Kannan wrote:
> Julian Foad wrote:
> > I think that fixed the last of the URLs being passed to
> > svn_dirent_get_absolute(), so we can now assert that its input is not a
> > URL.
>
> Thank you. Seems the 'assert' is not yet moved to the repo, so sending
> the patch herewith.
>
> [[[
> Lo
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Julian Foad wrote:
[..]
>
> Or late tonight, in fact.
>
> It's a good patch. Committed revision 887469.
>
> Thank you, Kannan.
>
> I think that fixed the last of the URLs being passed to
> svn_dirent_get_absolute(), so we can now assert that its in
On Fri, 2009-12-04 at 18:33 +, Julian Foad wrote:
> Thanks. I have to run away right now, so if nobody beats me to it I'll
> read this on Monday.
Or late tonight, in fact.
It's a good patch. Committed revision 887469.
Thank you, Kannan.
I think that fixed the last of the URLs being passed t
Thanks. I have to run away right now, so if nobody beats me to it I'll
read this on Monday.
- Julian
> Julian Foad wrote:
> [..]
> > I think that means there are two more places where this mistake is made:
> > in 'log.c' and in 'mergeinfo.c'. Would you be willing to fix those, too?
>
> Fix req
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Julian Foad wrote:
[..]
> I think that means there are two more places where this mistake is made:
> in 'log.c' and in 'mergeinfo.c'. Would you be willing to fix those, too?
Fix requires only in log.c Attaching the patch herewith.
[[[
Log:
Fix some
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Julian Foad wrote:
[..]
> After those fixes, if I assert "!svn_path_is_url(relative)" in
> svn_dirent_get_absolute(), now I get only the following few test
> failures:
>
> FAIL: log_tests.py 2: 'svn log' on an empty repository
> FAIL: log_tests.py 6
On Wed, 2009-12-02, Kannan wrote:
> The patch attached herewith fixes the way urls are handled with
> dirents in `svn merge'.
>
> [[[
> Log:
> Make `svn merge' treat urls gracefully, by checking the type of the
> path obtained and invoking `svn_dirent_get_absolute()' only if its a wc
> path
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Julian Foad wrote:
[..]
I suggest adding
SVN_ERR_ASSERT(!svn_path_is_url(relative));
[..]
>> "Assertions are forever." (A quote from the book "Writing Solid Code".)
>> The purpose of assertions is to catch programming errors (bugs
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Julian Foad wrote:
> Oh yes, so you did. However, that patch you initially proposed there
> passed a part of the URL to svn_dirent_get_absolute(), which is not
> right either.
Yes, then I had a different patch altogether.
>>> I suggest adding
>>>
>>
On Fri, 2009-11-27 at 18:45 +, Julian Foad wrote:
> On Fri, 2009-11-27 at 23:25 +0530, Kannan wrote:
> > -BEGIN PGP SIGNED MESSAGE-
> > Hash: SHA1
> >
> > Julian Foad wrote:
> > >> I spotted an error in the conversion to use abs-paths:
> > >>
> > >> [[[
> > >> static svn_error_t *
> >
On Fri, 2009-11-27 at 23:25 +0530, Kannan wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
>
> Julian Foad wrote:
> >> I spotted an error in the conversion to use abs-paths:
> >>
> >> [[[
> >> static svn_error_t *
> >> normalize_merge_sources(apr_array_header_t **merge_sources_p,
> > [...]
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Julian Foad wrote:
>> I spotted an error in the conversion to use abs-paths:
>>
>> [[[
>> static svn_error_t *
>> normalize_merge_sources(apr_array_header_t **merge_sources_p,
> [...]
>> SVN_ERR(svn_dirent_get_absolute(&source_abspath, source, pool))
12 matches
Mail list logo
