On 17/07/2015 10:59, Alexander Grahn wrote:
>>    \bool_if:nTF
>>      {
>>           \cs_if_exist_p:N  \pdftex_pdfoutput:D
>>        && \int_compare_p:nNn \pdftex_pdfoutput:D > \c_zero
>>      }
>>      { pdfmode }
>>      { dvips }
>>
>> will fail as the second test will not be skipped.  We note though that
>> this case is likely unusual: in general expl3 variables should be
>> declared before use so should. A quick sweep over TL2015 shows the
>> following packages would have to be updated following any change:
>>
>> - media9
>>  -ocgx2
>>
>> (Other packages do use \cs_if_exist_p:(N|c) but not in a way which
>> relies on lazy evaluation.)
>>
>> Are there other obvious issues if we switch from lazy evaluation? If the
>> change is to be made we intend to work with package authors in advance
>> to make sure that users are not impacted.
> 
> thank you  for pointing out this. Could you please give me specific
> advice on the changes I will have to do?
> 
> Kind regards,
> Alexander

Hi Alexander,

As I said, at present we've not changed anything precisely because we
are not sure whether dropping lazy evaluation is acceptable. (It's a
tricky balance: having an issue with ")" tokens is also not ideal.) The
reason I've raised on LaTeX-L is we want to be sure the proposed change
doesn't cause significant issues either for end-users or for programmers.

Assuming we do make a change and evaluate all parts of an expression
then doing

    \bool_if:nT
      {
            \cs_if_exist_p:N <some-macro>
        &&  \some_test:N <some-macro>
      }

won't work if the second test only works if <some-macro> is defined. In
your code you've got three places I can see where \cs_if_exist_p:N or
\cs_if_exist_p:c is used as the first test in an expression, so the
second one will fail. The options are then to rewrite as a set of nested
TF versions

    \cs_if_exist:NTF  \foo
      { }
      { }

or to use a bit of imagination to avoid that. Assuming the latter is
wanted, in media9 you could replace

    \bool_if:nTF
      {
           !\cs_if_exist_p:c {#1}
        || !\str_if_eq_x_p:nn {\tl_use:c{#1} } {#2}
      }

by

    \bool_if:nTF
      {
        !\cs_if_exist:cTF {#1}
          {
            \str_if_eq_x_p:nn { \tl_use:c {#1} } {#2}
          }
          { \c_false_bool }
      }

and replace

    \bool_if:nT
      {
             \cs_if_exist_p:N \fxocg@insert@OC
         && !\tl_if_blank_p:V \fxocg@insert@OC
      }

by

    \bool_if:nT
      {
        ! \cs_if_exist:NTF \fxocg@insert@OC
          { \tl_if_blank_p:V \fxocg@insert@OC }
          { \c_true_bool }
      }

In ocgx2.sty there is

    \bool_if:nT
      {
           !\cs_if_exist_p:c { OCGpdfobj#3 }
        || !\str_if_eq_x_p:nn { \use:c { OCGpdfobj#3 } } { \use:c {
ocgxii@#3 } }
      }

which could be redone as for the first case in media9. However,
reviewing the code I realise that as you have \use:c here you'll
actually get away with it (the second part will evaluate even if the
csname is undefined). If you do want a 'safe' change then

    \bool_if:nTF
      {
        ! \cs_if_exist:cTF { OCGpdfobj#3 }
          {
            \str_if_eq_x_p:nn { \use:c { OCGpdfobj#3 } } { \use:c {
ocgxii@#3 } }
          }
          { \c_false_bool }
      }

(BTW, I assume ocgxii@#3 is known to be defined!)

I hope this makes sense.
--
Joseph Wright