KolibriOS/kolibrios
15
25
Fork 61
Code Issues 135 Pull Requests 31 Actions Projects 1 Releases Wiki Activity

WIP: libc.obj: implement gets_s #438

Draft
Egor00f wants to merge 4 commits from Egor00f/kolibrios:add-`gets_s` into main
pull from: Egor00f/kolibrios:add-`gets_s`
merge into: KolibriOS:main
Conversation 3 Commits 4 Files Changed 7
+50 -10
Egor00f commented 2026-05-02 09:43:20 +00:00
Collaborator
Copy Link
Copy Source

Changes

new functions:

  • gets_s

gets is defined as

#define gets(str) gets_s(str, STDIO_MAX_MEM)

and

char* _gets(char* str)
{
    return gets(str);
}

gets DIDN'T BECOME SAFE

## Changes ### new functions: + `gets_s` ### `gets` is defined as ```c #define gets(str) gets_s(str, STDIO_MAX_MEM) ``` and ```c char* _gets(char* str) { return gets(str); } ``` `gets` DIDN'T BECOME SAFE
Egor00f force-pushed add-`gets_s` from a80783a903 to d7adcd1caa 2026-05-02 09:50:05 +00:00 Compare
Egor00f force-pushed add-`gets_s` from d7adcd1caa to c3d1506c18 2026-05-19 06:18:00 +00:00 Compare
Egor00f force-pushed add-`gets_s` from c3d1506c18 to f814987094 2026-05-19 06:22:41 +00:00 Compare
Doczom approved these changes 2026-05-19 06:51:30 +00:00
Burer approved these changes 2026-05-19 11:15:24 +00:00
Dismissed
Burer left a comment
Owner
Copy Link
Copy Source

Problems I see:

  • 🟥 stdio.h's gets macro references STDIO_MAX_MEM.
    Fixed: stdio.h now #include <limits.h>, so STDIO_MAX_MEM is available wherever <stdio.h> is; the plain #include <stdio.h>; gets(buf); case compiles again.

  • 🟥 C89/C99 conformance & portability — only the ABI half is fixed.
    ABI part fixed: gets is exported again (libc.def keeps gets+gets_s; EXPORTS[] = { "gets", &_gets }), so pre-built binaries link.
    Still broken at C-source level: stdio.h has only #define gets(str) gets_s(...) and no real char *gets(char*) declaration (the function is the internal _gets, which user code can't portably name). The standard (C89 §4.1.6 / C99 §7.1.4) requires the actual function to stay reachable when masked by a macro. So &gets, (gets)(buf), #undef gets; gets(...), and extern char *gets(char*); redeclarations still fail to compile. This still hurts porting C89/C99 software to KolibriOS — the original concern.
    Fix: declare a real prototype in stdio.h — DLLAPI char* gets(char* str); — and name the function gets (drop the _gets indirection); keep the macro alongside if safe-by-default is wanted. The _gets wrapper is exactly what keeps the source-level breakage.

  • 🟨 ABI note — resolved: gets symbol kept (via &_gets), pre-built binaries no longer break.

  • 🟨 The gets compat macro preserves the old unsafe behavior.
    gets(buf) → gets_s(buf, STDIO_MAX_MEM=4096) regardless of the real buffer size — no added safety for existing callers; benefit exists only when code explicitly calls gets_s(buf, sizeof buf). Acceptable as a migration shim, but state it so it isn't mistaken for a fix of existing call sites.

  • 🟨 Not full C11 Annex K gets_s.
    Improved (now clears str[0] on EINVAL/EIO), but still no constraint handler. Acceptable for a minimal libc — it's "gets_s-like"; note it in the PR description.

  • 🟨 gets.c dropped its own #include <limits.h> while still using INT_MAX.
    n > (size_t)INT_MAX now relies on INT_MAX arriving transitively via stdio.h → limits.h. Include-what-you-use regression — re-add #include <limits.h> to gets.c; it depends on it directly.

  • 🟨 New sample gets_s_test.c is not wired into the build.
    Unlike abort_test/atexit_test in PR 340, it isn't added to Makefile/build_all.sh, so it won't be auto-built/tested. Also trailing whitespace on return 0;.

Problems I see: - [x] **🟥 `stdio.h`'s `gets` macro references `STDIO_MAX_MEM`.** Fixed: `stdio.h` now `#include <limits.h>`, so `STDIO_MAX_MEM` is available wherever `<stdio.h>` is; the plain `#include <stdio.h>; gets(buf);` case compiles again. - [ ] **🟥 C89/C99 conformance & portability — only the ABI half is fixed.** ABI part fixed: `gets` is exported again (`libc.def` keeps `gets`+`gets_s`; `EXPORTS[] = { "gets", &_gets }`), so pre-built binaries link. Still broken at C-source level: `stdio.h` has only `#define gets(str) gets_s(...)` and **no real `char *gets(char*)` declaration** (the function is the internal `_gets`, which user code can't portably name). The standard (C89 §4.1.6 / C99 §7.1.4) requires the actual function to stay reachable when masked by a macro. So `&gets`, `(gets)(buf)`, `#undef gets; gets(...)`, and `extern char *gets(char*);` redeclarations still fail to compile. This still hurts porting C89/C99 software to KolibriOS — the original concern. Fix: declare a real prototype in `stdio.h` — `DLLAPI char* gets(char* str);` — and name the function `gets` (drop the `_gets` indirection); keep the macro alongside if safe-by-default is wanted. The `_gets` wrapper is exactly what keeps the source-level breakage. - [x] **🟨 ABI note** — resolved: `gets` symbol kept (via `&_gets`), pre-built binaries no longer break. - [ ] **🟨 The `gets` compat macro preserves the old unsafe behavior.** `gets(buf)` → `gets_s(buf, STDIO_MAX_MEM=4096)` regardless of the real buffer size — no added safety for existing callers; benefit exists only when code explicitly calls `gets_s(buf, sizeof buf)`. Acceptable as a migration shim, but state it so it isn't mistaken for a fix of existing call sites. - [ ] **🟨 Not full C11 Annex K `gets_s`.** Improved (now clears `str[0]` on `EINVAL`/`EIO`), but still no constraint handler. Acceptable for a minimal libc — it's "gets_s-like"; note it in the PR description. - [ ] **🟨 `gets.c` dropped its own `#include <limits.h>` while still using `INT_MAX`.** `n > (size_t)INT_MAX` now relies on `INT_MAX` arriving transitively via `stdio.h → limits.h`. Include-what-you-use regression — re-add `#include <limits.h>` to `gets.c`; it depends on it directly. - [ ] **🟨 New sample `gets_s_test.c` is not wired into the build.** Unlike `abort_test`/`atexit_test` in PR 340, it isn't added to `Makefile`/`build_all.sh`, so it won't be auto-built/tested. Also trailing whitespace on ` return 0; `.
👀 1
Burer requested changes 2026-05-19 11:15:34 +00:00
Burer left a comment
Owner
Copy Link
Copy Source

See comment above.

See comment above.
Egor00f requested review from Doczom 2026-05-19 16:42:19 +00:00
Egor00f requested review from Burer 2026-05-19 16:42:32 +00:00
Egor00f added 3 commits 2026-05-19 17:13:59 +00:00
libc.obj: implement gets_s ee6df29e4d
libc.obj: gets_s: add more checks cda139c12e
return back gets as function && update gets_s && include limits.h in stdio.h && add gets_s test
Build system / Check kernel codestyle (pull_request) Successful in 25s
Details
Build system / Build (pull_request) Successful in 14m54s
Details
035ed0790b 
add
```c
str[0] = '\0';
````
on error in `gets_s`
Egor00f force-pushed add-`gets_s` from 810204d56f to 035ed0790b 2026-05-19 17:13:59 +00:00 Compare
Egor00f added 1 commit 2026-05-19 17:17:11 +00:00
add test to tests build
Build system / Check kernel codestyle (pull_request) Successful in 32s
Details
Build system / Build (pull_request) Successful in 14m49s
Details
66da18501a
Egor00f marked the pull request as work in progress 2026-05-20 04:51:06 +00:00
Egor00f added the CCategory/Libraries
Priority
Low
 labels 2026-06-08 17:01:58 +00:00
Some checks are pending
Build system / Check kernel codestyle (pull_request) Successful in 32s
Details
Build system / Build (pull_request) Successful in 14m49s
Details
Test PR / Build *
Required
This pull request has changes conflicting with the target branch.
  • programs/develop/ktcc/libc.obj/samples/Makefile
  • programs/develop/ktcc/libc.obj/samples/build_all.sh
View command line instructions

Checkout

From your project repository, check out a new branch and test the changes.
git fetch -u http://git.kolibrios.org/Egor00f/kolibrios add-`gets_s`:Egor00f-add-`gets_s`
git checkout Egor00f-add-`gets_s`
Sign in to join this conversation.
Reviewers
No Reviewers
Doczom
Burer
Labels
Clear labels
C
Category/Applications
Category/Drivers
Category/General
Category/Kernel
Category/Libraries
Eolite
FASM
FS
GSoC
This issue in GSoC program
 HardwareTested
HLL
Influence/Settings
Influence/Text/TYPO
IRCC
Kernel
Kind
Breaking
Breaking change that won't be backward compatible
Kind
Bug
Something is not working
Kind
Build
Kind
Documentation
Documentation changes
Kind
Enhancement
Improve existing functionality
Kind
Feature
New functionality
Kind
Security
This is security issue
Kind
Testing
Issue or pull request related to testing
Pay for the code
Paid task
PR
Conflicts
PR conflicts with main
PR
Dependent
This PR is dependent on another PR
Priority
Critical
The priority is critical
Priority
High
The priority is high
Priority
Low
The priority is low
Priority
Medium
The priority is medium
PR
Ready to merge
Pull request is ready for merge
PR
Request changes
Changes requested in pull request
PR
Review required
Reviewed
Confirmed
Issue has been confirmed
Reviewed
Duplicate
This issue or pull request already exists
Reviewed
Invalid
Invalid issue
Reviewed
Won't Fix
This issue won't be fixed
Status
Abandoned
Somebody has started to work on this but abandoned work
Status
Blocked
Something is blocking this issue or pull request
Status
Need More Info
Feedback is required to reproduce issue or to continue work
No Label C Category/Libraries
Priority
Low
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
ace-dent (Andrew) Burer Doczom (Mikhail Frolov) dunkaist (Ivan B) Egor00f (Егор Ярин) hidnplayr IgorA Leency (Kiril Lipatov) mxlgv (Max Logaev) ramenu rgimad sdongles (Serhii Sakhno) Sweetbread (Gleb Zaharov)
No Assignees
3 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: KolibriOS/kolibrios#438
Delete Branch "Egor00f/kolibrios:add-`gets_s`"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?