Demo broken #35
Comments
|
Interesting. I think I know what is causing the name issue. I'm not sure what causes it to be rejected, though! That will take some more debugging. |
|
I went back and fixed several bugs. Can you retry the capture step and tell me if that works? (If it doesn't, please post the captured file.) If it does work, can you retry the "accept" call? (If it doesn't, please post the captured file.) |
|
This is what the captured file now (57faca2...) looks like: With that file as input, $ racket src/run.rkt -d accept bench/example.rkt example.com input: Could not find :script example.com If I comment-out the offending script lookup: $ git diff src/input.rkt
diff --git a/src/input.rkt b/src/input.rkt
index 04dde07..ff74001 100644
--- a/src/input.rkt
+++ b/src/input.rkt
@@ -50,7 +50,7 @@
(get-from ':sheets sheets)
(get-from ':layouts layouts)
- (get-from ':script scripts)
+ ; (get-from ':script scripts)
(set! properties (dict-set properties ':documents (dict-ref properties ':layouts)))
(set! properties
(dict-set properties ':fonts
I get: $ racket src/run.rkt -d accept bench/example.rkt example.com
[ 0.000s] Read 1 documents with 13 elements, 18 boxes, 114 rules, and 3 fonts
[ 0.733s] Produced 2152 constraints of 25875 terms
[ 1.758s] Prepared 4714 constraints of 80483 terms
[ 2.528s] Found core with 5 constraints
([VIEW :w 1280]
([BLOCK :x 0 :y 0 :w 1280 :h 6193/15 :elt 0]
([BLOCK :x 0 :y 80 :w 1280 :h 3793/15 :elt 7]
([BLOCK :x 308 :y 80 :w 664 :h 3793/15 :elt 8]
([BLOCK :x 340 :y 4003/30 :w 600 :h 38 :elt 9]
([LINE]
([TEXT :x 340 :y 4003/30 :w 1507/5 :h 38])))
([BLOCK :x 340 :y 2893/15 :w 600 :h 57 :elt 10]
([LINE]
([TEXT :x 340 :y 2893/15 :w 2962/5 :h 19]))
([LINE]
([TEXT :x 340 :y 3178/15 :w 5113/10 :h 19]))
([LINE]
([TEXT :x 340 :y 3463/15 :w 1849/20 :h 19])))
([BLOCK :x 340 :y 3988/15 :w 600 :h 19 :elt 11]
([LINE]
([INLINE :x 340 :y 3988/15 :w 9139/60 :h 19 :elt 12]
([TEXT :x 340 :y 3988/15 :w 9139/60 :h 19]))))))))
Rejected.
where the bold-white lengths are printed in red on my terminal. Oddly, if I leave out the This is with z3 4.8.7 but there are newer versions available eg 4.8.17 and 4.10.2. I know it's somewhat of a hit and miss, so I'm asking should I test with another specific z3 version? |
|
Very odd. Here's what I get: So, an "accept" without the It's pretty hit-or-miss indeed, because Cassius talks to Z3 over the command-line interface, and that changes frequently. So perhaps upgrading will help. |
|
I can confirm I get the exact same results as you now with z3 4.9.1. Whereas with z3 4.8.10 (which is what eg current Ubuntu LTS has), the invocation of z3 from cassius/racket appears to hang and not terminate, though oddly not with high CPU load. I still need to use the version of |
|
I think the z3 issue has low CPU load because the issue is a change in Z3's text protocol—probably Cassius is waiting for Z3 to say something but Z3 doesn't think there's more to say. It's probably fixable pretty easily. I'm not sure why the Both of these issues I'd be happy to merge fixes for, but I'm unlikely to find the time to debug, especially since they don't happen on my machine. (Cassius is no longer actively developed.) |
Attempting to capture example.com as advised ie running
results in a
bench/example.rktfile looking like this for me:ie. tokens after
define-stylesheet,define-browser, etc. are all missing, resulting in error messages when attempting to run Cassius command line tools on it eg:Manually editing the file and adding
doc-1,firefox, fixingdefine-problem, etc. such that it readswill result in a succesfull (albeit rejected)
acceptrun:Ubuntu version 20.04.4
Firefox version 103 (for Canonical)
geckodriver version 0.31.0
Python version 3.8.10
Selenium version 4.0.0a1
Racket version 7.2
Greets
The text was updated successfully, but these errors were encountered: