Hey Arfrever,
Sorry that I haven't been able to catch you on IRC, I saw you tried to
ping me a few times on week-ends but I was AFK and when coming back,
you were offline. Then I browsed the dev@ mailing list and found a
few emails for me that had slipped through. Apologies, going to
re-configure my MUA filters and catch up.
I saw you merged the svnpatch-diff branch back to trunk. Well done
and thanks for all your hard work on keeping it sync'ed too! It seems
some svnpatch tests on trunk have been failing and caused people to
unleash their anger. As far as I remember, all tests passed on my box
(GNU/linux) back then. I had documented somewhere (notes/svnpatch or
docstrings) that svnpatch relies on patch(1)-able systems to apply the
unidiff chunk on the working-copy so I'm not surprised they fail on a
patch(1)-less Windows box at least, although I remember someone (was
it you? or Bert?) had done a bit of testing on Windows. Would have
been great to have the Buildbot on the branches, or at least BOD
(Buildbot on demand) on the branches :-)
I guess it's a bit late now to answer your questions below but don't
hesitate to CC me again, hopefully I'll have those filters sorted out
and I'll be able to help in time!
Charles
On Mon, Feb 16, 2009 at 11:58 PM, Arfrever Frehtes Taifersar Arahesis
<arfrever.fta_at_gmail.com> wrote:
> Hello Charles Acknin,
>
> svn_wc.h on svnpatch-diff branch contains some new public functions:
>
> ========================================================================
>
> /**
> *
> * @defgroup svn_wc_svnpatch svnpatch related functions
> *
> * @{
> *
> */
>
> /* Output -- Writing */
>
> /* Append @a number into @a target stream. */
> svn_error_t *
> svn_wc_write_number(svn_stream_t *target,
> apr_pool_t *pool,
> const apr_uint64_t number);
>
> /* Append @a str into @a target stream. Is binary-able. */
> svn_error_t *
> svn_wc_write_string(svn_stream_t *target,
> apr_pool_t *pool,
> const svn_string_t *str);
>
> /* Append @a s cstring into @a target stream. */
> svn_error_t *
> svn_wc_write_cstring(svn_stream_t *target,
> apr_pool_t *pool,
> const char *s);
>
> /* Append @a word into @a target stream. */
> svn_error_t *
> svn_wc_write_word(svn_stream_t *target,
> apr_pool_t *pool,
> const char *word);
>
> /* Append a list of properties @a props into @a target. */
> svn_error_t *
> svn_wc_write_proplist(svn_stream_t *target,
> apr_hash_t *props,
> apr_pool_t *pool);
>
> /* Begin a list, appended into @target */
> svn_error_t *
> svn_wc_start_list(svn_stream_t *target);
>
> /* End a list, appended into @target */
> svn_error_t *
> svn_wc_end_list(svn_stream_t *target);
>
> /* Append a tuple into @target in a printf-like fashion.
> * @see svn_ra_svn_write_tuple() for further details with the format. */
> svn_error_t *
> svn_wc_write_tuple(svn_stream_t *target,
> apr_pool_t *pool,
> const char *fmt, ...);
>
> /* Append a command into @target, using the same format notation as
> * svn_wc_write_tuple(). */
> svn_error_t *
> svn_wc_write_cmd(svn_stream_t *target,
> apr_pool_t *pool,
> const char *cmdname,
> const char *fmt, ...);
>
> /* Input -- Reading */
>
> svn_error_t *
> svn_wc_read_item(svn_stream_t *from,
> apr_pool_t *pool,
> svn_ra_svn_item_t **item);
>
> svn_error_t *
> svn_wc_parse_tuple(apr_array_header_t *list,
> apr_pool_t *pool,
> const char *fmt, ...);
>
> svn_error_t *
> svn_wc_read_tuple(svn_stream_t *from,
> apr_pool_t *pool,
> const char *fmt, ...);
>
> /* Drive @a diff_editor against @a decoded_patch_file's clear-text
> * Editor Commands. */
> svn_error_t *
> svn_wc_apply_svnpatch(apr_file_t *decoded_patch_file,
> const svn_delta_editor_t *diff_editor,
> void *diff_edit_baton,
> apr_pool_t *pool);
>
> /* Run an external patch program against @a patch_path patch file. @a
> * outfile and @a errfile are respectively connected to the external
> * program's stdout and stderr pipes when executed. @a config is looked
> * up for the SVN_CONFIG_OPTION_PATCH_CMD entry to use as the patch
> * program. If missing or @a config is @c NULL, the function tries to
> * execute 'patch' literally, which should work on most *NIX systems at
> * least. This involves searching into $PATH. The external program is
> * given the patch file via its stdin pipe.
> *
> * The program is passed the '--force' argument when @a force is set.
> */
> svn_error_t *
> svn_wc_apply_unidiff(const char *patch_path,
> svn_boolean_t force,
> apr_file_t *outfile,
> apr_file_t *errfile,
> apr_hash_t *config,
> apr_pool_t *pool);
>
> /** @} end group: svnpatch related functions */
>
> ========================================================================
>
> I have some questions and suggestions:
>
> - Are all these functions specific to libsvn_wc?
> (If not, then they could be in libsvn_subr.)
>
> - If these functions are specific to svnpatch, then their names should
> contain "svnpatch", e.g. svn_wc_svnpatch_write_number().
>
> - Do these functions have to be public?
>
> - svn_wc_write_proplist() isn't used anywhere. Is this function needed?
>
> - svn_wc_read_item(), svn_wc_parse_tuple() and svn_wc_read_tuple()
> should have doc strings.
>
> - Could you improve doc strings of these functions? (e.g. add purpose
> of every function etc.)
>
> --
> Arfrever Frehtes Taifersar Arahesis
>
------------------------------------------------------
http://subversion.tigris.org/ds/viewMessage.do?dsForumId=462&dsMessageId=1560354
Received on 2009-04-06 12:01:34 CEST