess I still don't quite understand how things are supposed to work in situations like this. If you make changes in a branch and want them on top of what's in master, then do git-rebase master in _the branch_ and then merge the branch into master. And if you're using github to do pull requests and the like, then don't merge into master all. Simply create the pull request from the branch. That way, master only ever gets updated when the main repository gets updated, and you always have a clean version which matches the main repository. - Jonathan M Davis
Re: SysTime in a Struct
On Thursday, March 01, 2012 15:15:00 albatroz wrote:
Hi,
I have defined this struct
struct preEv {
string edate; //010112
string etime; //00:00:00
string etext; //
SysTime esystime;
this (this) {
SysTime esystime = SysTime(DateTime(
Clock.currTime.year,
to!int(this.edate[2..4]),
to!int(this.edate[0..2]),
to!int(etime[0..2]),
to!int(etime[3..5]),
to!int(etime[6..8])));
}
}
If I write to the sctruct and then print it I'm able to see the
SysTime variable with a value.
writeln(preEv) //previousEvents(140212, 05:13:26, 9 140212
05:13:26 d, 2012-Feb-14 05:13:26)
but if trying to get the value from the SysTime variable I get a
Segmentation fault. Trying to read any other variable inside this
struct will not be a problem.
writeln (preEv.esystime.day) // will compile but segfaults
On DMD32 D Compiler v2.058
Any correct way to do this?
A default-initialized SysTime is useless. It hasn't been properly initialized.
SysTime contains a time zone object, which is a class and must be initialized
at runtime (so directly initializing it won't work), and that object is null
in SysTime.init (and it can't be anything else, because you can't construct a
class at compile and have it persist to runtime). You need to actually
initialize a SysTime before using it. So, using the init value of struct which
has a SysTime as a member probably isn't a great idea.
- Jonathan M Davis
Re: Using lazily ?
Le jeudi 01 mars 2012 à 23:10 +0100, Timon Gehr a écrit :
S s;
size_t tokenLength = member1.length;
void main(){
foreach(char[] line; stdin.byLine())
foreach(m;__traits(allMembers,S)){
if(line[0..tokenLength] == m) mixin(s.~m) =
line[tokenLength
.. $].idup;
}
}
awesome :)
can use hasMember instead allMembers ?
Re: Using lazily ?
On 03/01/2012 02:25 PM, bioinfornatics wrote:
Le jeudi 01 mars 2012 à 23:10 +0100, Timon Gehr a écrit :
S s;
size_t tokenLength = member1.length;
void main(){
foreach(char[] line; stdin.byLine())
foreach(m;__traits(allMembers,S)){
if(line[0..tokenLength] == m) mixin(s.~m) =
line[tokenLength
.. $].idup;
}
}
awesome :)
can use hasMember instead allMembers ?
No, because both of those are compile-time features. Even if you used
hasMember, the answer will always be true:
if (__traits(hasMember, S, member1))
Yes, S has member1.
Note that Timon's inner foreach is a compile-time foreach, which is the
equivalent of the following three lines:
if(line[0..tokenLength] == member1) s.member1 = line[tokenLength ..
$].idup;
if(line[0..tokenLength] == member2) s.member2 = line[tokenLength ..
$].idup;
if(line[0..tokenLength] == member3) s.member3 = line[tokenLength ..
$].idup;
There is no inner foreach looop that is executed at runtime.
Ali
Re: Regarding std.array.Appender
Am 01.03.2012 03:40, schrieb Jonathan M Davis: On Wednesday, February 29, 2012 21:23:54 bearophile wrote: Jonathan M Davis: put is a function on output ranges, and Appender is an output range. Also, given that it doesn't define ~ (and it wouldn't really make sense for it to), it would be very weird IMHO to define ~=. I don't understand why that's weird. In Java you can't overload an append operator, so using a method is right. But for me it's weird that Appender doesn't use the D operator to _append_. I sometimes use add instead of put by mistake, forgetting the right method name, because I find it quite unnatural. If Appender needs a put, then I suggest to give it both put method and ~= operator. Would you define += without defining +? Or *= without defining *? It strikes me as a misuse of operator overloading if you have an opOpAssign without its corresponding opBinary. - Jonathan M Davis Consider matrix * vector and matrix *= vector for an opposite example.
Re: Using lazily ?
Ali:
Note that Timon's inner foreach is a compile-time foreach, which is the
equivalent of the following three lines:
I'd like it to be written:
static foreach (...) {...
In the meantime an annotation helps clarify the code for the person that will
read the code:
/*static*/ foreach (...) {...
Bye,
bearophile
Re: SysTime in a Struct
Are you trying to record the time when a prevEv is copied from another one? If not, I suggest not defining this(this). It is the postblit, to make things right for rare structs and only when the compiler generated copying is wrong for a that type. Or, are you just trying to define a type that contains time information? Hi Ali, just tring to define a type that holds this information. It was just an attempt to create a type DateTime with the values from the known strings, I thought it was possible to create the definition directly in the Struct, with no need for an external function. edate and etime are strings that I will read in to the struct, but for operations with time and dates I need to create/define a DateTime type. Thanks
Re: SysTime in a Struct
On 03/01/2012 03:46 PM, albatroz wrote:
Hi Ali, just tring to define a type that holds this information. It was
just an attempt to create a type DateTime with the values from the known
strings, I thought it was possible to create the definition directly in
the Struct, with no need for an external function.
edate and etime are strings that I will read in to the struct, but for
operations with time and dates I need to create/define a DateTime type.
From that description, it looks like you can hold edate etc. as members
and produce SysTime as needed. The following demonstrates how to convert
preEv to SysTime by opCast implicitly and by sys_time explicitly:
import std.stdio;
import std.conv;
import std.datetime;
struct preEv
{
string edate; //010112
string etime; //00:00:00
string etext; //
SysTime opCast(T : SysTime)() const
{
return SysTime(DateTime(
Clock.currTime.year,
to!int(this.edate[2..4]),
to!int(this.edate[0..2]),
to!int(etime[0..2]),
to!int(etime[3..5]),
to!int(etime[6..8])));
}
SysTime sys_time() const @property
{
return to!SysTime(this);
}
}
void main()
{
auto pe = preEv(010312, 15:53:00, The event);
// Explicit conversion
auto st0 = to!SysTime(pe);
writeln(st0);
// Casting
auto st1 = cast(SysTime)(pe);
writeln(st1);
// As a property
auto st2 = pe.sys_time;
writeln(st2);
}
If you think that you need to cache SysTime in the object itself, you
can do that for example in opCast.
On the other hand, if all you need to store is SysTime and etext, then
you need to create SysTime in the constructor:
import std.stdio;
import std.conv;
import std.datetime;
struct preEv
{
SysTime time;
string etext;
this (string edate, string etime, string etext)
{
this.time = SysTime(DateTime(
Clock.currTime.year,
to!int(edate[2..4]),
to!int(edate[0..2]),
to!int(etime[0..2]),
to!int(etime[3..5]),
to!int(etime[6..8])));
this.etext = etext;
}
}
void main()
{
auto pe = preEv(010312, 15:53:00, The event);
writeln(pe.time);
}
Ali
Re: SysTime in a Struct
On Thursday, March 01, 2012 16:00:09 Ali Çehreli wrote:
On 03/01/2012 03:46 PM, albatroz wrote:
Hi Ali, just tring to define a type that holds this information. It was
just an attempt to create a type DateTime with the values from the known
strings, I thought it was possible to create the definition directly in
the Struct, with no need for an external function.
edate and etime are strings that I will read in to the struct, but for
operations with time and dates I need to create/define a DateTime type.
From that description, it looks like you can hold edate etc. as members
and produce SysTime as needed. The following demonstrates how to convert
preEv to SysTime by opCast implicitly and by sys_time explicitly:
import std.stdio;
import std.conv;
import std.datetime;
struct preEv
{
string edate; //010112
string etime; //00:00:00
string etext; //
SysTime opCast(T : SysTime)() const
{
return SysTime(DateTime(
Clock.currTime.year,
to!int(this.edate[2..4]),
to!int(this.edate[0..2]),
to!int(etime[0..2]),
to!int(etime[3..5]),
to!int(etime[6..8])));
}
SysTime sys_time() const @property
{
return to!SysTime(this);
}
}
void main()
{
auto pe = preEv(010312, 15:53:00, The event);
// Explicit conversion
auto st0 = to!SysTime(pe);
writeln(st0);
// Casting
auto st1 = cast(SysTime)(pe);
writeln(st1);
// As a property
auto st2 = pe.sys_time;
writeln(st2);
}
If you think that you need to cache SysTime in the object itself, you
can do that for example in opCast.
On the other hand, if all you need to store is SysTime and etext, then
you need to create SysTime in the constructor:
import std.stdio;
import std.conv;
import std.datetime;
struct preEv
{
SysTime time;
string etext;
this (string edate, string etime, string etext)
{
this.time = SysTime(DateTime(
Clock.currTime.year,
to!int(edate[2..4]),
to!int(edate[0..2]),
to!int(etime[0..2]),
to!int(etime[3..5]),
to!int(etime[6..8])));
this.etext = etext;
}
}
void main()
{
auto pe = preEv(010312, 15:53:00, The event);
writeln(pe.time);
}
You know, you can create a TimeOfDay from 15:53:00 with
TimeOfDay.fromISOExtString. That won't work with the date, since it's not in
either the ISO or ISO Extended format, but it would work for the time.
I'd also point out that currTime isn't a property (since it takes an optional
TimeZone argument), so you really should be using parens when you call it.
Otherwise, once property enforcement is enabled, your code won't compile.
Also, you should use std.conv.to, not a cast, when converting, since
std.conv.to now supports calling user-defined opCasts, and there's less risk of
screwing up the cast if you use std.conv.to. So, defining an opCast is fine,
but
it should probably be used with std.conv.to rather than directly.
- Jonathan M Davis
Re: Dumb question about git
Unless you have an expectation that other people are already using the old version of your branch, just use 'git push blah -f' to overwrite the old version. It's not a big deal for patches and pull requests, but it would be a disaster if anyone did this to the master branch. H. S. Teoh hst...@quickfur.ath.cx wrote in message news:mailman.271.1330614611.24984.digitalmars-d-le...@puremagic.com... OK, so I'm new to git, and I ran into this problem: - I forked druntime on github and made some changes in a branch - Pushed the changes to the fork - Pulled upstream commits to master - Merged master with branch - Ran git rebase master, so that my changes appear on top of the latest upstream master. - Tried to push branch to my fork, but now it complains that I have non-fast-forward changes and rejects the push. What's the right thing to do here? Looks like I screwed up my branch history. How do I fix it? Thanks! T -- Real Programmers use cat a.out.
Char * character and string
Hello,
I am trying to work with SDL and one of their functions takes a
char * file as a function a parameter. However, i'm running into
trouble how to actually set this variable in my constructor.
I am getting a problem where if I use a pointer to a char and set
it as test.bmp I get an error stating cannot implicitly
convert expression (file) of type string to char*. After that I
decided to try to set file to 'test.bmp' instead, and in that
case I get: Unterminated character constant . Although I am
familiar with what this error is referring to, I do not know how
to add a terminator in D.
This is the function that I intend to use the filename in:
**Note the function LoadBMP is the one that REQUIRES a pointer to
a char
--
SDL_Surface * Load(char * file)
{
SDL_Surface * Temp = null;
if((Temp = SDL_LoadBMP(file)) == null)
return null;
Surface = SDLDisplayFormat(Temp);
SDL_FreeSurface(Temp);
return Surface;
}
This is the constructor that is giving me the error:
--
char * file;
this()
{
this.filename = test.bmp;
}
-
Re: Char * character and string
On Friday, March 02, 2012 05:51:14 Chris Pons wrote:
Hello,
I am trying to work with SDL and one of their functions takes a
char * file as a function a parameter. However, i'm running into
trouble how to actually set this variable in my constructor.
I am getting a problem where if I use a pointer to a char and set
it as test.bmp I get an error stating cannot implicitly
convert expression (file) of type string to char*. After that I
decided to try to set file to 'test.bmp' instead, and in that
case I get: Unterminated character constant . Although I am
familiar with what this error is referring to, I do not know how
to add a terminator in D.
This is the function that I intend to use the filename in:
**Note the function LoadBMP is the one that REQUIRES a pointer to
a char
--
SDL_Surface * Load(char * file)
{
SDL_Surface * Temp = null;
if((Temp = SDL_LoadBMP(file)) == null)
return null;
Surface = SDLDisplayFormat(Temp);
SDL_FreeSurface(Temp);
return Surface;
}
This is the constructor that is giving me the error:
--
char * file;
this()
{
this.filename = test.bmp;
}
-
Use std.string.toStringz (or std.utf.toUTFz if immutable(char)* doesn't cut
it).
Regardless, be careful of passing char* to C code. Even if the C code keeps
the pointer, it won't stop the GC from collecting it. The D code needs to have
reference to it of some kind (though a pointer in your struct as you appeart
to be doing should be plenty as long as an instance of the struct remains in
the D code - as opposed to passing it to the C code and then not having it in
the D code anymore, which would be no better than just passing the char* to
the C code without keeping a reference to it).
- Jonathan M Davis
Re: Char * character and string
Thank you for the reply. However, I've run into another problem.
I changed:
---
char * file;
this()
{
this.filename = test.bmp;
}
---
To:
char * file
this()
{
this.filename = toStringz(test.bmp);
}
---
I am getting this error:
Error 1 Error: cannot implicitly convert expression
(toStringz(test.bmp)) of type immutable(char)* to
char* D:\Documents\Projects\Test\Test\DPBall.d 10
Instead I tried toUTFz, which I used like this:
--
char * filename;
this()
{
this.filename = toUTFz(test.bmp);
}
--
I get these errors:
Error 1 Error: template std.utf.toUTFz(P,S) if (isSomeString!(S)
isPointer!(P) isSomeChar!(typeof(*P.init))
is(Unqual!(typeof(*P.init)) == Unqual!(ElementEncodingType!(S)))
is(immutable(Unqual!(ElementEncodingType!(S))) ==
ElementEncodingType!(S))) does not match any function template
declaration D:\Documents\Projects\Test\Test\DPBall.d 11
Error 2 Error: template std.utf.toUTFz(P,S) if (isSomeString!(S)
isPointer!(P) isSomeChar!(typeof(*P.init))
is(Unqual!(typeof(*P.init)) == Unqual!(ElementEncodingType!(S)))
is(immutable(Unqual!(ElementEncodingType!(S))) ==
ElementEncodingType!(S))) cannot deduce template function from
argument types
!()(string) D:\Documents\Projects\Test\Test\DPBall.d 11
Am I using these functions incorrectly?
Cocoa bindings?
Hi, Are there any actively-maintained Cocoa bindings for D? -- - Alex
Re: Char * character and string
On Friday, March 02, 2012 06:22:41 Chris Pons wrote:
Thank you for the reply. However, I've run into another problem.
I changed:
---
char * file;
this()
{
this.filename = test.bmp;
}
---
To:
char * file
this()
{
this.filename = toStringz(test.bmp);
}
---
I am getting this error:
Error 1 Error: cannot implicitly convert expression
(toStringz(test.bmp)) of type immutable(char)* to
char* D:\Documents\Projects\Test\Test\DPBall.d10
Instead I tried toUTFz, which I used like this:
--
char * filename;
this()
{
this.filename = toUTFz(test.bmp);
}
--
I get these errors:
Error 1 Error: template std.utf.toUTFz(P,S) if (isSomeString!(S)
isPointer!(P) isSomeChar!(typeof(*P.init))
is(Unqual!(typeof(*P.init)) == Unqual!(ElementEncodingType!(S)))
is(immutable(Unqual!(ElementEncodingType!(S))) ==
ElementEncodingType!(S))) does not match any function template
declaration D:\Documents\Projects\Test\Test\DPBall.d11
Error 2 Error: template std.utf.toUTFz(P,S) if (isSomeString!(S)
isPointer!(P) isSomeChar!(typeof(*P.init))
is(Unqual!(typeof(*P.init)) == Unqual!(ElementEncodingType!(S)))
is(immutable(Unqual!(ElementEncodingType!(S))) ==
ElementEncodingType!(S))) cannot deduce template function from
argument types
!()(string) D:\Documents\Projects\Test\Test\DPBall.d11
Am I using these functions incorrectly?
toStringz returns an immutable(char)*, so you can assign it to a const(char)*
or an immutable(char)* but not a char*. If you want to assign to a char*, then
you need to use toUTFz, which you are most definitely using incorrectly. The
documentation gives several examples on how to use it correctly, but if what
you want is a char*, then you'd do
this.filename = toUTFz!(char*)(test.bmp);
toUTFZ is a templated function which requires that you give it the type that
you want to convert to. You were trying to call it without giving the type.
- Jonathan M Davis
Re: Char * character and string
Ok, got it all sorted. Thank you for the guidance.
Re: Cocoa bindings?
On 2 March 2012 18:52, Alex Rønne Petersen xtzgzo...@gmail.com wrote: Hi, Are there any actively-maintained Cocoa bindings for D? -- - Alex Not as far as I know. You should make some! -- James Miller
