git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: [PATCHv4 8/8] clone: recursive and reference option triggers submodule alternates

From
Stefan Beller <sbeller@google.com>
Date
Aug 17, 2016, 21:41 UTC
Message-ID
<CAGZ79kaqy16PLm8aZkAzp6PtTPFDuN80CiPPi-bbU9tec3X_oA@mail.gmail.com>
In-Reply-To
<xmqqd1l784lf.fsf@gitster.mtv.corp.google.com>
On Wed, Aug 17, 2016 at 2:31 PM, Junio C Hamano <gitster@pobox.com> wrote:
Show 31 quoted lines
> Junio C Hamano <gitster@pobox.com> writes:
>
>> Stefan Beller <sbeller@google.com> writes:
>>
>>> +static void prepare_possible_alternates(const char *sm_name,
>>> +            struct string_list *reference)
>>> +{
>>> +    char *sm_alternate = NULL, *error_strategy = NULL;
>>> +    struct submodule_alternate_setup sas = SUBMODULE_ALTERNATE_SETUP_INIT;
>>> +
>>> +    git_config_get_string("submodule.alternateLocation", &sm_alternate);
>>> +    if (!sm_alternate)
>>> +            return;
>>> +
>>> +    git_config_get_string("submodule.alternateErrorStrategy", &error_strategy);
>>
>> I have to admit that I need to follow the codepath in config.c every
>> time to check, but I _think_ git_config_get_string() gives you your
>> own copy of the value.  As this function does not give ownership of
>> sm_alternate or error_strategy to something else, they are leaked
>> every time this function is called, I think.
>>
>>> +    sas.submodule_name = sm_name;
>>> +    sas.reference = reference;
>>> +    if (!strcmp(error_strategy, "die"))
>>> +            sas.error_mode = SUBMODULE_ALTERNATE_ERROR_DIE;
>
> Another thing I noticed but forgot to mention.  Can error_strategy
> be NULL here?  We are assuming sm_alternate can be, so I presume
> that it is sensible to protect against dereferencing a NULL here,
> too?
Oh!

Of course we need to fix that too. That slipped in because I assumed that those two variables go in tandem (if sm_alternate is set, then error_strategy is set as well), but that is not the case of course.

Which leads to another thing: we need to have a default for error_strategy. I think "die" is reasonable here. That also needs a documentation update.

Show 6 quoted lines
>
>>> +    if (!strcmp(error_strategy, "info"))
>>> +            sas.error_mode = SUBMODULE_ALTERNATE_ERROR_INFO;
>>> +    if (!strcmp(sm_alternate, "superproject"))
>>> +            foreach_alt_odb(add_possible_reference_from_superproject, &sas);
>>> +}
Previous: Junio C Hamano
Message 18 of 18 in “[PATCHv4 0/6] git clone: Marry --recursive and --reference”
  1. Stefan BellerAug 11, 2016
  2. 1/8 t7408: modernize styleStefan Beller, Aug 11, 2016
  3. 2/8 t7408: merge short tests, factor out testing methodStefan Beller, Aug 11, 2016
  4. 3/8 submodule--helper module-clone: allow multiple referencesStefan Beller, Aug 11, 2016
  5. 4/8 submodule--helper update-clone: allow multiple referencesStefan Beller, Aug 11, 2016
  6. 5/8 clone: factor out checking for an alternate pathStefan Beller, Aug 11, 2016
  7. Junio C HamanoAug 12, 2016
  8. Stefan BellerAug 15, 2016
  9. Junio C HamanoAug 15, 2016
  10. Stefan BellerAug 15, 2016
  11. 7/8 clone: implement optional referencesStefan Beller, Aug 11, 2016
  12. 6/8 clone: clarify option_reference as requiredStefan Beller, Aug 11, 2016
  13. 8/8 clone: recursive and reference option triggers submodule alternatesStefan Beller, Aug 11, 2016
  14. Junio C HamanoAug 12, 2016
  15. Junio C HamanoAug 17, 2016
  16. Stefan BellerAug 17, 2016
  17. Junio C HamanoAug 17, 2016
  18. Stefan BellerAug 17, 2016

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.