www.digitalmars.com         C & C++   DMDScript  

digitalmars.D - Nicer anchors

reply Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
Please destroy 
https://github.com/D-Programming-Language/dlang.org/pull/734 -- Andrei
Dec 29 2014
next sibling parent reply "Adam D. Ruppe" <destructionator gmail.com> writes:
Script? Gross, we should just change DDOC_ANCHOR in doc.c to 
output ident->toPrettyChars instead of ident->string.

I think it seriously is *that* simple.
Dec 29 2014
parent reply Andrei Alexandrescu <SeeWebsiteForEmail erdan.org> writes:
"Adam D. Ruppe" <destructionator gmail.com> wrote:

 Script? Gross, we should just change DDOC_ANCHOR in doc.c to output
 ident->toPrettyChars instead of ident->string.
 
 I think it seriously is *that* simple.


How would that achieve cross referencing?
Dec 29 2014
parent reply "Adam D. Ruppe" <destructionator gmail.com> writes:
On Tuesday, 30 December 2014 at 01:18:39 UTC, Andrei Alexandrescu 
wrote:

 How would that achieve cross referencing?


To be clear, what your change does is make links like this work, 
yes?



(BTW, nice leading dot there...)


The toPrettyChars function outputs the fully-qualified name too 
without needing a script.
Dec 29 2014
next sibling parent Nick Treleaven <ntrel-pub mybtinternet.com> writes:
On 30/12/2014 01:27, Adam D. Ruppe wrote:

 On Tuesday, 30 December 2014 at 01:18:39 UTC, Andrei Alexandrescu wrote:

 How would that achieve cross referencing?


 To be clear, what your change does is make links like this work, yes?




No, this has worked for some time.


 (BTW, nice leading dot there...)


Necessary for full backwards link compatibility.


 The toPrettyChars function outputs the fully-qualified name too without
 needing a script.


DDOC_ANCHOR is fully qualified (and its more complicated than just 
toPrettyChars).
Dec 30 2014
prev sibling parent reply Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
On 12/29/14 5:27 PM, Adam D. Ruppe wrote:

 On Tuesday, 30 December 2014 at 01:18:39 UTC, Andrei Alexandrescu wrote:

 How would that achieve cross referencing?


 To be clear, what your change does is make links like this work, yes?



 (BTW, nice leading dot there...)


 The toPrettyChars function outputs the fully-qualified name too without
 needing a script.


Look at the "Jump to:" anchors and compare them:

https://web.archive.org/web/20141129100337/http://dlang.org/phobos/std_datetime.html

vs.

http://dlang.org/phobos/std_datetime.html

There is no comparison. It's a shame it took me so long to finish this.


Andrei
Dec 30 2014
parent reply "Adam D. Ruppe" <destructionator gmail.com> writes:
On Tuesday, 30 December 2014 at 18:00:38 UTC, Andrei Alexandrescu 
wrote:

 Look at the "Jump to:" anchors and compare them:


Ah, I was looking in completely the wrong place and with 
javascript disabled nothing was there anyway.

Well, yeah, that's a step up from where it was. Not where it 
should be yet though, I think a list or table that gives a 
summary would be the next nice step.
Dec 30 2014
parent Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
On 12/30/14 10:24 AM, Adam D. Ruppe wrote:

 On Tuesday, 30 December 2014 at 18:00:38 UTC, Andrei Alexandrescu wrote:

 Look at the "Jump to:" anchors and compare them:


 Ah, I was looking in completely the wrong place and with javascript
 disabled nothing was there anyway.

 Well, yeah, that's a step up from where it was. Not where it should be
 yet though, I think a list or table that gives a summary would be the
 next nice step.


Then DO IT. https://www.youtube.com/watch?v=JoqDYcCDOTg -- Andrei
Dec 30 2014
prev sibling next sibling parent reply "H. S. Teoh via Digitalmars-d" <digitalmars-d puremagic.com> writes:
On Mon, Dec 29, 2014 at 04:19:53PM -0800, Andrei Alexandrescu via Digitalmars-d
wrote:

 Please destroy https://github.com/D-Programming-Language/dlang.org/pull/734
 -- Andrei


Please don't.

The *real* fix is to fix dmd to output nicer anchors, not patching
broken output with, of all things, a *script*.


T

-- 
Making non-nullable pointers is just plugging one hole in a cheese grater. --
Walter Bright
Dec 29 2014
parent Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
On 12/29/14 4:27 PM, H. S. Teoh via Digitalmars-d wrote:

 On Mon, Dec 29, 2014 at 04:19:53PM -0800, Andrei Alexandrescu via
Digitalmars-d wrote:

 Please destroy https://github.com/D-Programming-Language/dlang.org/pull/734
 -- Andrei


 Please don't.

 The *real* fix is to fix dmd to output nicer anchors, not patching
 broken output with, of all things, a *script*.


You'd need a much better argument to be convincing. Each platform has 
its own way of doing referencing; LaTeX runs multiple times, and for 
HTML using scripts seems appropriate to me. I'd see generation of 
platform-independent indexes a major effort. -- Andrei
Dec 29 2014
prev sibling next sibling parent reply "Vladimir Panteleev" <vladimir thecybershadow.net> writes:
On Tuesday, 30 December 2014 at 00:19:54 UTC, Andrei Alexandrescu 
wrote:

 Please destroy 
 https://github.com/D-Programming-Language/dlang.org/pull/734 -- 
 Andrei


I'm not sure using numbers to distinguish overloads is a good 
idea. Such links would break easily.

Since you seem to have time to work on the presentation of the 
documentation, I must ask: what's the current status of DDox? 
IIRC, the effort was mainly hampered by there being no one to 
update the software on the website, and you had neither time to 
do it yourself nor time to delegate the responsibility to someone 
else.
Dec 29 2014
next sibling parent reply Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
On 12/29/14 6:18 PM, Vladimir Panteleev wrote:

 On Tuesday, 30 December 2014 at 00:19:54 UTC, Andrei Alexandrescu wrote:

 Please destroy
 https://github.com/D-Programming-Language/dlang.org/pull/734 -- Andrei


 I'm not sure using numbers to distinguish overloads is a good idea. Such
 links would break easily.

 Since you seem to have time to work on the presentation of the
 documentation, I must ask: what's the current status of DDox? IIRC, the
 effort was mainly hampered by there being no one to update the software
 on the website, and you had neither time to do it yourself nor time to
 delegate the responsibility to someone else.


Some support for dub exists but needs some tender love and care. dub is 
present under github.com/D-Programming-Language, but (unlike is the case 
for dmd/druntime/phobos) the dub project is not automatically checked 
out and built as a dependency.

As a rule of thumb, anyone should get the project and be able to type 
"make" and get things build. I've done my best to achieve that and keep 
it working, breaking things on occasion but not too often. Sadly others 
tend to be more cavalier about changing makefiles, and with every change 
to them things seem to get worse in one way or another. Building things 
with dub is not easily achievable right now. Right now it fails on my 
machine with this message: 
http://paste.ofcode.org/W4QYryFttXLUm9z3GV2RbN. To fix things now I'd 
need to essentially embark on a project. I hope to convince Martin Nowak 
or someone else to do so.

There is a possible response on my part that is unsettlingly appealing, 
and that is frustration. After all _anyone_ concerned with the state of 
ddox could check for themselves in less than a minute by attempting to 
build the docs. The irritation pattern that seems to be present in this 
group as of late only invites more of it.

The better route out of this is to kindly invite contributors who added 
ddox support, and all others who are interested, to improve ddox build 
to the point of making it robustly usable. I trust you are out there 
reading this. Thanks.


Andrei
Dec 29 2014
next sibling parent "Vladimir Panteleev" <vladimir thecybershadow.net> writes:
On Tuesday, 30 December 2014 at 04:27:33 UTC, Andrei Alexandrescu 
wrote:

 appealing, and that is frustration. After all _anyone_ 
 concerned with the state of ddox could check for themselves in 
 less than a minute by attempting to build the docs. The 
 irritation pattern that seems to be present in this group as of 
 late only invites more of it.


Well, to fix something, one would first need to know that it's 
broken. (It would take me more than a minute to check as I don't 
use dub or have it installed.)


 The better route out of this is to kindly invite contributors 
 who added ddox support, and all others who are interested, to 
 improve ddox build to the point of making it robustly usable. I 
 trust you are out there reading this. Thanks.


I think you are entirely right in that such issues should be 
resolved not by D project leaders, but delegated to Dub/DDox 
maintainers. But we should be aware of possible communication 
issues regarding in whose court the ball is at the moment.

I've filed an issue, hope this helps:

https://github.com/rejectedsoftware/ddox/issues/69
Dec 30 2014
prev sibling parent reply "Vladimir Panteleev" <vladimir thecybershadow.net> writes:
On Tuesday, 30 December 2014 at 04:27:33 UTC, Andrei Alexandrescu 
wrote:

 There is a possible response on my part that is unsettlingly 
 appealing, and that is frustration. After all _anyone_ 
 concerned with the state of ddox could check for themselves in 
 less than a minute by attempting to build the docs. The 
 irritation pattern that seems to be present in this group as of 
 late only invites more of it.


According to Martin, the problem is on your side after all. 
Please update Dub.
Jan 02 2015
parent Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
On 1/2/15 8:01 PM, Vladimir Panteleev wrote:

 On Tuesday, 30 December 2014 at 04:27:33 UTC, Andrei Alexandrescu wrote:

 There is a possible response on my part that is unsettlingly
 appealing, and that is frustration. After all _anyone_ concerned with
 the state of ddox could check for themselves in less than a minute by
 attempting to build the docs. The irritation pattern that seems to be
 present in this group as of late only invites more of it.


 According to Martin, the problem is on your side after all. Please
 update Dub.


Nope. https://github.com/D-Programming-Language/dlang.org/pull/742 -- Andrei
Jan 02 2015
prev sibling parent reply Nick Treleaven <ntrel-pub mybtinternet.com> writes:
On 30/12/2014 02:18, Vladimir Panteleev wrote:

 I'm not sure using numbers to distinguish overloads is a good idea. Such
 links would break easily.


Note: that isn't part of Andrei's pull.

Using numbers is the only workable way, other schemes would also break 
easily, probably moreso. I implemented it so that ditto overloads don't 
get a separate number, to avoid some breakage when adding dittos.
Dec 30 2014
parent reply Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
On 12/30/14 8:59 AM, Nick Treleaven wrote:

 On 30/12/2014 02:18, Vladimir Panteleev wrote:

 I'm not sure using numbers to distinguish overloads is a good idea. Such
 links would break easily.


 Note: that isn't part of Andrei's pull.


Yah, I was surprised when I saw'em.


 Using numbers is the only workable way, other schemes would also break
 easily, probably moreso. I implemented it so that ditto overloads don't
 get a separate number, to avoid some breakage when adding dittos.


Agreed. I think the best solution is to simply group overloads together. 
They already do related things in any given module, so consolidating 
documentation makes sense anyway.


Andrei
Dec 30 2014
parent "H. S. Teoh via Digitalmars-d" <digitalmars-d puremagic.com> writes:
On Tue, Dec 30, 2014 at 09:55:17AM -0800, Andrei Alexandrescu via Digitalmars-d
wrote:

 On 12/30/14 8:59 AM, Nick Treleaven wrote:

On 30/12/2014 02:18, Vladimir Panteleev wrote:

I'm not sure using numbers to distinguish overloads is a good idea. Such
links would break easily.


Note: that isn't part of Andrei's pull.


 
 Yah, I was surprised when I saw'em.
 

Using numbers is the only workable way, other schemes would also break
easily, probably moreso. I implemented it so that ditto overloads don't
get a separate number, to avoid some breakage when adding dittos.


 
 Agreed. I think the best solution is to simply group overloads together.
 They already do related things in any given module, so consolidating
 documentation makes sense anyway.


[...]

Note that consolidating docs for overloads doesn't work as well as it
should, due to:

	https://issues.dlang.org/show_bug.cgi?id=13270


T

-- 
Customer support: the art of getting your clients to pay for your own
incompetence.
Dec 30 2014
prev sibling parent reply Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
On 12/29/14 4:19 PM, Andrei Alexandrescu wrote:

 Please destroy
 https://github.com/D-Programming-Language/dlang.org/pull/734 -- Andrei


Improved anchors are up on the site. Please try'em out! -- Andrei
Dec 29 2014
parent reply "Kapps" <opantm2+spam gmail.com> writes:
On Tuesday, 30 December 2014 at 05:30:26 UTC, Andrei Alexandrescu 
wrote:

 On 12/29/14 4:19 PM, Andrei Alexandrescu wrote:

 Please destroy
 https://github.com/D-Programming-Language/dlang.org/pull/734 
 -- Andrei


 Improved anchors are up on the site. Please try'em out! -- 
 Andrei


http://dlang.org/phobos/std_algorithm.html
Just has "quickindex, quickindex" displayed instead of any 
anchors.
Dec 29 2014
parent Andrei Alexandrescu <SeeWebsiteForEmail erdani.org> writes:
On 12/29/14 11:11 PM, Kapps wrote:

 On Tuesday, 30 December 2014 at 05:30:26 UTC, Andrei Alexandrescu wrote:

 On 12/29/14 4:19 PM, Andrei Alexandrescu wrote:

 Please destroy
 https://github.com/D-Programming-Language/dlang.org/pull/734 -- Andrei


 Improved anchors are up on the site. Please try'em out! -- Andrei


 http://dlang.org/phobos/std_algorithm.html
 Just has "quickindex, quickindex" displayed instead of any anchors.


Ah, thanks. Fixed on the site, here's the pr: 
https://github.com/D-Programming-Language/dlang.org/pull/735 -- Andrei
Dec 30 2014