And .... it works. THANKS.
On Friday, March 24, 2017 at 11:09:20 PM UTC+1, Paul Atlan wrote:
>
> Thanks Joel, I was slowly getting my way to something like this.
> I'd figured out the weirdness of using when/splice and when, although I
> hadn't grokked that '(@ ...) is a pollen tagged x--expression ...
> I also had to figure out that, outside of the "string append" function, we
> need to escape the \date{ string :-)
>
> I'm at the point where I've actually split the latex-source into three
> substrings that I splice together. Your solution is obviously much more
> elegant - I'll patch that in.
>
> Thanks again to both for taking the time to answer, and sorry Matthew for
> the cryptic clues :-)
>
> On Friday, March 24, 2017 at 10:09:15 PM UTC+1, Joel Dueck wrote:
>>
>> It's a bit esoteric for sure. The problem is something I didn't think of
>> when I saw Matthew's remarks because I didn't fill in the context. Short
>> story is, you can't actually use when/splice directly inside a call to
>> string-append.
>>
>> When its "if" condition evaluates to anything other than #f, ◊when/splice
>> produces an X-expression with the @ symbol in front. E.g.
>> when/splice[#t]{xyz} produces '(@ "xyz") In the context of a Pollen
>> tagged X-expression, @ means "splice this x-expression into the
>> surrounding one" -- or essentially, "drop the @ and the outermost
>> parentheses".
>>
>> However, this when/splice is not directly inside a Pollen tagged
>> X-expression. Instead it's inside a list of arguments to a string-append
>> function call. And string-append expects everything you send it to be a
>> string and nothing else. So you really can't use when/splice in there
>> since it returns an X-expression and not a string.
>>
>> (Also: using symbol->string doesn't work because it expects a symbol as
>> input, and '(@ "\\date{"...) isn't a symbol but a list.)
>>
>> The thing to be sure of, then, is that any function calls inside
>> string-append produce only strings.
>>
>> So, one good approach would be to prepare a value, outside of
>> string-append, which evaluates to either the entire \date{} command or
>> an empty string, then use that:
>>
>> ◊(define maybe-date (hash-ref metas 'doc-publish-date #f))
>> ◊(define date-command (if maybe-date
>> (string-append "\\date{" (pubdate->english
>> maybe-date) "}")
>> ""))
>>
>> Then inside the string-append call, just replace the whole \date{} line
>> with:
>>
>> ◊date-command
>>
>> Hopefully I've got that right...
>>
>> On Thursday, March 23, 2017 at 6:29:25 PM UTC-5, Paul Atlan wrote:
>>
>>> Thanks Matthew and Joel,
>>> * yes, I took the brackets literally - oops
>>> * I know I could just skip the date thing, but where's the teaching
>>> moment in that ....
>>>
>>> ... as shown by the fact that I'm still struggling:
>>> * I've defined the "maybe-date" outside of the ◊(define latex-source
>>> ◊string-append{} loop. So that works.
>>> * however,
>>> ◊when/splice[maybe-date]{
>>> \date{◊(pubdate->english maybe-date)}}
>>> gives me:
>>>
>>> string-append: contract violation
>>> expected: string?
>>> given: '(@ "\\date{" "Saturday, February 20th, 2016" "}")
>>> argument position: 74th
>>> other arguments...:... [165 total] ...
>>>
>>>
>>> I guess instead of a string I'm outputting "something" (a quote?) that's
>>> a concatenation o strings ...
>>> I thought I could convert this into a simple string by using
>>> symbol->string but no joy.
>>> I'm pretty sure I'm missing something obvious about lists and strings
>>> here, but I can't figure out what ...
>>>
>>>
>>> On Thursday, March 23, 2017 at 3:25:23 AM UTC+1, Joel Dueck wrote:
>>>>
>>>> Matthew’s answer is great. I would just add that the main reason I made
>>>> us of dates in the metas was that I planned to have an RSS feed for
>>>> chapters, and things syndicated in a feed need to have a date stamp. But
>>>> the feed is also coded so as to omit anything that doesn’t have a
>>>> 'doc-publish-date in the metas.
>>>>
>>>> So if you won’t be publishing an RSS feed, and if your book doesn’t
>>>> particularly need dates on each chapter, you could just remove any
>>>> reference to the date in the templates.
>>>>
>>>> On Wednesday, March 22, 2017 at 12:20:29 AM UTC-5, Paul Atlan wrote:
>>>>>
>>>>> I'm slowly winding my way up the learning curve by adapting Joel
>>>>> Dueck's work to a personal web-book I want to write.
>>>>> I've hit a snag due to my limited understanding of racket, probably.
>>>>>
>>>>>
>>>>> The template.pdf.p files contains the following line to generate a
>>>>> nicely formatted LaTex date:
>>>>> \date{◊(pubdate->english (hash-ref metas 'doc-publish-date))}
>>>>>
>>>>> *metas *is generated from meta values in the source files, with one
>>>>> of the keys being "doc-publish-date".
>>>>>
>>>>> *pubdate->English * is a helper function to format everything nicely.
>>>>>
>>>>> This works perfectly if I don't forget the "doc-publish-date" key in
>>>>> the source files. If I do (or don't want to set it), the *hash-ref*
>>>>> fails and stops the rendering process.
>>>>>
>>>>> The *hash-ref* documentation (hash-ref
>>>>> <http://docs.racket-lang.org/reference/hashtables.html?q=hash-ref%20#%28def._%28%28quote._~23~25kernel%29._hash-ref%29%29>)
>>>>> mentions
>>>>> that one can write:
>>>>>
>>>>> hash-ref hash key [failure-result]
>>>>>
>>>>> were *failure-result* is either a procedure or is a value and gets
>>>>> returned instead of the missing key-value.
>>>>>
>>>>> I've been trying to use this in order to fail gracefully when
>>>>> *doc-publish-date
>>>>> *is missing:
>>>>> * either by returnnig a blank value (so the LaTex output will be "
>>>>> \date() "
>>>>> * or, even better, by skipping the whole \date function if the
>>>>> *hash-ref* function fails.
>>>>>
>>>>> Unfortunately all my efforts to set the failure-result have failed.
>>>>> I've tried e.g.
>>>>> \date{◊(pubdate->english (hash-ref metas 'doc-publish-date [] ))}
>>>>> in order to return an empty string, but racket chokes by telling me []
>>>>> is not a procedure.
>>>>>
>>>>> And I can't find reference to something resembling if-error in the
>>>>> racket documentation.
>>>>>
>>>>> Any pointers would be greatly appreciated!
>>>>>
>>>>> P.
>>>>>
>>>>>
>>>>>
>>>>>
>>>>
--
You received this message because you are subscribed to the Google Groups
"Pollen" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/d/optout.
