value bug with wildcard?

ramagottfried commented 3 years ago

not sure what's going on here exactly -- any ideas? is it a bug? I think it might be.


----------begin_max5_patcher----------
1028.3oc6X0zaaiCD8r8uBAdpovqiHE0GbA5g8xdt2qKLnkncUWYJAJJutMH
+26PRIaoX65jTErMXCPhr3vY37Fx2Lbrua5Dzpx8hZj2e58IuISta5jIVQFA
SZGOAskuOsfWaUCUNOsbaUYs.MyMqVrWam412COuagzy610kkv63Ex6msPda
FWyOyTcKPQtTjV1HsqRTqvJknVH0bcdob4Y0nrQWHz5uUIbvGg79b6Txls4R
XRKhIsBWWJ0q4oVs867BWm9kb4lkJQp1sLwwQyYAggznj3vYdIAwy8m4giYl
ORfA88hCCV2faklm41lV80+fD2Eh07chrkqZjYEhkEB4F8Wr1PnmSAy9kCLA
.BXIf6wwlGFf3aeXj4imAAx0+iBJSA6ojCRnv+QmnX.HZT7HNx4QrOwrH31G
C.E0Lc3ol168PR2x.CXV7D4dw+0Pj38YyY68SmZdL6kI85r4PAWMGJ3kMGJA
ylSOlCEw5mCQYOobH1UxghRdKEZbHdh8UJf8kILJdB6yVD+Cd63EMh2s.XiK
P2ftDM5uaJJ9HO8eDZzrAiNO8B+KQuhihmCePnDC8J.eQ5E4T5UfOZbSRwuZ
RRiHgCRRSXO+jzneqRRY+e71A7EZ9pQU.ifDV7BzSsYr3qxQieY4nAQrAMiQ
5yQgCrmDIM7ZciQouwRub2XGVFGVnsv3jkAeXYFFlu0V2q815nXxnciQvas0
8aQaclqG9f2FglmkA7pZQ86t4v0Dcs6Yz5lyxIIWkSR9uoGwGvUiCY85QjRe
J8HRH86QzZmcS3A+rIVPXjO7zntrQk1EyceiFuitOSTqyk1MtdJA8kZU5rm3
OVO0sH+TOY9QJtH25QGSQOFOQFgXpaQ94dJbLhofmWL4HH7ppcBUcq1Vm.Il
esTYFZqsg.xsankygThc4c5aI1HtBX0ZfR2nr.CsOhhblBoyJYStMwXpI7.W
Vq+VwCIkR9VmoqJfjKuRo2+ZVRWtLpYqP1LbaX0l04EEokENnM4vLPQi1bWj
a1YshsivN25OOIhDSMEulmD6Glj3dKgFXJ0cLqq0LRmYDlOlYp6NOvmk3Scu
AhNqYNqv1ae5+3nZb4lBKXIwPShcRqTkUkptiPvCrC52nK2n3Y4BWUKeqb69
ZW8RqmWlKMGqhC3NfQvVbSBoLSQc3s9g6ziKioRyJgZ39sobV2gDvo9KUNu3
XEvmue6UsrN+6s1A0De.jp3JHhsLG6QaKwXaSgNuaenexyA9TJuxrO5YuIwY
TlXMGr6z3qqb8m7HWBX3QCX0P9hAXY46xgjj5qAtKu4OD5AWB5giNzykqK8T
hM+ZP+E4XWmqM0XFo8T7E.Lw+4AXWoWayOFHTW05IaORSue5O.rovElA
-----------end_max5_patcher-----------

ramagottfried commented 3 years ago

(I just edited the example slightly to make it clearer)

ramagottfried commented 3 years ago

ah! the plot thickens: /* is a bundle, but /1 is a string


----------begin_max5_patcher----------
887.3oc2V0zaaiCD8r8uBAdIsEtNhTeXoBjC6kdt2qKLnkncYWYRARprNaP9
uuyPJ4XmXi3j3Br6BXSINbFNONyaFp6GOhrTuUXIQeI56QiFc+3Qi7hPAi5m
Ohrguspga8pQzSqk11F9cjIgUchsN+JWSgw6mqhhtdkVCuSmqdXxb00clFX1
bPg4Dbt6tVAH3J6UC6QiTIpzcJ+Fk0KT24ZDNT4.9HjnezujpaiTAK5gDsW3
JsxshW40NtWVK2U8SoZ8BinxE1lTJaZZRVVZdwrrIQErjowShnYw3ihrow66
k.FNvMx5PbX4u9bRFAk8v3w3vjyNBVo2zpshyKBdzXTxv4yHrBki6jZ0hipw
4FEYutnXFkNEhd4Ek9v1LbLs70D7X4CmLK+VQ8hkcp5FwhFgZs6mnV4EGa8Z
tiGfPB3+xBH0QmgCXRL1OfxhoS.3+x+RAkSmgXemjT3e9yTLADcQ7HMO3QZL
C2DZ+vAfJEWN64lF8iKKe6S+mgu8jp12CuK4eU7N1+i4chssFf7UKPEeF4Cu
V3ln0BGutFnUVg8Ceb20C2DgOb701OfJ9wixJYuHqjcJV4W6ZZ9Fu5OENxjC
l863NlYYkXyRZVtmtldR5J6HzU192w3syGDdxk1dPfxOLgX0clpgyb+1Aozc
tuVXcRkOvsmRv8ZdkNZR+r8T9Y3oA3797TxaySgfIus8Vgw1qs2I.O9WZCN0
2Jf.DgvTe9gXD2JGz2SBHbCv.bP5uy3AFYadJIXJv9MpNomDMFOdfKst6ZdZ
BTw2DLcYCPDizpn+B2x.umzsQn5NLLrb8JYSSktI.sQ6VApw544jvpS5E6mQ
CtMdZQNaVJVqOsXVbVQQ3shzDryviLzdyXClwJiokXapoIwkEwog2.QG0rfU
T+WXs+vipwUqa7fkMC+BrdosFcq1LjBAOTtS+NmdsgWKEgJ7Xubebcn8h2yK
jJLsJ1g6jRF0iaVVZI1CDda+i63G2FrpbovbX7FK8GRR.m5OLRdyicKd69cu
NKV4e2aGz+3IPpkafSrm43Ss8DiMcMN4PbX+hmc7oJdKFGi7MdCFUKVwA6d9
4an012iXmBXzKFvrP8BBrZ4sRnHw9Rf6zA+Cgdxofd1EG5R0JcjQr98A8eKo
cmzg8XtPwT5I.LK9sA3PqW+2JfPv116I+mTL9gw+Cj80WUK
-----------end_max5_patcher-----------

maccallum commented 3 years ago

Ok, so this is extremely confusing, but not a bug. The input bundle is

/* : { /foo : 1}

and the program is

/url = getaddresses(),
/data = value(/url)

The first line of the program produces a binding of /url = "/*", so now the working bundle is

/* : { /foo : 1},
/url = "/*"

Then, when value(/url) is evaluated in the second line, the address (string) /url is compared to the pattern in the first message of the working bundle, which is /*, which produces a match, resulting in value( { /foo : 1 } ), which the value function just bails on.

There is a bug here--it's that value doesn't report an error. But in terms of pattern matching, it's unfortunately doing the right thing, even if it's confusing

ramagottfried commented 3 years ago

why does value do pattern matching with the wild card? I think only match should do pattern matching, otherwise this means you can't operate on wildcard addresses, which would be unfortunate.

ramagottfried commented 3 years ago

did you see the second issue?

<pre><code>
----------begin_max5_patcher----------
843.3oc2VszaaDBD9r8uBzdJox0YW12UpG5kdt2Sphv6hcHECq.1TmFk+6kA
Xsie07xpRsRV3kgY36aF9F18gwihlIWQ0QnOgtDMZzCiGMxYBLLJLeTzRxpF
NQ6bKRNsko63j6il3WsiXZtgIVbsh1X76TU7zrz77rhpx7InhX7z3Inj7X3u
zzownuGhctTXlSZnPXwAah9kLAmZbvkDLxZ8fO61OllGswSYuYWW8lL22Q8j
IJB8cXkGGOFFl7hyyF4xNolNflgtxkdQWjXGe3JABcwboDP9JwiCdwYBZirW
3bM83knjx5o1ZisDAEkpRXLq90TZv6WZvEu9RSXIM4NZ60y5Esb50bpXg4Fv
mhpCsdKwP76PpMGpqrGtIkv.bLG6F.awISrT+4+kYcNqDx+0VxfRydNlZMcR
PLoviXRLF1jjvvVjJCVNe+PO0poObBTSa2vchEUo+aKpv+GKpnq5TVkUKEbb
OkUuhi9LZA0PZaUTslpO67IWIt.NqrK.+YHKzmANd9Akb3WpjKOsBtNKIuvo
4xdm2wiwGVygOll6q8b92HM+fZhlr0rsJrNN4xvcdkmK+.6aWs0xdUy.DC8B
nMoVKUaXBhgIEO0Ir2oCdh9hQp3uFR3W.RvKc2AIewjz0cGUoCd6.wJRuUpf
ot97H6Qsep6zKRQuiM3uSODQTVwkwpr5UNhEspHKxGpUZqD8LmLYLjdVH0l6
46d.JHK8gNiaO2QRA5mvV5UQfHZFUscc.jkCQYSxunXD9latflnFIWptlIf7
K3V7TbZMFZzsOkmUCc51mJqxRg18cT8Z1uBwkfmtt34oT+RpneaF8twb1h4L
N2sEgcNr018NzpD4WcRvraVx.NUE3xL3xooUkw4UU9m1CmPX30zqNNo1Quz3
5p3L+SVSGLrPAw8EgOcXiaDwBtir3R3KFCV6TxNoZPVZQnds+8F4BEokQ8WZ
E6r6ToqK2cDkcUmxwUFBBik8bCaHlm17rVO0P5.LQtaU8A0RmSrwsubZ3VtK
Q3ioCR1UG7lIl11u.DqkcGy1jneNxcbs91TO8XTO+jSclXtDonKdeT+O0k8l
ongYf6XNQ0zjiPXb7aiv9qdceH.PAcW.I22KL9ww+FfRI.B3
-----------end_max5_patcher-----------
</code></pre>

ramagottfried commented 3 years ago

or is that similar, that typetags also does pattern matching? wow,... still really not sure that o.expr should do pattern matching without calling match

maccallum commented 3 years ago

@ramagottfried, I think you're right that pattern matching in o.expr should be explicit, rather than implicit. I just checked the history, and this behavior has been there forever, so this would be a breaking change, although, I'm not sure anyone relies on it. I'm going to add this to the 1.4 list--I think it's too late to consider something as deep as this for 1.3.

ramagottfried commented 3 years ago

@maccallum cool, that sounds good -- I'm fairly sure that not many people are using the hidden pattern matching functionality, since as far as I know there is no example of this in the documentation, I only stumbled on it when trying to operate on a wildcard address (that was then being sent to another process where the matching happens).

CNMAT / CNMAT-odot

value bug with wildcard? #421