[do-not-upstream] GNU toolchain Picolibc fixes (arc-2025.09) #24
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…umbers Handled invalid hexadecimal sequence passed to the `vfscanf` function by using `isxdigit` to check for the full hex range and detect invalid entry. Handled if the exponent in the floating point is a valid entry by using `isdigit` Signed-off-by: Mostafa Salman <[email protected]>
…ccuracy re-implement `casin/casinh` functions to match the standard relationship between these functions which is `casin(z) = -i * casinh(i*z)` this is numerically more stable with less cancelation errors avoiding the direct `sqrt(1 - z^2)` problematic calculation leading to more accurate results. Signed-off-by: Mostafa Salman <[email protected]>
for `rr ~= 1.0`, compute `log(rr)` as `0.5*log1p(rr^2 - 1)`, where `r^2 - 1` is calculated accurately to avoid cancellation errors. because direct `log(rr)` loses precision when `rr ~= 1.0` Signed-off-by: Mostafa Salman <[email protected]>
Special handling for purely real inputs where `x >= 1` to use the real `acoshf()` function which is more accurate for these cases and use the complex formula for other cases Signed-off-by: Mostafa Salman <[email protected]>
…inputs `catan()` functions needed special case handling for pure imaginary inputs instead of triggering the overflow case when it shouldn't Signed-off-by: Mostafa Salman <[email protected]>
According to the ISO/IEC_9899_1999, section:7.19.5.3 Opening file with `r+` mode just open text file for update (reading and writing) but dosen't create it. if the file isn't exist `open()` returns NULL. So, to create and open a file to update (reading and writing), we should open with "a+" or "w+" modes. Signed-off-by: Mostafa Salman <[email protected]>
…doesn't support stdin" This reverts commit [81e6fa2](picolibc@81e6fa2) As This commit introduced a workaround specifically for ARM that is no longer necessary or desirable. Signed-off-by: Mostafa Salman <[email protected]>
…ve values of width before this change, if width is zero code get stuck in an infite loop through negative values of width as it is defined as a signed int. Signed-off-by: Mostafa Salman <[email protected]>
- Extends the handling of the `%n` specifier to support different integer types, including `%ln` (long), `%lln` (long long), `%hn` (short) and `%hhn` (char). This ensures correct and consistent behavior for various specifier types. - Fix in vfprintf with %n when buffer has limited size before this change, the %n specifier returns the number of characters that would have been written if the buffer was large enough, instead of the actual number of characters written. Also, Added a gaurd if %n format specifier is enabled to protect printf functions from %n security issues Signed-off-by: Mostafa Salman <[email protected]>
| @@ -1249,7 +1266,9 @@ int vfprintf (FILE * stream, const CHAR *fmt, va_list ap_orig) | |||
| base = 2; | |||
| #endif | |||
| } else { | |||
| my_putc('%', stream); | |||
| while (--width > 0) { | |||
There was a problem hiding this comment.
This change seems unrelated to the rest of the patch?
There was a problem hiding this comment.
I read through the POSIX spec and your interpretation seems reasonable for this. However, I then created a testcase for this and discovered that glibc does exactly what the upstream picolibc code does -- it writes the number of characters that were generated by the operation, even if some of those are not actually written to the output because of the buffer size limit.
I'm struggling to understand how to read the POSIX spec in a way that makes glibc compliant, and yet glibc compatibility is at least as important as POSIX compliance for picolibc.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR is based on branch
arc-2025.09and contains a set of targeted fixes required specifically for the GNU toolchain. Tsome of these changes are not meant to be merged upstream, as they address toolchain-specific behaviors that aren't relevant to the official Picolibc repository.Includes the following commits
Note:
These commits should remain isolated from upstream synchronization to avoid merge conflicts and unintended code changes in the upstream branch.