On 2013-04-03, 20:04, Jonathan M Davis wrote:

> On Wednesday, April 03, 2013 19:59:37 Brad Anderson wrote:
>> Perhaps 34k is too large but 2k is laughable.
>
> I really should strip out the unit tests and documentation to see what the
> line count of actual code is, as something like 75% of that is unit tests and
> documentation, and IIRC, std.datetime provides most of the functionality that
> Boost does plus some more, though it does some weird, complicated stuff with
> its header files from what I recall. I'd hate to be the maintainer of Boost's
> datetime stuff.

Removed all comments, unittests, and empty lines from std.datetime. File
went from 34070 to 5843 lines.


-- 
Simen

On 4/3/13 2:55 PM, Jacob Carlborg wrote:
> On 2013-04-03 19:39, Andrei Alexandrescu wrote:
>
>> The way I see it, the first is terrible and the second asks for better
>> focus on a data-driven approach.
>
> Stupid me, posting on Ruby.

I was referring to the repeatability of the code used in testing, which is language-independent.

Andrei

On Wednesday, 3 April 2013 at 19:28:56 UTC, Simen Kjaeraas wrote:
> On 2013-04-03, 20:04, Jonathan M Davis wrote:
>
>> On Wednesday, April 03, 2013 19:59:37 Brad Anderson wrote:
>>> Perhaps 34k is too large but 2k is laughable.
>>
>> I really should strip out the unit tests and documentation to see what the
>> line count of actual code is, as something like 75% of that is unit tests and
>> documentation, and IIRC, std.datetime provides most of the functionality that
>> Boost does plus some more, though it does some weird, complicated stuff with
>> its header files from what I recall. I'd hate to be the maintainer of Boost's
>> datetime stuff.
>
> Removed all comments, unittests, and empty lines from std.datetime. File
> went from 34070 to 5843 lines.

cloc doesn't support /+ comments... But using your number, cloc, and some math

loc: 5843
comments: 6255
unittest: 16503
blank: 5469

On Wed, 03 Apr 2013 14:42:12 -0400, Walter Bright <newshound2@digitalmars.com> wrote:

> On 4/3/2013 9:49 AM, Dmitry Olshansky wrote:
>> +1
>
> Stylistic nit:
>
> When writing a one-liner post like this, please do not quote the entire preceding post, especially if it is long. We have great forum software, and the newsreaders as well are great at navigating the threads.

I couldn't disagree more.  The given +1 had 4 lines of context.  There was some straggling text after it, but this was only an additional 5 lines.

My newsreader highlights replied-to text in different colors depending on the level of indent.  I can immediately pick out new replies, and if I don't want to read the re-posted stuff, I don't have to, unless I want to for context.

Newsreaders are known not to thread things properly, and some people's posts don't thread properly ANYWHERE.  Context is important.

> Not to pick on you, but I see this a lot here from many of our participants and finally felt compelled to speak up!

I find posts that are solely about how you didn't "post properly" annoying.  Kind of like compulsively telling someone they didn't use correct grammar (for which I have to fight my instincts in order to remain married).  Sorry, I had to say something ;)

-Steve

On 4/3/13 11:24 PM, Steven Schveighoffer wrote:
> On Wed, 03 Apr 2013 14:42:12 -0400, Walter Bright
> <newshound2@digitalmars.com> wrote:
>
>> On 4/3/2013 9:49 AM, Dmitry Olshansky wrote:
>>> +1
>>
>> Stylistic nit:
>>
>> When writing a one-liner post like this, please do not quote the
>> entire preceding post, especially if it is long. We have great forum
>> software, and the newsreaders as well are great at navigating the
>> threads.
>
> I couldn't disagree more. The given +1 had 4 lines of context. There was
> some straggling text after it, but this was only an additional 5 lines.

I'm with Walter. The top context was fine for that message. The bottom was not seeing as the poster had nothing to say about it. Deleting the bottom is good common courtesy.

Walter himself used to leave vast amounts of trailing context in our communication, and it saved me significant time when he started to consistently trim it. With trailing chaff, essentially every reader needs to scroll down to find "is there anything more this guy wanted to add"? Some don't even insert an empty line.

> My newsreader highlights replied-to text in different colors depending
> on the level of indent. I can immediately pick out new replies, and if I
> don't want to read the re-posted stuff, I don't have to, unless I want
> to for context.

Mine too, but that doesn't make the problem go away.

> Newsreaders are known not to thread things properly, and some people's
> posts don't thread properly ANYWHERE. Context is important.

Yes, just not trailing chaff.

>> Not to pick on you, but I see this a lot here from many of our
>> participants and finally felt compelled to speak up!
>
> I find posts that are solely about how you didn't "post properly"
> annoying. Kind of like compulsively telling someone they didn't use
> correct grammar (for which I have to fight my instincts in order to
> remain married). Sorry, I had to say something ;)

Such posts are good because netiquette is not as widespread and as agreed upon as grammar.


Andrei

On 2013-04-03 22:50, Andrei Alexandrescu wrote:

> I was referring to the repeatability of the code used in testing, which
> is language-independent.

I think the first one is far more readable then the one using the loop.

-- 
/Jacob Carlborg

On 2013-04-03 21:28, Simen Kjaeraas wrote:

> Removed all comments, unittests, and empty lines from std.datetime. File
> went from 34070 to 5843 lines.

Heheh, that's more reasonable. That's also why I don't like to have unit tests inline.

-- 
/Jacob Carlborg

On 2013-04-04 03:47, Jesse Phillips wrote:

> cloc doesn't support /+ comments... But using your number, cloc, and
> some math

std.datetime contains mostly /+ and // comments. It only contains a single /* comment.

-- 
/Jacob Carlborg

On Thursday, 4 April 2013 at 14:31:36 UTC, Jacob Carlborg wrote:
> On 2013-04-04 03:47, Jesse Phillips wrote:
>
>> cloc doesn't support /+ comments... But using your number, cloc, and
>> some math
>
> std.datetime contains mostly /+ and // comments. It only contains a single /* comment.

I realize that, reason I had to use math. Cloc reports 11598 (something near that) then I know subtracted the actual loc gives me the /+ comments.

On 4/4/13 10:26 AM, Jacob Carlborg wrote:
> On 2013-04-03 22:50, Andrei Alexandrescu wrote:
>
>> I was referring to the repeatability of the code used in testing, which
>> is language-independent.
>
> I think the first one is far more readable then the one using the loop.

I understand. And I think you are very wrong about that.

Andrei