Skip to content

C
ccan
Switch branch/tag
  1. 28 May, 2015 2 commits
    • David Gibson's avatar
      asearch: Add context pointer to asearch comparison callback · e7519047
      David Gibson authored 
      asearch, like the standard library bsearch, takes a comparison callback.
Like bsearch() that callback doesn't include a user supplied context
pointer.  As well as being generally limiting, that makes the comparison
functions used by asearch gratuitously different from those used by the
asort module.

This patch alters this.  Note that this is an incompatible change to the
asearch interface.  I think this is worth it to correct the oversight, but
others might disagree.  At present the only user within ccan is ntdb, which
is corrected to match.

This means actually supplying a binary search implementation, rather than
relying on bsearch() from the standard library.  We follow the lead of the
asort module and steal^H^H^H^H^Hadapt the implementation from glibc.
Signed-off-by: default avatarDavid Gibson <david@gibson.dropbear.id.au>
Signed-off-by: default avatarRusty Russell <rusty@rustcorp.com.au>
      e7519047
    • David Gibson's avatar
      asearch: Correct license tag to LGPL 2.1+ · 8ad4e453
      David Gibson authored 
      The _info file for asearch currently gives the license as just "LGPL".
That gives a warning from ccanlint, because it interprets that as the
latest LGPL 3 whereas asearch.h and the license link indicate it's actually
LGPL 2.1+.
Signed-off-by: default avatarDavid Gibson <david@gibson.dropbear.id.au>
Signed-off-by: default avatarRusty Russell <rusty@rustcorp.com.au>
      8ad4e453
  3. 27 May, 2015 4 commits
  5. 25 May, 2015 9 commits
  7. 24 May, 2015 1 commit
  9. 20 May, 2015 2 commits
  11. 19 May, 2015 4 commits
  13. 14 May, 2015 1 commit
  15. 12 May, 2015 1 commit
  17. 05 May, 2015 1 commit
  19. 25 Apr, 2015 1 commit
  21. 24 Apr, 2015 4 commits
  23. 13 Apr, 2015 1 commit
    • Delio Brignoli's avatar
      tal_stack · 73930140
      Delio Brignoli authored 
      Hi Rusty,

Thanks for reviewing the patch. V2 is attached, see my comments below.

> On 31 Mar 2015, at 02:36, Rusty Russell <rusty@rustcorp.com.au> wrote:
>
> Delio Brignoli <brignoli.delio@gmail.com> writes:
>> Hi All,
>>
>> tal_stack implements a (trivial) stack of tal contexts. Would this be a worthy addition to CCAN? (not necessarily in its current form).

[…]

>        This is cute; I’ve seen similar used in Samba.  It's

Indeed, it was inspired by talloc_stack.h ;-)

[…]

> You are missing a _info file: I would create that, and put your example
> in an Example: section there.

I moved the module and tests under can/tal/stack and added a LICENSE and _info.

> Other random advice:
> 1) You should also document the tal_newframe function (particularly note
>   that you're expected to tal_free the result, and that it will free
>   any future unfreed frames).  And note that it’s not threadsafe.

Done.

> 2) You probably want tal_newframe to be a macro, and hand file and line
>   thought to the tal_alloc_ call.  That makes debugging nicer when
>   you iterate the tree.

Done. The macro is calling a tal_newframe_() function because I’d rather not make the module’s stack variable ‘public’.

> 3) Consider whether you want to declare a dummy type 'struct tal_stack'.
>   Probably pretty unnecessary since it’s quite clear.

Skipped this one. We can declare it later if we change our minds.

Thanks
—
Delio

From c2ceb9258d97b0dcb72e7b6986cfd2bd394b254e Mon Sep 17 00:00:00 2001
From: Delio Brignoli <dbrignoli@audioscience.com>
Date: Sun, 15 Mar 2015 13:26:40 +0100
Subject: [PATCH] tal_stack: new module - V2
Signed-off-by: default avatarRusty Russell <rusty@rustcorp.com.au>
      73930140
  25. 10 Apr, 2015 1 commit
    • Peter Barker's avatar
      ccanlint: avoid segfault when module_builds' linking fails · bdb8d751
      Peter Barker authored 
      In the case that the objects built but linking failed, module_builds.c
called score_file_error with a NULL ccan_file object and 0 for line
number.

score_file_error assumed that the ccan_file object it is passed was
not-NULL when appending file errors to the score's aggregate error
string.  It attempted to dereference it to get "fullname".

score_error was factored out from score_file_error.  It takes a
"source" parameter, which is the file's full name (and possibly line
number) in the score_file_error case, and the ccan module name in the
case of link failure.
      bdb8d751
  27. 30 Mar, 2015 8 commits
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7