0005080: type application in quotations

(0005605)
jaked (reporter)
2010-07-13 19:26

I have the same problem with << $x$ -> unit >>; I get" EOI expected after [quotation of type] (in [quotation of type])". It seems to happen with leading antiquotations.

(0005606)
jaked (reporter)
2010-07-13 19:35

Maybe not the same bug, but as it is for <:ctyp< >> I'll put it here: for <:ctyp< $lid:id$ : mutable $t$ >> I get "illegal begin of quotation of type".

(0005634)
ygrek (reporter)
2010-08-03 12:22

This affects deriving as well. First two cases can be workarounded with extra parentheses :

- | `Function (f, t) -> <:ctyp< $expr f$ -> $expr t$ >>
+ | `Function (f, t) -> <:ctyp< ($expr f$ -> $expr t$) >>

- | [x] -> <:ctyp< $expr x$ $f$ >>
- | x::xs -> app (<:ctyp< $expr x$ $f$ >>) xs
+ | [x] -> <:ctyp< ($expr x$ $f$) >>
+ | x::xs -> app (<:ctyp< ($expr x$ $f$) >>) xs

(0005804)
xclerc (developer)
2011-02-15 11:25

The patch proposed for 0005104 seems to also solve this issue.

(0005822)
xclerc (developer)
2011-02-17 12:09

Nicolas Pouillard provided me with a cleaner patch that has
been committed in "version/3.12". After some testing, it
seems to both resolve the issue and yield no regression.

Please test it, and re-open the issue if I missed something.

(0005959)
doligez (manager)
2011-06-02 01:15

The current version (3.12.1+dev11) breaks lablgtk with the following error message.

File "gtkSignal.ml4", line 92, characters 10-13:
Parse error: [opt_when_expr] expected after [patt_as_patt_opt] (in [match_case0])

This happens with and without the patch that you (xclerc) provided to me in private.

(0005960)
xclerc (developer)
2011-06-02 12:00
edited on: 2011-06-02 12:09

As a first step, I reduce the bug case to:
"let f = function | `INT x -> true | _ -> false"

camlp4of no more accepts polymorphic variants with a parameter.

Investigation continues...

(0005961)
xclerc (developer)
2011-06-02 12:23

OK, my bad.
The problem stems from the dirty hack I set for PR#4732.

Issue History
Date Modified	Username	Field	Change
2010-06-20 05:23	glondu	New Issue
2010-06-20 05:23	glondu	Status	new => assigned
2010-06-20 05:23	glondu	Assigned To	=> ertai
2010-07-13 19:26	jaked	Note Added: 0005605
2010-07-13 19:35	jaked	Note Added: 0005606
2010-08-03 12:22	ygrek	Note Added: 0005634
2011-02-15 11:24	xclerc	Relationship added	related to 0005104
2011-02-15 11:24	xclerc	Assigned To	ertai => xclerc
2011-02-15 11:25	xclerc	Note Added: 0005804
2011-02-15 11:25	xclerc	Status	assigned => feedback
2011-02-17 12:09	xclerc	Note Added: 0005822
2011-02-17 12:09	xclerc	Status	feedback => resolved
2011-02-17 12:09	xclerc	Fixed in Version	=> 3.12.1+dev
2011-02-17 12:09	xclerc	Resolution	open => fixed
2011-06-02 01:15	doligez	Note Added: 0005959
2011-06-02 01:16	doligez	Status	resolved => assigned
2011-06-02 12:00	xclerc	Note Added: 0005960
2011-06-02 12:09	xclerc	Note Edited: 0005960
2011-06-02 12:23	xclerc	Note Added: 0005961
2012-05-31 16:55	xclerc	Status	assigned => resolved
2012-05-31 16:55	xclerc	Fixed in Version	3.12.1+dev => 4.00.0+dev

Notes
(0005605) jaked (reporter) 2010-07-13 19:26	I have the same problem with << $x$ -> unit >>; I get" EOI expected after [quotation of type] (in [quotation of type])". It seems to happen with leading antiquotations.

(0005606) jaked (reporter) 2010-07-13 19:35	Maybe not the same bug, but as it is for <:ctyp< >> I'll put it here: for <:ctyp< $lid:id$ : mutable $t$ >> I get "illegal begin of quotation of type".

(0005634) ygrek (reporter) 2010-08-03 12:22	This affects deriving as well. First two cases can be workarounded with extra parentheses : - \| `Function (f, t) -> <:ctyp< $expr f$ -> $expr t$ >> + \| `Function (f, t) -> <:ctyp< ($expr f$ -> $expr t$) >> - \| [x] -> <:ctyp< $expr x$ $f$ >> - \| x::xs -> app (<:ctyp< $expr x$ $f$ >>) xs + \| [x] -> <:ctyp< ($expr x$ $f$) >> + \| x::xs -> app (<:ctyp< ($expr x$ $f$) >>) xs

(0005804) xclerc (developer) 2011-02-15 11:25	The patch proposed for 0005104 seems to also solve this issue.

(0005822) xclerc (developer) 2011-02-17 12:09	Nicolas Pouillard provided me with a cleaner patch that has been committed in "version/3.12". After some testing, it seems to both resolve the issue and yield no regression. Please test it, and re-open the issue if I missed something.

(0005959) doligez (manager) 2011-06-02 01:15	The current version (3.12.1+dev11) breaks lablgtk with the following error message. File "gtkSignal.ml4", line 92, characters 10-13: Parse error: [opt_when_expr] expected after [patt_as_patt_opt] (in [match_case0]) This happens with and without the patch that you (xclerc) provided to me in private.

(0005960) xclerc (developer) 2011-06-02 12:00 edited on: 2011-06-02 12:09	As a first step, I reduce the bug case to: "let f = function \| `INT x -> true \| _ -> false" camlp4of no more accepts polymorphic variants with a parameter. Investigation continues...

(0005961) xclerc (developer) 2011-06-02 12:23	OK, my bad. The problem stems from the dirty hack I set for PR#4732.