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

Re: [PATCH 3/8] pathspec: change PATHSPEC_STRIP_SUBMODULE_SLASH_EXPENSIVE flag

From
BWBrandon Williams <bmwill@google.com>
Date
May 10, 2017, 16:16 UTC
Message-ID
<20170510161633.GA41649@google.com>
In-Reply-To
<xmqqefvxxnpq.fsf@gitster.mtv.corp.google.com>
On 05/10, Junio C Hamano wrote:
Show 25 quoted lines
> Brandon Williams <bmwill@google.com> writes:
> 
> > It's confusing to have two different 'strip submodule slash' flags which
> > do subtly different things.  Both
> > PATHSPEC_STRIP_SUBMODULE_SLASH_EXPENSIVE and
> > PATHSPEC_STRIP_SUBMODULE_SLASH_CHEAP will accomplish the same task of
> > striping a slash from a pathspec which matches a submodule entry in the
> > index.  The only difference is that
> > PATHSPEC_STRIP_SUBMODULE_SLASH_EXPENSIVE will perform additional checks
> > and die if a pathspec has a leading path component which corresponds to
> > a submodule.  This additional functionality should be split out into its
> > own flag.
> >
> > To this end, rename the PATHSPEC_STRIP_SUBMODULE_SLASH_EXPENSIVE flag to
> > PATHSPEC_SUBMODULE_LEADING_PATH and change its behavior to only die if a
> > path descends into a submodule.  In addition add the
> > PATHSPEC_STRIP_SUBMODULE_SLASH_CHEAP flag to callers which relied on the
> > old slash stripping functionality.
> 
> "PATHSPEC_SUBMODULE_LEADING_PATH" feels like an unfinished sentence
> to me.  Do I understand your description correctly if I say it is
> about "checking" the leading path to see if it overlaps with a
> submodule path?  IOW, I am wondering if the name should have the
> word CHECK somewhere in it.
> 

You're probably right, the name has something left to be desired. I chose it simply because it conforms to another flag "PATHSPEC_SYMLINK_LEADING_PATH" which dies if there is a symlink in the leading path.

And you are correct, it checks if the leading path overlaps with a submodule path. If it strictly matches a submodule path that's alright though. The point of this flag is to disallow paths which descend into submodules. One such use case is to prevent a user from trying to 'git add' a file from a submodule.

-- 
Brandon Williams
Previous: Junio C HamanoNext: Brandon Williams
Message 5 of 29 in “convert pathspec.c to take an index parameter”
  1. 0/8 convert pathspec.c to take an index parameterBrandon Williams, May 9, 2017
  2. 2/8 submodule: add die_in_unpopulated_submodule functionBrandon Williams, May 9, 2017
  3. 3/8 pathspec: change PATHSPEC_STRIP_SUBMODULE_SLASH_EXPENSIVE flagBrandon Williams, May 9, 2017
  4. Junio C HamanoMay 10, 2017
  5. Brandon WilliamsMay 10, 2017
  6. 1/8 pathspec: provide a more descriptive die messageBrandon Williams, May 9, 2017
  7. 5/8 pathspec: convert strip_submodule_slash to take an indexBrandon Williams, May 9, 2017
  8. Junio C HamanoMay 10, 2017
  9. 7/8 pathspec: convert init_pathspec_item to take an indexBrandon Williams, May 9, 2017
  10. 6/8 pathspec: convert find_pathspecs_matching_against_index to take an indexBrandon Williams, May 9, 2017
  11. 4/8 pathspec: rename PATHSPEC_STRIP_SUBMODULE_SLASH_CHEAPBrandon Williams, May 9, 2017
  12. Junio C HamanoMay 10, 2017
  13. 8/8 pathspec: convert parse_pathspec to take an indexBrandon Williams, May 9, 2017
  14. Junio C HamanoMay 10, 2017
  15. Brandon WilliamsMay 10, 2017
  16. Junio C HamanoMay 11, 2017
  17. Johannes SixtMay 11, 2017
  18. Junio C HamanoMay 11, 2017
  19. Brandon WilliamsMay 11, 2017
  20. Junio C HamanoMay 12, 2017
  21. 0/6 convert pathspec.c to take an index parameterBrandon Williams, May 11, 2017
  22. 2/6 submodule: add die_in_unpopulated_submodule functionBrandon Williams, May 11, 2017
  23. 4/6 ls-files: prevent prune_cache from overeagerly pruning submodulesBrandon Williams, May 11, 2017
  24. 3/6 pathspec: remove PATHSPEC_STRIP_SUBMODULE_SLASH_EXPENSIVE flagBrandon Williams, May 11, 2017
  25. 6/6 pathspec: convert find_pathspecs_matching_against_index to take an indexBrandon Williams, May 11, 2017
  26. 5/6 pathspec: remove PATHSPEC_STRIP_SUBMODULE_SLASH_CHEAPBrandon Williams, May 11, 2017
  27. 1/6 pathspec: provide a more descriptive die messageBrandon Williams, May 11, 2017
  28. Junio C HamanoMay 12, 2017
  29. Brandon WilliamsMay 12, 2017

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.