You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Original bug ID: 5607 Reporter:@gasche Assigned to:@Octachron Status: resolved (set by @Octachron on 2017-10-16T06:56:25Z) Resolution: fixed Priority: normal Severity: feature Target version: 4.07.0+dev/beta2/rc1/rc2 Fixed in version: 4.06.0 +dev/beta1/beta2/rc1 Category: documentation Tags: manual Related to:#6225 Monitored by:@hcarty
Bug description
Users, even familiar with the language, mostly do not not about variance annotations, and in particular their interaction with the relaxed value restriction. The problem they observe is that expressions at an abstract type representing a polymorphic data structure do not get generalized as expected.
I think that would be a good addition in the FAQ, in the spirit of "A function obtained through partial application is not polymorphic enough".
On this matter, I think that the relaxed value restriction and weakly polymorphic type should absolutely be described in the reference manual. It makes very little sense that the manual currently does not even mention the value restriction.
Original bug ID: 5607
Reporter: @gasche
Assigned to: @Octachron
Status: resolved (set by @Octachron on 2017-10-16T06:56:25Z)
Resolution: fixed
Priority: normal
Severity: feature
Target version: 4.07.0+dev/beta2/rc1/rc2
Fixed in version: 4.06.0 +dev/beta1/beta2/rc1
Category: documentation
Tags: manual
Related to: #6225
Monitored by: @hcarty
Bug description
Users, even familiar with the language, mostly do not not about variance annotations, and in particular their interaction with the relaxed value restriction. The problem they observe is that expressions at an abstract type representing a polymorphic data structure do not get generalized as expected.
I think that would be a good addition in the FAQ, in the spirit of "A function obtained through partial application is not polymorphic enough".
Additional information
This problem has been discussed twice recently:
on the caml-list
https://sympa-roc.inria.fr/wws/arc/caml-list/2012-05/msg00032.html
on reddit:
http://www.reddit.com/r/ocaml/comments/t8g2e/pdf_relaxing_the_value_restriction_by_jacques/c4kf68s
My reddit message might be a good basis for a FAQ answer; I hope to receive additional feedback on which wording would be best.
The text was updated successfully, but these errors were encountered: