I decided that I wasn't happy ignoring errors here, even if consistency checks
can be made earlier: r31729.
With this change, I can also remove an assert(): r31730.
Blair
Blair Zajac wrote:
> Thanks, so I'll add checking code later.
>
> However, if an invalid revision kind does appear there, there's no
> mechanism to bubble up an error through write_entry. Should we have one?
>
> Without the ability to propagate an error, I'm going to have the file
> external just not write the file externals information into the entries
> file, which will leave it as a normal switched file to successive svn
> operations. If a user sees a file external turn into a normal switched
> file, then hopefully they'll report a bug.
>
> C. Michael Pilato wrote:
>> Seems to me like it'd make more sense to validate the entry bits quite
>> a bit earlier, like in svn_wc__modify_entry() or one of its helper
>> functions. Entries are cached, so writing them out can happen later
>> than when they are modified. You don't want to allow the opportunity
>> to refer to a cached bogus entry that wasn't caught simply because it
>> hasn't been flushed to disk yet.
>>
>>
>> Blair Zajac wrote:
>>> I am using a svn_opt_revision_t to store the file external revision
>>> number, but I only allow the svn_opt_revision_unspecified,
>>> svn_opt_revision_head and svn_opt_revision_number kinds.
>>>
>>> I'd like to have write_entry() check the validity of the
>>> svn_opt_revision_t value before it writes it out, but currently
>>> write_entry() returns void.
>>>
>>> I was thinking of having write_entry() return a svn_error_t * and
>>> then I could bubble up any errors.
>>>
>>> Is this a good place to put a check that the rest of my code has
>>> properly set the svn_wc_entry_t fields?
>>>
>>> Blair
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe_at_subversion.tigris.org
> For additional commands, e-mail: dev-help_at_subversion.tigris.org
>
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe_at_subversion.tigris.org
For additional commands, e-mail: dev-help_at_subversion.tigris.org
Received on 2008-06-13 17:25:33 CEST