Squeak does not have an implementation of Integer>>#asByteArray about seaside HOT 9 CLOSED

We could actually just copy the implementation from Pharo, which uses primitive 62 to compute the number of bytes in large integers.

I suppose I would implement #greaseAsByteArray on Integer and subclasses, as required, where the Pharo version would simply dispatch to Integer>>#asByteArray, right @jbrichau?

from seaside.

@theseion hm... I think we should do that yes. I had the same issue in GemStone (see 95d0f2e) but now I do not understand entirely because Integer>>asByteArray still does not exist in GemStone. I guess there is some other package being loaded that adds it... so there is an issue here for both Squeak and GemStone currently, which makes Grease the best place to tackle the difference.

from seaside.

Well... not the same issue but I should have the same issue actually after that commit... which is strange.

from seaside.

There may be more to this than initially meets the eye. Squeak really should have #asByteAray for integers since it is sent in Magnitude>>#putOn: which gets used in writing values to assorted streams. Some of which can be set to binary (file streams, for example) andwhich would fail with any sort of number.

from seaside.

It turns out that #asByteArray is really rather trivial for LargePositiveInteger -
myLargePositiveInteger as: ByteArray
will do it.
Hacking that trivially into #entityTagFor: to test makes all the nice formatting come back.

from seaside.

Yes, it's weird that that method isn't there. I looked back as far as I could (5.2 only, unfortunately) and it's not there. There's a PR in the works that adds Grease support for all platforms (GemStone has the same issue).

from seaside.

I've added a suitable #asByteArray for squeak 6.1 trunk.

from seaside.

That's great, thanks. Did you use the implementation from Pharo (which used for Seaside)?

from seaside.

from seaside.