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

[PATCH v3 00/12] do not overuse strbuf_split*()

From
Junio C Hamano <gitster@pobox.com>
Date
Aug 3, 2025, 06:52 UTC
Message-ID
<20250803065304.3325286-1-gitster@pobox.com>
In-Reply-To
<20250801220423.1230969-1-gitster@pobox.com>

strbuf is a very good data structure to work with string data without having to worry about running past the end of the string.

But an array of strbuf is often a wrong data structure. You rarely have need to be able to edit multiple strings represented by such an array simultaneously. And strbuf_split*() that produces result in such a shape is a misdesigned API function.

The most common use case of strbuf_split*() family of functions seems to be to trim away the whitespaces around each piece of split string. With modern string_list_split*(), it is often no longer necessary.

This series builds on top of the other series that extends string list API to allow string_list_split() to take more than one delimiter bytes, and to optionally trim the resulting string pieces.

I do not plan to eradicate all the uses of strbuf_split*() myself, not because I found some valid use cases in the existing code (I haven't yet), but these patches would give interested others enough material to study and mimic to continue the effort and I can safely leave it as #leftoverbits to rewrite them.

Relative to v2, this iteration v3 adds one more clean-up step to correct a callee that insists on taking a while strbuf when it can work with any NUL-terminated strings, and comes with a handful of typofixes.

Junio C Hamano (12):
  wt-status: avoid strbuf_split*()
  clean: do not pass strbuf by value
  clean: do not use strbuf_split*() [part 1]
  clean: do not pass the whole structure when it is not necessary
  clean: do not use strbuf_split*() [part 2]
  merge-tree: do not use strbuf_split*()
  notes: do not use strbuf_split*()
  config: do not use strbuf_split()
  environment: do not use strbuf_split*()
  sub-process: do not use strbuf_split*()
  trace2: trim_trailing_newline followed by trim is a no-op
  trace2: do not use strbuf_split*()
 builtin/clean.c      | 74 ++++++++++++++++++++--------------------
 builtin/merge-tree.c | 30 +++++++++--------
 builtin/notes.c      | 23 +++++++------
 config.c             | 23 ++++++-------
 environment.c        | 19 +++++++----
 sub-process.c        | 15 ++++-----
 trace2/tr2_cfg.c     | 80 +++++++++++++++-----------------------------
 wt-status.c          | 31 ++++++-----------
 8 files changed, 129 insertions(+), 166 deletions(-)
Range-diff against v2:
 1:  27de3d9a92 =  1:  2efe707054 wt-status: avoid strbuf_split*()
 2:  8f096e5a2d =  2:  899ff9c175 clean: do not pass strbuf by value
 3:  768b08907e =  3:  7a4acc3607 clean: do not use strbuf_split*() [part 1]
 -:  ---------- >  4:  4985f72ea5 clean: do not pass the whole structure when it is not necessary
 4:  0f8583e798 =  5:  4f60672f6f clean: do not use strbuf_split*() [part 2]
 5:  cefc2ec9f5 =  6:  d33091220d merge-tree: do not use strbuf_split*()
 6:  1c8ea097f6 !  7:  566e910495 notes: do not use strbuf_split*()
    @@ Metadata
      ## Commit message ##
         notes: do not use strbuf_split*()
     
    -    When reading the copy instruction from the standard input, the
    -    program reads a line, splits it into tokens at whitespace, and trims
    -    each of the tokens before using.  We no longer need to use strbuf
    -    just to be able to trimming, as string_list_split*() family now can
    -    trim while splitting a string.
    +    When reading copy instructions from the standard input, the program
    +    reads a line, splits it into tokens at whitespace, and trims each of
    +    the tokens before using.  We no longer need to use strbuf just to be
    +    able to trim, as string_list_split*() family now can trim while
    +    splitting a string.
     
    -    Retire the use of strbuf_split().
    +    Retire the use of strbuf_split() from this code path.
     
         Note that this loop is a bit sloppy in that it ensures at least
         there are two tokens on each line, but ignores if there are extra
 7:  a472688ec1 !  8:  dcecac2580 config: do not use strbuf_split()
    @@ Commit message
         config: do not use strbuf_split()
     
         When parsing an old-style GIT_CONFIG_PARAMETERS environment
    -    variable, the code parses the key=value pair by spliting them at '='
    -    into an array of strbuf's.  As strbuf_split() leafes the delimiter
    +    variable, the code parses key=value pairs by splitting them at '='
    +    into an array of strbuf's.  As strbuf_split() leaves the delimiter
         at the end of the split piece, the code has to manually trim it.
     
         If we split with string_list_split(), that becomes unnecessary.
    -    Retire the use of strbuf_split().
    +    Retire the use of strbuf_split() from this code path.
     
         Note that the max parameter of string_list_split() is of
         an ergonomically iffy design---it specifies the maximum number of
 8:  2b9957f31c =  9:  b894d4481f environment: do not use strbuf_split*()
 9:  4a5599836d = 10:  d6fd08bd76 sub-process: do not use strbuf_split*()
10:  cf6ecd2090 ! 11:  cb8e82a641 trace2: trim_trailing_newline followed by trim is a no-op
    @@ Commit message
         of a string.  If the code plans to call strbuf_trim() immediately
         after doing so, the code is better off skipping the EOL trimming in
         the first place.  After all, LF/CRLF at the end is a mere special
    -    case of whitespaces at the right end of the string, which will be
    -    removed by strbuf_rtrim().
    +    case of whitespaces at the end of the string, which will be removed
    +    by strbuf_rtrim() anyway.
     
         Signed-off-by: Junio C Hamano <gitster@pobox.com>
     
11:  c2578b6b1c ! 12:  838fe56920 trace2: do not use strbuf_split*()
    @@ Metadata
      ## Commit message ##
         trace2: do not use strbuf_split*()
     
    -    tr2_cfg_load_patterns() and tr2_load_env_vars() functions are copied
    -    and pasted pair of functions that each reads an environment
    -    variable, split the value at ',' boundaries and trims the resulting
    -    string pieces into an array of strbufs.  But the code paths that
    -    later use these strbufs take no advantage of the strbuf-ness of the
    -    result (they do not benefit from <ptr,len> representation to avoid
    -    having to run strlne(<ptr>), for example).
    +    tr2_cfg_load_patterns() and tr2_load_env_vars() functions are
    +    functions with very similar structure that each reads an environment
    +    variable, splits its value at the ',' boundaries, and trims the
    +    resulting string pieces into an array of strbufs.
    +
    +    But the code paths that later use these strbufs take no advantage of
    +    the strbuf-ness of the result (they do not benefit from <ptr,len>
    +    representation to avoid having to run strlen(<ptr>), for example).
     
         Simplify the code by teaching these functions to split into a string
    -    list instead.
    +    list instead; even the trimming comes for free ;-).
     
         Signed-off-by: Junio C Hamano <gitster@pobox.com>
     
     
-- 
2.50.1-633-g69dfdd50af
Previous: Patrick SteinhardtNext: Junio C Hamano
Message 60 of 72 in “string_list_split*() updates”
  1. 0/5 string_list_split*() updatesJunio C Hamano, Jul 31, 2025
  2. 1/5 string-list: report programming error with BUGJunio C Hamano, Jul 31, 2025
  3. Eric SunshineJul 31, 2025
  4. Junio C HamanoJul 31, 2025
  5. 2/5 string-list: align string_list_split() with its _in_place() counterpartJunio C Hamano, Jul 31, 2025
  6. Eric SunshineJul 31, 2025
  7. 3/5 string-list: unify string_list_split* functionsJunio C Hamano, Jul 31, 2025
  8. 4/5 string-list: optionally trim string pieces split by string_list_split()Junio C Hamano, Jul 31, 2025
  9. 5/5 diff: simplify parsing of diff.colormovedwsJunio C Hamano, Jul 31, 2025
  10. Eric SunshineJul 31, 2025
  11. 0/7 string_list_split*() updatesJunio C Hamano, Jul 31, 2025
  12. 1/7 string-list: report programming error with BUGJunio C Hamano, Jul 31, 2025
  13. 2/7 string-list: align string_list_split() with its _in_place() counterpartJunio C Hamano, Jul 31, 2025
  14. shejialuoAug 1, 2025
  15. Junio C HamanoAug 1, 2025
  16. shejialuoAug 1, 2025
  17. Junio C HamanoAug 1, 2025
  18. 3/7 string-list: unify string_list_split* functionsJunio C Hamano, Jul 31, 2025
  19. shejialuoAug 1, 2025
  20. 4/7 string-list: optionally trim string pieces split by string_list_split*()Junio C Hamano, Jul 31, 2025
  21. shejialuoAug 1, 2025
  22. Junio C HamanoAug 1, 2025
  23. shejialuoAug 1, 2025
  24. Junio C HamanoAug 1, 2025
  25. shejialuoAug 2, 2025
  26. Patrick SteinhardtAug 1, 2025
  27. Junio C HamanoAug 1, 2025
  28. 5/7 diff: simplify parsing of diff.colormovedwsJunio C Hamano, Jul 31, 2025
  29. Patrick SteinhardtAug 1, 2025
  30. 6/7 string-list: optionally omit empty string pieces in string_list_split*()Junio C Hamano, Jul 31, 2025
  31. Eric SunshineJul 31, 2025
  32. shejialuoAug 1, 2025
  33. Patrick SteinhardtAug 1, 2025
  34. Junio C HamanoAug 1, 2025
  35. 7/7 string-list: split-then-remove-empty can be done while splittingJunio C Hamano, Jul 31, 2025
  36. Patrick SteinhardtAug 1, 2025
  37. 0/7 string_list_split*() updatesJunio C Hamano, Aug 1, 2025
  38. 1/7 string-list: report programming error with BUGJunio C Hamano, Aug 1, 2025
  39. 2/7 string-list: align string_list_split() with its _in_place() counterpartJunio C Hamano, Aug 1, 2025
  40. Jeff KingAug 2, 2025
  41. Junio C HamanoAug 2, 2025
  42. Jeff KingAug 2, 2025
  43. 3/7 string-list: unify string_list_split* functionsJunio C Hamano, Aug 1, 2025
  44. 4/7 string-list: optionally trim string pieces split by string_list_split*()Junio C Hamano, Aug 1, 2025
  45. Jeff KingAug 2, 2025
  46. Junio C HamanoAug 2, 2025
  47. Jeff KingAug 2, 2025
  48. 5/7 diff: simplify parsing of diff.colormovedwsJunio C Hamano, Aug 1, 2025
  49. 6/7 string-list: optionally omit empty string pieces in string_list_split*()Junio C Hamano, Aug 1, 2025
  50. 7/7 string-list: split-then-remove-empty can be done while splittingJunio C Hamano, Aug 1, 2025
  51. 0/7 string_list_split*() updatesJunio C Hamano, Aug 3, 2025
  52. 1/7 string-list: report programming error with BUGJunio C Hamano, Aug 3, 2025
  53. 2/7 string-list: align string_list_split() with its _in_place() counterpartJunio C Hamano, Aug 3, 2025
  54. 3/7 string-list: unify string_list_split* functionsJunio C Hamano, Aug 3, 2025
  55. 4/7 string-list: optionally trim string pieces split by string_list_split*()Junio C Hamano, Aug 3, 2025
  56. 5/7 diff: simplify parsing of diff.colormovedwsJunio C Hamano, Aug 3, 2025
  57. 6/7 string-list: optionally omit empty string pieces in string_list_split*()Junio C Hamano, Aug 3, 2025
  58. 7/7 string-list: split-then-remove-empty can be done while splittingJunio C Hamano, Aug 3, 2025
  59. Patrick SteinhardtAug 4, 2025
  60. 00/12 do not overuse strbuf_split*()Junio C Hamano, Aug 3, 2025
  61. 01/12 wt-status: avoid strbuf_split*()Junio C Hamano, Aug 3, 2025
  62. 02/12 clean: do not pass strbuf by valueJunio C Hamano, Aug 3, 2025
  63. 03/12 clean: do not use strbuf_split*() [part 1]Junio C Hamano, Aug 3, 2025
  64. 04/12 clean: do not pass the whole structure when it is not necessaryJunio C Hamano, Aug 3, 2025
  65. 05/12 clean: do not use strbuf_split*() [part 2]Junio C Hamano, Aug 3, 2025
  66. 06/12 merge-tree: do not use strbuf_split*()Junio C Hamano, Aug 3, 2025
  67. 07/12 notes: do not use strbuf_split*()Junio C Hamano, Aug 3, 2025
  68. 08/12 config: do not use strbuf_split()Junio C Hamano, Aug 3, 2025
  69. 09/12 environment: do not use strbuf_split*()Junio C Hamano, Aug 3, 2025
  70. 10/12 sub-process: do not use strbuf_split*()Junio C Hamano, Aug 3, 2025
  71. 11/12 trace2: trim_trailing_newline followed by trim is a no-opJunio C Hamano, Aug 3, 2025
  72. 12/12 trace2: do not use strbuf_split*()Junio C Hamano, Aug 3, 2025

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.