Packaging Next browser (Common Lisp) [work in progress]

Done

Details

3 participants

Andy Patterson
Pierre Neidhardt
Pierre Neidhardt

Owner: unassigned

Submitted by: Pierre Neidhardt

Severity: normal

Debbugs page

Pierre Neidhardt wrote on 19 Sep 2018 00:15

Recipients:(address . bug-guix@gnu.org)

Message-ID:87o9cuez56.fsf@ambrevar.xyz

I've started working on packaging a collection of Common Lisp

dependencies for Next browser:

http://next.atlas.engineer/

My progress can be found on the wip-next-browser branch.

Pierre Neidhardt

https://ambrevar.xyz/

Pierre Neidhardt wrote on 10 Oct 2018 06:21

Recipients:(address . 32769@debbugs.gnu.org)(name . Andy Patterson)(address . ajpatter@uwaterloo.ca)

Message-ID:87va6a7x9l.fsf@ambrevar.xyz

The good news: all Next browser dependencies are now packaged under the

branch wip-nextbrowser.

The bad news: Next browser fails to build.

Toggle snippet (117 lines); compiling file "/gnu/store/39vh5ghdn9x7zwb6ky7law678wkz20l2-next-browser-0.0.8-1.ccc289b/share/common-lisp/sbcl-source/next/gtk/next/source/document-mode.lisp" (written 10 OCT 2018 01:12:05 PM):
; compiling (IN-PACKAGE :NEXT)
; compiling (DEFVAR *DOCUMENT-MODE-MAP* ...)
; compiling (DEFCLASS DOCUMENT-MODE ...)
; compiling (DEFPARENSTATIC SCROLL-TO-TOP ...)
; compiling (DEFPARENSTATIC SCROLL-TO-BOTTOM ...)
; compiling (DEFPARENSTATIC SCROLL-DOWN ...)
; compiling (DEFPARENSTATIC SCROLL-UP ...)
; compiling (DEFPARENSTATIC SCROLL-LEFT ...)
; compiling (DEFPARENSTATIC SCROLL-RIGHT ...)
; compiling (DEFUN ENSURE-ZOOM-RATIO-RANGE ...)
; compiling (DEFPAREN %ZOOM-IN-PAGE ...)
; compiling (DEFINE-COMMAND ZOOM-IN-PAGE ...)
; compiling (DEFPAREN %ZOOM-OUT-PAGE ...)
; compiling (DEFINE-COMMAND ZOOM-OUT-PAGE ...)
; compiling (DEFPAREN %UNZOOM-PAGE ...)
; compiling (DEFINE-COMMAND UNZOOM-PAGE ...)
; compiling (DEFINE-COMMAND HISTORY-BACKWARDS ...)
; compiling (DEFINE-COMMAND HISTORY-FORWARDS ...)
; compiling (DEFUN HISTORY-FOWARDS-QUERY-COMPLETE ...)
; compiling (DEFINE-COMMAND HISTORY-FORWARDS-QUERY ...)
; compiling (DEFUN ADD-OR-TRAVERSE-HISTORY ...)
; compiling (DEFINE-COMMAND SET-URL-NEW-BUFFER ...)
; compiling (DEFUN SET-URL-BUFFER ...)
; compiling (DEFUN SET-URL-NEW-BUFFER ...)
; file: /gnu/store/39vh5ghdn9x7zwb6ky7law678wkz20l2-next-browser-0.0.8-1.ccc289b/share/common-lisp/sbcl-source/next/gtk/next/source/document-mode.lisp
; in: DEFUN SET-URL-NEW-BUFFER
;     (DEFUN NEXT::SET-URL-NEW-BUFFER
;            (NEXT::INPUT-URL &OPTIONAL NEXT::DISABLE-HISTORY)
;       (LET ((NEXT::NEW-BUFFER (NEXT::GENERATE-NEW-BUFFER "default" #)))
;         (NEXT::SET-VISIBLE-ACTIVE-BUFFER NEXT::NEW-BUFFER)
;         (NEXT::SET-URL-BUFFER NEXT::INPUT-URL NEXT::NEW-BUFFER
;          NEXT::DISABLE-HISTORY)))
; --> PROGN
; ==>
;   (EVAL-WHEN (:COMPILE-TOPLEVEL)
;     (SB-C:%COMPILER-DEFUN 'NEXT::SET-URL-NEW-BUFFER NIL T))
;
; caught WARNING:
;   Duplicate definition for SET-URL-NEW-BUFFER found in one file.
;   See also:
;     The ANSI Standard, Section 3.2.2.3

; compiling (DEFUN SETUP-URL ...)
; compiling (DEFUN SET-URL ...)
; compiling (DEFINE-COMMAND SET-URL-CURRENT-BUFFER ...)
; compiling (DEFINE-COMMAND SET-URL-FROM-BOOKMARK ...)
; compiling (DEFUN SETUP-ANCHOR ...)
; compiling (DEFINE-COMMAND GO-ANCHOR ...)
; compiling (DEFINE-COMMAND GO-ANCHOR-NEW-BUFFER ...)
; compiling (DEFINE-COMMAND GO-ANCHOR-NEW-BUFFER-FOCUS ...)
; compiling (DEFUN DOCUMENT-MODE ...)
; compiling (DEFMETHOD SETUP ...)

; /gnu/store/39vh5ghdn9x7zwb6ky7law678wkz20l2-next-browser-0.0.8-1.ccc289b/lib/sbcl/next/source/document-mode-tmpOPCILR65.fasl written
; compilation finished in 0:00:00.058
Unhandled UIOP/LISP-BUILD:COMPILE-FILE-ERROR in thread #<SB-THREAD:THREAD "main thread" RUNNING
                                                          {10005585B3}>:
  COMPILE-FILE-ERROR while compiling #<CL-SOURCE-FILE "next" "document-mode">

Backtrace for: #<SB-THREAD:THREAD "main thread" RUNNING {10005585B3}>
0: (SB-DEBUG::DEBUGGER-DISABLED-HOOK #<UIOP/LISP-BUILD:COMPILE-FILE-ERROR {1004DDACA3}> #<unused argument> :QUIT T)
1: (SB-DEBUG::RUN-HOOK *INVOKE-DEBUGGER-HOOK* #<UIOP/LISP-BUILD:COMPILE-FILE-ERROR {1004DDACA3}>)
2: (INVOKE-DEBUGGER #<UIOP/LISP-BUILD:COMPILE-FILE-ERROR {1004DDACA3}>)
3: (ERROR UIOP/LISP-BUILD:COMPILE-FILE-ERROR :CONTEXT-FORMAT "~/asdf-action::format-action/" :CONTEXT-ARGUMENTS ((#<ASDF/LISP-ACTION:COMPILE-OP > . #<ASDF/LISP-ACTION:CL-SOURCE-FILE "next" "document-mode">)))
4: (UIOP/LISP-BUILD:CHECK-LISP-COMPILE-RESULTS NIL T T "~/asdf-action::format-action/" ((#<ASDF/LISP-ACTION:COMPILE-OP > . #<ASDF/LISP-ACTION:CL-SOURCE-FILE "next" "document-mode">)))
5: ((SB-PCL::EMF ASDF/ACTION:PERFORM) #<unused argument> #<unused argument> #<ASDF/LISP-ACTION:COMPILE-OP > #<ASDF/LISP-ACTION:CL-SOURCE-FILE "next" "document-mode">)
6: ((LAMBDA NIL :IN ASDF/ACTION:CALL-WHILE-VISITING-ACTION))
7: ((:METHOD ASDF/ACTION:PERFORM-WITH-RESTARTS :AROUND (T T)) #<ASDF/LISP-ACTION:COMPILE-OP > #<ASDF/LISP-ACTION:CL-SOURCE-FILE "next" "document-mode">) [fast-method]
8: ((:METHOD ASDF/PLAN:PERFORM-PLAN (T)) #<ASDF/PLAN:SEQUENTIAL-PLAN {1002497DE3}>) [fast-method]
9: ((FLET SB-C::WITH-IT :IN SB-C::%WITH-COMPILATION-UNIT))
10: ((:METHOD ASDF/PLAN:PERFORM-PLAN :AROUND (T)) #<ASDF/PLAN:SEQUENTIAL-PLAN {1002497DE3}>) [fast-method]
11: ((:METHOD ASDF/OPERATE:OPERATE (ASDF/OPERATION:OPERATION ASDF/COMPONENT:COMPONENT)) #<ASDF/BUNDLE:COMPILE-BUNDLE-OP > #<ASDF/SYSTEM:SYSTEM "next/gtk"> :PLAN-CLASS NIL :PLAN-OPTIONS NIL) [fast-method]
12: ((SB-PCL::EMF ASDF/OPERATE:OPERATE) #<unused argument> #<unused argument> #<ASDF/BUNDLE:COMPILE-BUNDLE-OP > #<ASDF/SYSTEM:SYSTEM "next/gtk">)
13: ((LAMBDA NIL :IN ASDF/OPERATE:OPERATE))
14: ((:METHOD ASDF/OPERATE:OPERATE :AROUND (T T)) #<ASDF/BUNDLE:COMPILE-BUNDLE-OP > #<ASDF/SYSTEM:SYSTEM "next/gtk">) [fast-method]
15: ((SB-PCL::EMF ASDF/OPERATE:OPERATE) #<unused argument> #<unused argument> ASDF/BUNDLE:COMPILE-BUNDLE-OP "next/gtk")
16: ((LAMBDA NIL :IN ASDF/OPERATE:OPERATE))
17: ((:METHOD ASDF/OPERATE:OPERATE :AROUND (T T)) ASDF/BUNDLE:COMPILE-BUNDLE-OP "next/gtk") [fast-method]
18: (ASDF/SESSION:CALL-WITH-ASDF-SESSION #<CLOSURE (LAMBDA NIL :IN ASDF/OPERATE:OPERATE) {1002351FBB}> :OVERRIDE T :KEY NIL :OVERRIDE-CACHE T :OVERRIDE-FORCING NIL)
19: ((LAMBDA NIL :IN ASDF/OPERATE:OPERATE))
20: (ASDF/SESSION:CALL-WITH-ASDF-SESSION #<CLOSURE (LAMBDA NIL :IN ASDF/OPERATE:OPERATE) {1004329D1B}> :OVERRIDE NIL :KEY NIL :OVERRIDE-CACHE NIL :OVERRIDE-FORCING NIL)
21: ((:METHOD ASDF/OPERATE:OPERATE :AROUND (T T)) ASDF/BUNDLE:COMPILE-BUNDLE-OP "next/gtk") [fast-method]
22: (SB-INT:SIMPLE-EVAL-IN-LEXENV (ASDF/OPERATE:OPERATE (QUOTE ASDF/BUNDLE:COMPILE-BUNDLE-OP) "next/gtk") #<NULL-LEXENV>)
23: (EVAL (ASDF/OPERATE:OPERATE (QUOTE ASDF/BUNDLE:COMPILE-BUNDLE-OP) "next/gtk"))
24: (SB-IMPL::PROCESS-EVAL/LOAD-OPTIONS ((:EVAL . "(require :asdf)") (:EVAL . "(let ((*package* (find-package :asdf))) (load \"/gnu/store/39vh5ghdn9x7zwb6ky7law678wkz20l2-next-browser-0.0.8-1.ccc289b/share/common-lisp/sbcl-source/next/gtk/next/next.asd\"))") (:EVAL . "(asdf:operate (quote asdf:compile-bundle-op) \"next/gtk\")") (:QUIT)))
25: (SB-IMPL::TOPLEVEL-INIT)
26: ((FLET SB-UNIX::BODY :IN SAVE-LISP-AND-DIE))
27: ((FLET "WITHOUT-INTERRUPTS-BODY-27" :IN SAVE-LISP-AND-DIE))
28: ((LABELS SB-IMPL::RESTART-LISP :IN SAVE-LISP-AND-DIE))

unhandled condition in --disable-debugger mode, quitting
;
; compilation unit aborted
;   caught 1 fatal ERROR condition
;   caught 1 WARNING condition
;   caught 8 STYLE-WARNING conditions
Backtrace:
           7 (primitive-load "/gnu/store/6v5dy1rnk8m3h154q4c8mjwgq7y…")
In ice-9/eval.scm:
   191:35  6 (_ _)
In ice-9/boot-9.scm:
    152:2  5 (with-fluid* _ _ _)
    152:2  4 (with-fluid* _ _ _)
In srfi/srfi-1.scm:
    640:9  3 (for-each #<procedure 98e3a0 at /gnu/store/k4nbgbgygrv…> …)
In /gnu/store/k4nbgbgygrv0lk97d5phx8fwrxr0ah9r-module-import/guix/build/gnu-build-system.scm:
   799:31  2 (_ _)
In /gnu/store/k4nbgbgygrv0lk97d5phx8fwrxr0ah9r-module-import/guix/build/asdf-build-system.scm:
    124:4  1 (build #:outputs _ #:inputs _ #:asd-file _ # _)
In unknown file:
           0 (scm-error misc-error #f "~A ~S ~S" ("lisp-eval-pro…" …) …)

ERROR: In procedure scm-error:
lisp-eval-program failed! "/gnu/store/rdvy6q3nskb0r01rwjrxj3935d850r2d-sbcl-1.4.4/bin/sbcl" ((require :asdf) (let ((*package* (find-package :asdf))) (load "/gnu/store/39vh5ghdn9x7zwb6ky7law678wkz20l2-next-browser-0.0.8-1.ccc289b/share/common-lisp/sbcl-source/next/gtk/next/next.asd")) (asdf:operate (quote asdf:compile-bundle-op) "next/gtk"))

Andy, any clue on this?

Pierre Neidhardt

https://ambrevar.xyz/

Andy Patterson wrote on 11 Oct 2018 00:16

Recipients:(name . Pierre Neidhardt)(address . mail@ambrevar.xyz)(address . 32769@debbugs.gnu.org)

Message-ID:20181011031649.5863f9a8@mailservices.uwaterloo.ca

Hi,

On Wed, 10 Oct 2018 15:21:26 +0200

Pierre Neidhardt <mail@ambrevar.xyz> wrote:

Toggle quote (9 lines)

> The good news: all Next browser dependencies are now packaged under

> the branch wip-nextbrowser.

> The bad news: Next browser fails to build.

> [...]

> Andy, any clue on this?

It's probably the

Toggle quote (3 lines)

> ; caught WARNING:

> ; Duplicate definition for SET-URL-NEW-BUFFER found in one file.

since sbcl (at least on guix?) treats some warnings as errors for some

reason.

Hoping that helps,

Andy

Pierre Neidhardt wrote on 11 Oct 2018 01:50

Recipients:(name . Andy Patterson)(address . ajpatter@uwaterloo.ca)(address . 32769@debbugs.gnu.org)

Message-ID:87va68uatd.fsf@ambrevar.xyz

I'll try this. Note that Next builds fine via Quicklisp. Any idea why?

Pierre Neidhardt

https://ambrevar.xyz/

Pierre Neidhardt wrote on 11 Oct 2018 03:44

Recipients:(name . Andy Patterson)(address . ajpatter@uwaterloo.ca)(address . 32769@debbugs.gnu.org)

Message-ID:87sh1cu5ic.fsf@ambrevar.xyz

You were right, this was the issue. I've posted a patch upstream, it should be

fixed soon.

Now Next compiles. Hurray!!! :D

There is one last step though: the .asd bunble works, but not the

"build-program".

See the wip-nextbrowser for what I've tried. I've re-used the code from

Stumpwm.

I narrowed down the issue to the "generate-executable" function:

Toggle snippet (19 lines)define* (generate-executable out-file #:key
                              dependencies
                              dependency-prefixes
                              entry-program
                              type
                              #:allow-other-keys)
  "Generate an executable by using asdf operation TYPE, containing whithin the
image all DEPENDENCIES, and running ENTRY-PROGRAM in the case of an
executable.  Link in any asd files from DEPENDENCY-PREFIXES to ensure
references to those libraries are retained."
  let* ((bin-directory (dirname out-file))
         (name (basename out-file)))
    (mkdir-p bin-directory)
    (with-directory-excursion bin-directory
      (generate-executable-wrapper-system name dependencies)
      (generate-executable-entry-point name entry-program))
			...

The problem is that Next's system is called "next/gtk", but we cannot use that

as an out-file. If I use something else, say "next", then compilation fails

with

Toggle snippet (3 lines)

Component :NEXT not found, required by #<SYSTEM "next-exec">

I guess there is a flaw in the logic here. The system name should be

#:asd-system-name

and not

(name (basename out-file))

What shall we do?

Pierre Neidhardt

https://ambrevar.xyz/

Andy Patterson wrote on 11 Oct 2018 22:51

Recipients:(name . Pierre Neidhardt)(address . mail@ambrevar.xyz)(address . 32769@debbugs.gnu.org)

Message-ID:20181012015120.686812a8@mailservices.uwaterloo.ca

Hi Pierre,

On Thu, 11 Oct 2018 12:44:43 +0200

Pierre Neidhardt <mail@ambrevar.xyz> wrote:

Toggle quote (5 lines)

> You were right, this was the issue. I've posted a patch upstream, it

> should be fixed soon.

> Now Next compiles. Hurray!!! :D

Nice.

Toggle quote (29 lines)> 
> There is one last step though: the .asd bunble works, but not the
> "build-program".
> See the wip-nextbrowser for what I've tried.  I've re-used the code
> from Stumpwm.
> 
> I narrowed down the issue to the "generate-executable" function:
> 
> [...]
> 
> The problem is that Next's system is called "next/gtk", but we cannot
> use that as an out-file.  If I use something else, say "next", then
> compilation fails with
> 
> --8<---------------cut here---------------start------------->8---
>   Component :NEXT not found, required by #<SYSTEM "next-exec">
> --8<---------------cut here---------------end--------------->8---
> 
> I guess there is a flaw in the logic here.  The system name should be
> 
> 	#:asd-system-name
> 
> and not
> 
> 	(name (basename out-file))
> 
> What shall we do?
> 

That seems right. I guess we'll need to update that and then maybe

also normalize the generated asd. Doesn't seem like it'd be too

difficult; there just weren't enough examples of program-generating

systems for this to get caught.

Thanks for the update,

Andy

Pierre Neidhardt wrote on 11 Oct 2018 23:44

Recipients:(name . Andy Patterson)(address . ajpatter@uwaterloo.ca)(address . 32769@debbugs.gnu.org)

Message-ID:87ftxbslyk.fsf@ambrevar.xyz

Let me know if you'd like to do it or if I should give it a shot :)

Pierre Neidhardt

https://ambrevar.xyz/

Andy Patterson wrote on 12 Oct 2018 15:25

Recipients:(name . Pierre Neidhardt)(address . mail@ambrevar.xyz)(address . 32769@debbugs.gnu.org)

Message-ID:20181012182510.693c28f6@mailservices.uwaterloo.ca

Hi,

On Fri, 12 Oct 2018 08:44:35 +0200

Pierre Neidhardt <mail@ambrevar.xyz> wrote:

Toggle quote (3 lines)

> Let me know if you'd like to do it or if I should give it a shot :)

Please feel free. I'll probably get around to it this weekend if you

don't have time before then, if that's ok with you.

Thanks,

Andy

Pierre Neidhardt wrote on 14 Oct 2018 01:27

Recipients:(name . Andy Patterson)(address . ajpatter@uwaterloo.ca)(address . 32769@debbugs.gnu.org)

Message-ID:878t31kk59.fsf@ambrevar.xyz

I won't have time this weekend, so you can go ahead. Otherwise I'll do it next

week.

Cheers!

Pierre Neidhardt

https://ambrevar.xyz/

Andy Patterson wrote on 14 Oct 2018 23:05

Recipients:(name . Pierre Neidhardt)(address . mail@ambrevar.xyz)(address . 32769@debbugs.gnu.org)

Message-ID:20181015020505.2387a8c7@mailservices.uwaterloo.ca

Hi Pierre,

On Sun, 14 Oct 2018 10:27:46 +0200

Pierre Neidhardt <mail@ambrevar.xyz> wrote:

Toggle quote (3 lines)

> I won't have time this weekend, so you can go ahead. Otherwise I'll

> do it next week.

I investigated the problem and I think the build-program procedure is

ok. It has the option to specify which dependencies you want (maybe

that's what you used?), and I think that's the right option.

You said that you'd tried passing "next" in some way - in this case the

system that the executable will depend on is "next-gtk". There's a

couple of things here - firstly we need a system for "next" and a

separate one for "next/gtk". The other is that the resulting systems

will be normalized, so we'll end up with "next-gtk" as the system that

we want to use a dependency for the next-browser executable.

I made these changes and found that I still had problems building

next. I managed to figure out that it's because next exercises some

less common asd options which makes the bundle that it outputs not

conform to the expectations of the build system. We'll have to make

some changes to accommodate that, but for now, here's a patch that makes

next build.

I wasn't able to really test whether or not it works - it throws an

exception whenever I press a key. I was able to find out that its

*active-buffer* variable is nil, which it expects not to be. If you

want to debug the problems, I recommend doing something like

$ guix environment --ad-hoc sbcl sbcl-slynk:image next-browser:lib &&

slynk.image

* (slynk:create-server :port 4545)

and then connecting from sly. You might also pass

(sb-ext:enable-debugger) to the entry program if you prefer that

interface but I haven't tried. You can then run

|> (asdf:load-system 'next-gtk)

|> (next:start)

to get started.

Hoping that helps,

Andy

From 452bb2919ab1db6d819973d655407b1a335b2741 Mon Sep 17 00:00:00 2001

* gnu/packages/web-browsers.scm (next-browser-core): New variable.

(next-browser): Inherit from next-browser-core.

---

gnu/packages/lisp.scm | 1 +

gnu/packages/web-browsers.scm | 59 ++++++++++++++++++++++++-----------

2 files changed, 42 insertions(+), 18 deletions(-)

Toggle diff (116 lines)diff --git a/gnu/packages/lisp.scm b/gnu/packages/lisp.scm
index fcd3156a7..1b390b66d 100644
--- a/gnu/packages/lisp.scm
+++ b/gnu/packages/lisp.scm
@@ -30,6 +30,7 @@
   #:use-module ((guix licenses) #:prefix license:)
   #:use-module (guix packages)
   #:use-module (guix download)
+  #:use-module (guix hg-download)
   #:use-module (guix git-download)
   #:use-module (guix utils)
   #:use-module (guix build-system gnu)
diff --git a/gnu/packages/web-browsers.scm b/gnu/packages/web-browsers.scm
index 0df9180d7..c6844c868 100644
--- a/gnu/packages/web-browsers.scm
+++ b/gnu/packages/web-browsers.scm
@@ -25,6 +25,7 @@
 (define-module (gnu packages web-browsers)
   #:use-module ((guix licenses) #:prefix license:)
   #:use-module (guix packages)
+  #:use-module (guix utils)
   #:use-module (gnu packages)
   #:use-module (gnu packages compression)
   #:use-module (gnu packages databases)
@@ -324,10 +325,10 @@ access.")
 GUI.  It is based on PyQt5 and QtWebKit.")
     (license license:gpl3+)))
 
-(define-public next-browser
+(define-public next-browser-core
   (let ((commit "ccc289b44610ada4ae9c875910af7720e85b546b"))
     (package
-      (name "next-browser")
+      (name "next-browser-core")
       (version (git-version "0.0.8" "1" commit))
       (source (origin
                 (method git-fetch)
@@ -338,11 +339,11 @@ GUI.  It is based on PyQt5 and QtWebKit.")
                  (search-patches "next-fix-duplicate-function.patch"))
                 (sha256
                  (base32
-                  "1pcpb274zb0qm26rrglgcgzg2d2v7j2aja5685swwyq9rcqlbdf4"))))
+                  "0rccs9w49p577l12kgykccxrjkr03rf1cfz9mvxbjzc8g82dya2k"))
+                (file-name (git-file-name name version))))
       (build-system asdf-build-system/sbcl)
       ;; TODO: Move Common Lisp libraries to "native-inputs"?
       ;; TODO: Does runtime depends on gsettings-desktop-schemas and glib-networking?
-      (outputs '("out" "lib"))
       (inputs
        `(("alexandria" ,sbcl-alexandria)
          ("cl-strings" ,sbcl-cl-strings)
@@ -352,26 +353,48 @@ GUI.  It is based on PyQt5 and QtWebKit.")
          ("cl-sqlite" ,sbcl-cl-sqlite)
          ("parenscript" ,sbcl-parenscript)
          ("cl-json" ,sbcl-cl-json)
-         ("unix-opts" ,sbcl-unix-opts)
-         ("cl-cffi-gtk" ,sbcl-cl-cffi-gtk)
-         ("cl-webkit" ,sbcl-cl-webkit)
-         ("lparallel" ,sbcl-lparallel)))
+         ("unix-opts" ,sbcl-unix-opts)))
       (arguments
        `( ;; #:tests? #f
          #:asd-file "next/next.asd"
-         #:asd-system-name "next/gtk"
-         #:phases
-         (modify-phases %standard-phases
-           (add-after 'create-symlinks 'build-program
-             (lambda* (#:key outputs #:allow-other-keys)
-               (build-program
-                (string-append (assoc-ref outputs "out") "/bin/next-browser")
-                outputs
-                #:entry-program '((next:start) 0)))))
-         ))
+         #:asd-system-name "next"))
       (home-page "http://next.atlas.engineer/")
       (synopsis "Emacs-inspired web browser in extensible in Common Lisp")
       (description "Next is a keyboard-oriented, extensible web-browser inspired
 by Emacs and designed for power users.  The application has familiar
 key-bindings, is fully configurable and extensible in Lisp.")
       (license license:expat))))
+
+(define-public next-browser
+  (package
+    (inherit next-browser-core)
+    (name "next-browser")
+    (outputs '("out" "lib"))
+    (inputs
+     `(("next" ,next-browser-core)
+       ("cl-cffi-gtk" ,sbcl-cl-cffi-gtk)
+       ("cl-webkit" ,sbcl-cl-webkit)
+       ("lparallel" ,sbcl-lparallel)))
+    (arguments
+     (substitute-keyword-arguments (package-arguments next-browser-core)
+       ((#:asd-system-name _ #f) "next/gtk")
+       ((#:phases phases '%standard-phases)
+        `(modify-phases ,phases
+           (add-before 'cleanup 'move-bundle
+             (lambda* (#:key outputs #:allow-other-keys)
+               (define lib (assoc-ref outputs "lib"))
+               (define actual-fasl (string-append
+                                    lib
+                                    "/lib/sbcl/next/source/next-gtk.fasl"))
+               (define expected-fasl (string-append
+                                      lib
+                                      "/lib/sbcl/gtk--system.fasl"))
+               (copy-file actual-fasl expected-fasl)
+               #t))
+           (add-after 'create-symlinks 'build-program
+             (lambda* (#:key outputs #:allow-other-keys)
+               (build-program
+                (string-append (assoc-ref outputs "out") "/bin/next-browser")
+                outputs
+                #:entry-program '((next:start) 0)
+                #:dependencies '("next-gtk"))))))))))
-- 
2.19.1

Pierre Neidhardt wrote on 15 Oct 2018 02:34

Recipients:(name . Andy Patterson)(address . ajpatter@uwaterloo.ca)(address . 32769@debbugs.gnu.org)

Message-ID:87o9bvk0yx.fsf@ambrevar.xyz

Hi Andy!

Thanks for the hard work, this is pretty elegant indeed!

Question: do we need both "out" and "lib" outputs?

My understanding is that "lib" is useful when one wants to start the program

from a REPL. Correct?

Toggle quote (3 lines)

> I wasn't able to really test whether or not it works - it throws an

> exception whenever I press a key.

Same here. It does not do that when built manually with Quicklisp. Although I

recall seeing something similar when building manually with the wrong arguments.

I also suspect that some CFFI dependency (gtk?) is not packaged correctly and

some paths are missing.

I'll look into it.

Pierre Neidhardt

https://ambrevar.xyz/

Pierre Neidhardt wrote on 15 Oct 2018 03:04

Recipients:(name . Andy Patterson)(address . ajpatter@uwaterloo.ca)(address . 32769@debbugs.gnu.org)

Message-ID:87murfjzkx.fsf@ambrevar.xyz

It could also be a dependency version that's different from the one pulled by
Quicklisp.

-- 
Pierre Neidhardt
https://ambrevar.xyz/

Andy Patterson wrote on 15 Oct 2018 20:11

Recipients:(name . Pierre Neidhardt)(address . mail@ambrevar.xyz)(address . 32769@debbugs.gnu.org)

Message-ID:20181015231120.17d2bf7a@mailservices.uwaterloo.ca

Hi Pierre,

On Mon, 15 Oct 2018 11:34:14 +0200

Pierre Neidhardt <mail@ambrevar.xyz> wrote:

Toggle quote (4 lines)

> Question: do we need both "out" and "lib" outputs?

> My understanding is that "lib" is useful when one wants to start the

> program from a REPL. Correct?

"out" will always depend on "lib" for programs.  We want to build it
the same way as other systems so that all of the source for the
program remains inspect-able at run-time, even if it's just a
one-liner.

--
Andy

Pierre Neidhardt wrote on 19 Oct 2018 09:05

Recipients:(name . Andy Patterson)(address . ajpatter@uwaterloo.ca)(address . 32769@debbugs.gnu.org)

Message-ID:87sh11q5vm.fsf@ambrevar.xyz

I haven't managed to figure what's happening. According to John Mercouris (Next

developer), the state of CFFI-GTK is rather poor and it will be hard to track

down exactly what's wrong.

Next browser will soon move to another interface which does not rely on cffi-gtk

so it won't be a problem.

Andy, if you have an immediate solution, feel free to chime in, otherwise Next

will have to wait a few more weeks before hitting the Guix repository.

I've merged all the 78 (?) new Lisp packages onto master.

I've moved the Next package draft to wip-next-browser3.

Pierre Neidhardt

https://ambrevar.xyz/

Pierre Neidhardt wrote on 2 Dec 2018 04:32

control message for bug #32769

Recipients:(address . control@debbugs.gnu.org)

Message-ID:87ftvgun19.fsf@atlas.engineer

tags 32769 fixed

close 32769

Your comment

This issue is archived.

To comment on this conversation send an email to 32769@patchwise.org

To respond to this issue using the mumi CLI, first switch to it

mumi current 32769

Then, you may apply the latest patchset in this issue (with sign off)

mumi am -- -s

Or, compose a reply to this issue

mumi compose

Or, send patches to this issue

mumi send-email *.patch

is:open	open issues
is:done	closed issues
submitter:<who>	search issue submitter
author:<who>	search by message author
date:yesterday..now	search by issue date
mdate:3m..2d	search by message date