Ulli Horlacher | 15 May 23:37 2008
Picon

groovy-mode?


Is there a groovy_mode for jed?

groovy is a language based on the java virtual machine, see
http://groovy.codehaus.org/

--

-- 
Ullrich Horlacher              Server- und Arbeitsplatzsysteme
Rechenzentrum                  E-Mail: horlacher <at> rus.uni-stuttgart.de
Universitaet Stuttgart         Tel:    ++49-711-685-65868
Allmandring 30                 Fax:    ++49-711-682357
70550 Stuttgart (Germany)      WWW:    http://www.rus.uni-stuttgart.de/

Jörg Sommer | 18 May 21:37 2008
Picon

Patches for jed

Hi John,

I've some patches I would like to see included in your source tree.

http://git.debian.org/?p=pkg-jed/jed.git;a=commitdiff;h=7760b2b2a1437beac98ebe9f93dd813cbc5b611b

    Split off the non‐binary part from the install

    The Debian people prefer to install the non‐binary files with its own
    rule to create different packages: one with the non‐binary files and
    others with the binary files for each architecture.

http://git.debian.org/?p=pkg-jed/jed.git;a=commitdiff;h=d1558cb5d7b7e1cd4d8519976efb85453d6c1fc3

    Remove dynamic linking checks and options

    The dynamic linking stuff isn't used by Jed directly. The SLang library
    does the work. The bad side of these dynamic linking stuff is that it
    introduces an unnecessary dependency on libdl. It doesn't use anything
    from libdl, but by linking to it it adds a dependency to the ELF header.

http://git.debian.org/?p=pkg-jed/jed.git;a=commitdiff;h=746c81ae56a120cd29133028fae1b7f07c9aaa49

    Don't create JED_ROOT/bin, if it end up empty

    If getmail is not built, the directory JED_ROOT/bin would be empty after
    installation. This isn't nice.

http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=473617

(Continue reading)

John E. Davis | 19 May 06:25 2008
Picon

Re: Patches for jed

Jörg Sommer <joerg <at> alea.gnuu.de> wrote:
> I've some patches I would like to see included in your source tree.

I added the patches except for the dynamic linking one:

> http://git.debian.org/?p=pkg-jed/jed.git;a=commitdiff;h=d1558cb5d7b7e1cd4d8519976efb85453d6c1fc3
>
>     Remove dynamic linking checks and options
>
>     The dynamic linking stuff isn't used by Jed directly. The SLang library
>     does the work. The bad side of these dynamic linking stuff is that it
>     introduces an unnecessary dependency on libdl. It doesn't use anything
>     from libdl, but by linking to it it adds a dependency to the ELF header.

The problem is that the patch removes the --export-dynamic flag that
some modules require if they use functions provided by jed.  As I
recall, Paul has some modules that do this.  Is there any particular
reason why the patch removed this?

Thanks,
--John

Jörg Sommer | 19 May 12:11 2008
Picon

Libraries linked with Jed

Hallo John,

"John E. Davis" <davis <at> space.mit.edu> wrote:
> Jörg Sommer <joerg <at> alea.gnuu.de> wrote:
>> I've some patches I would like to see included in your source tree.
>
> I added the patches

Thanks.

> except for the dynamic linking one:
>
>> http://git.debian.org/?p=pkg-jed/jed.git;a=commitdiff;h=d1558cb5d7b7e1cd4d8519976efb85453d6c1fc3
>>
>>     Remove dynamic linking checks and options
>>
>>     The dynamic linking stuff isn't used by Jed directly. The SLang library
>>     does the work. The bad side of these dynamic linking stuff is that it
>>     introduces an unnecessary dependency on libdl. It doesn't use anything
>>     from libdl, but by linking to it it adds a dependency to the ELF header.
>
> The problem is that the patch removes the --export-dynamic flag that
> some modules require if they use functions provided by jed.  As I
> recall, Paul has some modules that do this.  Is there any particular
> reason why the patch removed this?

I didn't see that impact. My main intention was to remove the dependency
to libdl, because jed doesn't use it. A while ago, Debian started to
evaluate dependencies on binary level, i.e. they look at the ELF header
and generate the list of needed libraries (and packages) from this. Jed
(Continue reading)

Jörg Sommer | 19 May 12:28 2008
Picon

More patches

Hi again,

diff -urNad 0.99.19~/doc/tm/rtl/mini.tm 0.99.19/doc/tm/rtl/mini.tm
--- 0.99.19~/doc/tm/rtl/mini.tm 2006-02-13 18:51:58.000000000 +0100
+++ 0.99.19/doc/tm/rtl/mini.tm  2007-06-09 23:07:39.141893086 +0200
 <at>  <at>  -101,9 +101,11  <at>  <at> 
 #v-
   reads a string from the user and then searches forward for it and if
   not found, it resumes the search from the beginning of the buffer.
-  Note: If the user aborts the function \var{mini_read} by pressing the
+
+\notes
+  If the user aborts the function \var{read_mini} by pressing the
   keyboard quit character (e.g., Ctrl-G), an error is signaled.  This
-  error can be caught by an \var{ERROR_BLOCK} and the appropriate action
+  error can be caught by a \var{try-catch} statement and the appropriate action
   taken. Also if the mini-buffer is already in use, this function should
   not be called.  The variable \var{MINIBUFFER_ACTIVE} may be checked to
   determine if this is the case or not.

http://git.debian.org/?p=pkg-jed/jed.git;a=blob;hb=jo-0.99.19;f=debian/patches/fix-update-before-key-hook.dpatch

    The LASTKEY variable is not updated before the _jed_before_key_hooks
    hooks are called, so they see a wrong key sequencs.

http://git.debian.org/?p=pkg-jed/jed.git;a=blob;hb=jo-0.99.19;f=debian/patches/fix-help.sl.dpatch

    * describe_bindings() doesn't take care of CASE_SEARCH with leads to the
      problem that "ESC [ A" and "ESC [ a" is replace by the same key
    * describe_bindings() gets confused by newlines in the key definiton,
(Continue reading)

G. Milde | 19 May 14:36 2008
Picon
Picon

Re: More patches

On 19.05.08, J�rg Sommer wrote:
> Hi again,

> diff -urNad 0.99.19~/doc/tm/rtl/mini.tm 0.99.19/doc/tm/rtl/mini.tm

> -  error can be caught by an \var{ERROR_BLOCK} and the appropriate action
> +  error can be caught by a \var{try-catch} statement and the appropriate action

Should be ``\var{try} - \var{catch}`` statement, as try and catch arr S-Lang
keywords that should be marked (like ERROR_BLOCK in the old version).

> http://git.debian.org/?p=pkg-jed/jed.git;a=blob;hb=jo-0.99.19;f=debian/patches/extend-help.sl.dpatch

>     * Improves expand_keystring() to return Alt- instead of ESC, if ALT_CHAR
>       is non-zero, and print Space instead of the space character

Please test for ALT_CHAR == 27 instead of != 0:

 DESCRIPTION
  If this variable is non-zero, characters pressed in combination the
  `Alt' key will generate a two character sequence: the first
  character is the value of `ALT_CHAR' itself followed by the
  character pressed.  For example, if `Alt-X' is pressed and
  `ALT_CHAR' has a value of 27, the characters `ESC X' will be
  generated.

> http://git.debian.org/?p=pkg-jed/jed.git;a=blob;hb=jo-0.99.19;f=debian/patches/remove_jed_library_path.dpatch

>     In version 0.99.19-82 the require and provides functions where dropped
>     from jed and the functions from slsh stepped in. They have a different
(Continue reading)

Jörg Sommer | 20 May 01:37 2008
Picon

Re: More patches

Hallo G.,

"G. Milde" <milde <at> users.sourceforge.net> wrote:
> On 19.05.08, Jörg Sommer wrote:
>
>> diff -urNad 0.99.19~/doc/tm/rtl/mini.tm 0.99.19/doc/tm/rtl/mini.tm
>
>> -  error can be caught by an \var{ERROR_BLOCK} and the appropriate action
>> +  error can be caught by a \var{try-catch} statement and the appropriate action
>
> Should be ``\var{try} - \var{catch}`` statement, as try and catch arr S-Lang
> keywords that should be marked (like ERROR_BLOCK in the old version).

Yes, you are right.

>> http://git.debian.org/?p=pkg-jed/jed.git;a=blob;hb=jo-0.99.19;f=debian/patches/extend-help.sl.dpatch
>
>>     * Improves expand_keystring() to return Alt- instead of ESC, if ALT_CHAR
>>       is non-zero, and print Space instead of the space character
>
> Please test for ALT_CHAR == 27 instead of != 0:
>
>  DESCRIPTION
>   If this variable is non-zero, characters pressed in combination the
>   `Alt' key will generate a two character sequence: the first
>   character is the value of `ALT_CHAR' itself followed by the
>   character pressed.  For example, if `Alt-X' is pressed and
>   `ALT_CHAR' has a value of 27, the characters `ESC X' will be
>   generated.

(Continue reading)

G. Milde | 20 May 10:43 2008
Picon
Picon

Re: More patches

On 19.05.08, J�rg Sommer wrote:
> "G. Milde" <milde <at> users.sourceforge.net> wrote:
> > On 19.05.08, Jörg Sommer wrote:

> >> http://git.debian.org/?p=pkg-jed/jed.git;a=blob;hb=jo-0.99.19;f=debian/patches/remove_jed_library_path.dpatch
> >
> >>     In version 0.99.19-82 the require and provides functions where dropped
> >>     from jed 
...
> > Sounds a lot like incompatible with jed-extra. --> if so, and if we are
> > going to upload 0.99 to unstable, it should conflict with the current
> > jed-extra version.

> This patch is still in pre117 and I discovered no problems. Do you have
> any ideas where it might crash?

It will not necessarily crash, but `different argument list than the old
jed functions and they do not care about jed library path` sounds like an
incompatible change at the first glance. I wonder whether libdir.sl and
make_ini.sl still work as expected.

I could expect other incompatibilities too, not necessarily serious and
crashing but still annoying. 

I have to remember, that a lot of incompatibilities are fixed already
thanks to your testing of the jed 0.99 -- jed-extra combination.

Could you run the jedmodes unit tests with 0.99 and tell me about any
errors (or just mail a testreport)? There is a utility script in
jed-extra/trunk/utils/do-unittests.sh.
(Continue reading)

Jörg Sommer | 21 May 01:19 2008
Picon

Re: More patches

Hallo G.,

"G. Milde" <milde <at> users.sourceforge.net> wrote:
> On 19.05.08, Jörg Sommer wrote:
>> "G. Milde" <milde <at> users.sourceforge.net> wrote:
>> > On 19.05.08, Jörg Sommer wrote:
>
>> >> http://git.debian.org/?p=pkg-jed/jed.git;a=blob;hb=jo-0.99.19;f=debian/patches/remove_jed_library_path.dpatch
>> >
>> >>     In version 0.99.19-82 the require and provides functions where dropped
>> >>     from jed 
> ...
>> > Sounds a lot like incompatible with jed-extra. --> if so, and if we are
>> > going to upload 0.99 to unstable, it should conflict with the current
>> > jed-extra version.
>
>> This patch is still in pre117 and I discovered no problems. Do you have
>> any ideas where it might crash?
>
> It will not necessarily crash, but `different argument list than the old
> jed functions and they do not care about jed library path` sounds like an
> incompatible change at the first glance. I wonder whether libdir.sl and
> make_ini.sl still work as expected.

Do you remeber this thread?
http://lists.alioth.debian.org/pipermail/pkg-jed-devel/2007-June/001457.html

> Could you run the jedmodes unit tests with 0.99 and tell me about any
> errors (or just mail a testreport)? There is a utility script in
> jed-extra/trunk/utils/do-unittests.sh.
(Continue reading)

G. Milde | 21 May 12:39 2008
Picon
Picon

Re: More patches

On 20.05.08, J�rg Sommer wrote:

> Do you remeber this thread?
> http://lists.alioth.debian.org/pipermail/pkg-jed-devel/2007-June/001457.html

No. Thanks for remainding me.

If I got this right, than 

* the "changed syntax of require()" problem is fixed,

* jed 0.99 now uses slang_load_path() instead of jed_library_path() for
  both, require() and evalfile().

* a patch in site.sl will alias (s|g)et_jed_library_path() to
  (s|g)et_slang_load_path().

Then this issue should be solved. (There might be others.)

Please correct me if this assumption is wrong.  

> > Could you run the jedmodes unit tests with 0.99 and tell me about any
> > errors (or just mail a testreport)? There is a utility script in
> > jed-extra/trunk/utils/do-unittests.sh.

> I will do so, tomorrow.

Fine.

Thanks,
(Continue reading)


Gmane