I think no `<= lfun exists. There is a manual page at

http://pike.lysator.liu.se/generated/manual/modref/ex/lfun_3A_3A.html

which lists all of them.

On 08/27/14 09:47, Stephen R. van den Berg wrote:

It appears that when Pike is presented with a comparison like:

a<=b

And given the availability of `< and `<= and `== operators, it insists on using `< first, and if that fails it uses `==. Instead of using merely `<= directly.

Am I missing something here or is that a missed optimisation?

Its not the same list; one describes the functions which implement the operators and the other ones are lfuns. They are related in that the ones from predef:: will call the corresponding lfun. In the documentation you linked, a cross reference exists.

Maybe the overall structure is not so clear (i.e. predef::`[] vs `[]) and the documentation should be improved?

On 08/27/14 19:40, Stephen R. van den Berg wrote:

Arne Goedeke wrote:

...

I think no `<= lfun exists. There is a manual page at

...

http://pike.lysator.liu.se/generated/manual/modref/ex/lfun_3A_3A.html

...

which lists all of them.

Ok, fair enough. It is confusing though that this list is present twice. There undoubtedly is a good reason for that, but it confusing to the reader and there should be crossreferences from one to the other or similar.

E.g. compare: http://pike.lysator.liu.se/generated/manual/modref/ex/predef_3A_3A/_backtick... http://pike.lysator.liu.se/generated/manual/modref/ex/lfun_3A_3A/_backtick_5...

According to http://pike.lysator.liu.se/generated/manual/modref/ex/lfun_3A_3A.html, there doesn't exist any `<= lfun, so then it's hardly surpising it isn't used...

I would be ok with removing the `> lfun to reduce the API.